Background
Bacterial meningitis in sub-Saharan Africa has mainly been caused by
N. meningitidis (Nm) with localized outbreaks or large epidemics occurring every year in the dry season [
1]–[
4]. The number of cases usually increases in January and drops in April-May with the first rainfall.
Nm is normally colonising the upper respiratory tract in humans without causing disease [
3],[
5],[
6]. The bacteria circulate in the human population by transmitting from person to person through droplets and close contact. However, occasionally Nm may infect the blood and cause meningitis and/or septicaemia, with a possible fatal outcome occurring within hours.
Nm is classified into serogroups according to the structure of its polysaccharide capsule. In the meningitis belt, epidemics have essentially been caused by Nm of serogroups A (NmA), but serogroups W (NmW) and X (NmX) have also caused large outbreaks [
7]–[
9].
As the capsule is immunogenic and probably the most important virulence factor, capsular polysaccharide of serogroups A, C, W and Y has been used as vaccine antigens. Until very recently, A/C and A/C/W polysaccharide vaccines have been the only vaccines available for use in sub-Saharan Africa. Due to the relative short protection provided by polysaccharide vaccines and their inability to elicit a good antibody response in children below 2 years of age, these vaccines have been used mainly to stop ongoing outbreaks [
10]. Following the success of
Haemophilus influenzae type b and pneumococcal conjugate vaccines, Nm vaccines coupling a carrier protein to the main disease-causing polysaccharide serogroups were developed. In addition to eliciting strong and long lasting immune response in children below 2 years [
5],[
11], conjugate vaccines have also been shown to confer herd protection by reducing the carriage prevalence, thus interrupting transmission [
12]–[
16].
A monovalent serogroup A tetanus toxoid-conjugated vaccine, MenAfriVac, has been developed with the goal of eliminating the devastating NmA epidemics in Africa [
17]–[
19]. The vaccine was shown to be safe and immunogenic [
20] and its low price made it appropriate for routine vaccination and mass vaccination campaigns in Africa. MenAfriVac was prequalified by the World Health Organization (WHO) in June 2010 to be given as a single dose in the age group 1–29 years.
Being a conjugate vaccine, it was hoped that MenAfriVac will have an effect on NmA carriage, similarly to what has been found in the UK following vaccination campaign with a monovalent serogroup C conjugate on NmC carriage [
14],[
15]. Given the fact that not all age-groups were eligible to receive the vaccine, impact of vaccination on meningococcal carriage would be beneficial for the whole population by reducing transmission. To demonstrate the impact of this new vaccine in an African setting, meningococcal carriage studies were planned ahead of MenAfriVac implementation in the meningitis belt.
Burkina Faso was the first country to vaccinate the whole 1-29-year-old population in 2010 [
21]. Mali and Niger introduced the vaccine in 2010–2011 and other countries of the meningitis belt rapidly followed; over 150 million people had received the vaccine by the end of 2013. The impact of MenAfriVac on NmA disease and carriage, and the evidence of herd immunity, were first demonstrated in Burkina Faso. The risk for meningitis significantly decreased in all age groups, not only in the 1–29 year-old vaccinated population, and based on laboratory-confirmed cases, the risk for NmA meningitis was reduced by 99.8% [
22]. Up to 13 months after vaccination, NmA carriage remained undetected among both vaccinated and unvaccinated populations [
23] and there was no evidence of capsule replacement of the ST-2859 clone previously responsible for NmA disease and carriage in Burkina Faso [
24]. Similar results were recently reported from Chad that experienced an epidemic in 2010; outbreaks of NmA meningitis occurred only in non-vaccinated districts and NmA carriage was significantly lower 4–6 months after vaccination [
25]. As the vaccine is not yet introduced in childhood vaccination programs, the impact of this public health intervention relies in large part on the herd protection afforded by the vaccine. It is thus important to document how long this protection will last through continued laboratory-based surveillance and follow-up studies on carriage.
