Of the 329 patients included in the study, 49 (14.8%) tested positive for respiratory viruses (Table ). Of these, five (1.5%) were positive for H1N1pdm09, 23 (6.9%) for influenza B, 10 (3.0%) for HRSV and 11 (3.3%) for PIV-1 and −3 (Table and Figure B). The average age of positive patients was 41 months (range, 1 month to 9 years), with no difference in the age or gender distribution. All samples were negative for human metapneumovirus.
Respiratory viruses identified in swab samples from 329 patients with influenza-like illness in Bangui, Central African Republic, by age group
The first case of H1N1pdm09 was detected in the CAR on 26 July 2010 in a child aged 36 months, followed by a second case on 11 September in an infant aged 16 months, and three cases were found in young children on 4, 11 and 22 October (Figure B). Of the five children, only one was admitted to intensive care for respiratory distress; no deaths were recorded. A 253-bp fragment of the haemagglutinin gene of the five H1N1pdm09 strains [accession numbers: CY092425, CY092426, CY092427, CY092428, CY092429] showed high sequence similarity (99–100%) to influenza A/California/04/2009 (H1N1) (data not shown).
The 23 specimens positive for influenza B represented 82.1% of the influenza viruses detected. Ten of the patients were hospitalized; one infant aged 9 months died 2 days after admission to intensive care with a clinical picture of severe acute respiratory illness. Influenza B virus was implicated in respiratory infections throughout the year, with two peaks: a first (11 positive samples) at the beginning of the rainy season and a second (8 positive samples) in November at the end of the rainy season (Figure B).
Partial genome analysis of nine amplified samples of influenza B virus [accession numbers: HE803088, HE803089, HE803090, HE803091, HE803092, HE803093, HE803094, HE803095] demonstrated high nucleotide similarity (98%) to vaccine strain B/Victoria/02/1987, except for one that belonged to the Yamagata lineage (data not shown).
HRSV was detected in 10 patients (Table ) with a median age of 28 months, mainly in November (Figure B). One infant aged 10 months died 1 day after admission to intensive care with a clinical picture of fever, cough, rhinitis, bronchiolitis, dyspnoea and myalgia. Fever and rhinitis were recorded in 9 of the 10 patients, and all had cough. No viral co-infections were reported. Analysis of the nucleotide sequences of six isolates [accession numbers: HE803082, HE803083, HE803084, HE803085, HE803086, HE803087] showed sequence similarity to both genotypes A and B.
PIV-3 was detected in seven patients (median age, 16 months; range, 5–48 months) and PIV-1 in four infants (median age, 11 months; range, 3–24 months) between April and November 2010 (Figure B).
Respiratory viruses were found in 14.8% of the 329 collected samples, showing the presence of H1N1pdm09 infections in the CAR for the first time. We found that influenza B, PIV-1, PIV-3 and HRSV were also involved in ILI in the country. Most of the cases were detected during the rainy season.
H1N1pdm09 infection in the CAR was first described in July 2010. In other African countries, the virus was shown to have been introduced by travellers [12
], but its source in the CAR has not been elucidated. All four cases detected were indigenous, with no history of travel or contact with a person returning from a country with declared cases. Extensive investigations of contacts of the confirmed cases did not reveal any other cases, suggesting low dissemination of H1N1pdm09 in the country. The clinical picture of H1N1pdm09 infection was similar to that of seasonal influenza circulating before the pandemic. This is in accordance with studies showing relatively low transmissibility and severity of H1N1pdm09 [15
]. Consequently, H1N1pdm09 did not appear to have had a significant public health impact in this area of the world.
In the present study, influenza B virus was the most commonly detected respiratory virus (n=23), whereas in similar studies in Africa HRSV was the most frequent causative virus of ILI [17
]. Influenza B virus caused respiratory infections throughout the year, with two peaks: at the beginning of the rainy season (June–July) and at the end of the rainy season (November). Antigenically and genetically distinct lineages of influenza B virus, influenza B/Victoria and B/Yamagata viruses have circulated in the CAR [20
HRSV is a major cause of ILI among infants and children worldwide [21
] and is the most frequently detected respiratory virus in both developed and developing countries [17
]. In this study, HRSV was found in only 2.7% of the samples and represented 18.3% of the viruses detected. The difference from other studies might be due to a different epidemiology of HRSV in the CAR, a landlocked country in central Africa, or to different inclusion criteria or detection techniques. Most of HRSV cases detected in our study occurred between November and February, corresponding to the dry season in the CAR.
Another interesting finding was the relatively low prevalence of respiratory viruses in the children with ILI, which were present in only 14.9% of samples; in similar studies, as many as 50% of samples contained respiratory viruses [17
]. In the CAR, vaccines against Streptococcus pneumoniae
and Haemophilus influenzae
were introduced in the enlarged programme of vaccination only recently (H. influenzae
in September 2008 and S. pneumoniae
in July 2011). In a study in Bangui in 1995, these bacteria were found in hundreds of ill children aged less than 5 years [27
]. It is therefore likely that these bacteria are still very common or even the main causes of respiratory infections in the CAR, as in other countries before the introduction of vaccines, when most severe respiratory infections were due to bacterial infections, and S. pneumoniae
and H. influenzae
were the commonest bacterial causes [28
]. This may explain, at least partly, the differences between our results and those of other studies. It is also possible that viruses other than these we looked for were involved, such as human rhinovirus, human bocavirus, human coronavirus or adenovirus.
The main limitation of our study is the limited sample size, which prevented us from investigating associations between clinical outcome and viral etiology. Another weakness is a bias toward younger patients, so that we could not assess whether a particular age group is at greater risk for a specific infection. Owing to the design of the study, it is also difficult to determine prevalence from the results. Therefore, further studies are needed to evaluate the burden of all respiratory viruses infections in the general population of the CAR.
Our study does highlight the importance of the clinical, epidemiological and virological network for influenza surveillance in the CAR [31
We reported here the first data on the etiology of ILI in the CAR. This will help central African clinicians to provide better care and treatment for patients presenting with ILI, including better use of antibiotics. Further studies with more patients are needed to confirm the burden of viral respiratory diseases in the CAR. Collection of samples from healthy control children may also enable comment on virus detection and disease association. Another suggestion for future studies is to collect data about underlying health status of children as risk factors such as HIV infection, malaria, malnutrition, etc. might have an impact on the acute respiratory infections [33
]. The temporal patterns detected should be assessed over many years in order to identify long-term seasonal patterns.