To study the long term impact of MenAfriVac vaccination on carriage, an existing multicentre carriage study in Burkina Faso was extended and conducted in October-November 2012 using the same protocol as in previous years. Results from this carriage sampling together with molecular characterization of invasive isolates collected through surveillance are presented here.
Discussion
This study describes meningococcal carriage in Burkina Faso in October-November 2012, two years after the introduction of MenAfriVac. Molecular characteristics of the carriage isolates were compared to those of invasive isolates collected during the 2012 epidemic season. The significant reduction of NmA carriage after MenAfriVac vaccination still persisted. Serogroup W dominated both among carriage and patient isolates.
Although the small collection of invasive isolates that we analyzed was not a random selection, it reflected to some extent the characteristics of meningococcal disease isolates in Burkina Faso in 2012. Our collection was composed of 75% NmW and 20% NmX, which is comparable to the WHO surveillance data from the 2012 epidemic season showing that among the laboratory confirmed cases caused by Nm, 83.5% were NmW and 16.4% were NmX [
36]. Analysis of a larger sample of invasive isolates collected from Burkina Faso in 2012 newly confirmed that all the NmW isolates belonged to the ST-11 complex [
37]. A single NmY case was identified and this isolate was included in our strain collection. The surveillance system in Burkina Faso was strong in this period: of 6957 cases of suspect meningitis cases reported in 2012, 3292 CSF samples were collected, of which 1105 were reported with a laboratory result [
36]. No cases of NmA meningitis were reported during the 2012, 2013 and 2014 epidemic seasons, before and after the carriage study [
36].
This carriage study was the tenth sampling campaign performed in the same districts of Burkina Faso since 2009 [
23],[
24],[
26],[
27]. Efforts were made to keep all conditions as similar as possible to those in previous samplings. We used the same protocol, reagents, equipment and QC system. The teams were composed of essentially the same persons and retraining was conducted prior to start. The sensitivity and specificity of laboratory analysis in Burkina Faso, as extracted from the QC system, were comparable to previous campaigns [
23],[
26],[
27] and show that variation of carriage prevalence was not due to variation in the quality of the analysis. Furthermore, the carriage prevalence of
N. lactamica was within the range obtained from all the previous campaigns [
38], providing yet another indication that the sampling was well conducted.
The age group included in the carriage study was the same as the age group targeted for vaccination in 2010. Since the entire country was vaccinated within a short timeframe, the carriage study was not designed with a non-vaccinated control group. However, since MenAfriVac was not part of the Expanded Program of Immunization and the study was performed two years after the national mass vaccination campaign, 21.2% of the study participants were considered non-vaccinated (17.6% reported to not be vaccinated; 3.6% did not remember), and 70% of the 1-3-year-olds were not vaccinated. The low vaccination coverage compared to previous studies [
23],[
39] and the age distribution of non-vaccinated was consistent with the time delay between mass-vaccination in 2010 and the survey. Immigration might also contribute to explain the low coverage but we did not collect such data. The reliability of the participants’ responses to whether they had been vaccinated with MenAfriVac is questionable, however, because the survey was conducted two years after vaccination and only 54.5% of those responding “yes” presented a vaccination card. A relatively small number of 1- and 2-year-olds were reported to have received MenAfriVac, despite the fact that they were not eligible for vaccination in 2010. It is possible that these children in fact had been vaccinated in another country introducing the vaccine at a later time point, but it is more likely a recall bias.
Carriage of NmA was still low two years after the introduction of MenAfriVac, and together with the historically low NmA disease incidence post-MenAfriVac vaccination, our data support evidences of prolonged herd immunity established after the mass vaccination campaign [
23]. Despite the fact that 70% of the 1-3-year olds were unvaccinated, we did not find any NmA carrier in this age group. In comparison, pre-vaccination NmA carriage prevalence in the 1-3-year-olds was 0.30% [
26]. Surprisingly, the single NmA carriage isolate was ST-7 and not ST-2859, the only genotype responsible for NmA disease in Burkina Faso since 2004 [
35], and also the only NmA genotype found among carriers in the period 2008–2011 [
26],[
27]. The child who carried this ST-7 strain had not travelled in the period prior to sampling, but the household had recently received visitors from the Ivory Coast, and colonization of the child by a NmA ST-7 isolate might be a result of transmission from one of the visitors.
The introduction of a monovalent vaccine could result in capsule switch of the ST-2859 clone as a consequence of uptake of genes coding for another capsule without the loss of virulence, as has been described for other Nm clones [
40]–[
43]. We have previously reported that we did not observe any capsule switch up to 13 months after vaccination [
24] and we show here that this observation was still valid two years after vaccination.
The most striking result in this carriage study was the dominance of NmW ST-11 P1.5,2;F1-1, a clone that was responsible for a large NmW outbreak in Burkina Faso in 2000–2002. ST-11 was last detected among patients in Burkina Faso in 2005 and 2006 [
35],[
44] and was not found among carriers in 2008–2010 [
24],[
26], before it reappeared during the epidemic season of 2011 [
24]. It has previously been suggested that the re-introduction of ST-11 happened from Mali [
24], and our results support this hypothesis since the western district of Dandé had significantly higher NmW carriage prevalence when compared to Bogodogo and Kaya. Dandé is close to Mali and NmW caused districts on both sides of the Mali-Burkina Faso border to reach the epidemic threshold in 2012 [
36]. A reappearance of NmW ST-11 outbreaks ten years after the 2002 outbreak [
9],[
45] is consistent with the cyclic waves of meningococcal epidemics [
1],[
2]. Before MenAfriVac vaccination, the carriage prevalence of NmA was low and the ST-11 clone did not circulate. Thus, it is less likely that a vaccine-induced replacement of NmA ST-2859 with NmW ST-11 could explain the high proportion of NmW carriage and disease two years after vaccination.
The second most frequent clone among carriers and patients was NmX, ST-181 P1.5-1,10-1;F1-31. This same clone caused many cases of meningitis during the epidemic season of 2010, before MenAfriVac introduction [
35]. NmX carriage was high in October-November 2010 and decreased throughout 2011 [
23]. We show that carriage of NmX continued to decrease in 2012, but that the district of Kaya still had the highest prevalence of NmX. One of the culture-negative invasive NmX isolates had the unusual PorA variant P1.7,16-46. A nested PCR of the MLST loci attributed this isolate to the ST-181 complex, as the other NmX isolates.
Serogroup Y ST-4375 of the ST-23 complex was the dominating carriage clone circulating in 2009, before vaccine introduction [
26], but its carriage prevalence gradually declined in 2010–2011 [
23] and only 0.6% of the participants carried ST-4375 in 2012. The decline in carriage of the ST-23 complex meningococci is presented in Figure
1. Compared to other virulent clones, the ST-4375 clone is seldom associated with disease and has been only sporadically identified in patients in Burkina Faso in the past ten years, last in 2008 [
35]. Although various STs of the ST-23 complex have been causing an increasing number of cases in the US and Europe [
46]–[
48], only a few NmY cases of disease have been reported in Burkina Faso, or in other countries of the meningitis belt [
35].
Since the start of these carriage campaigns, the dominating serogroup has been Y (in 2009), X (in 2010–2011) and W (in 2012) [
23],[
26]. Carriage prevalence of each of these serogroups varied significantly with age only in the period when that serogroup was dominating. In 2012 when serogroup W dominated, there was a clear peak of carriage prevalence among the 5–9 year old; in 2011 when serogroup X was dominating, maximum carriage was in 5–14 year old [
23], and in 2009 when serogroup Y dominated, the peak was in the age group 10–14 years [
26].
We have also shown that the same genotypes were found among carriers and patients and that increased carriage of virulent clones was closely associated with increased disease incidence. In Burkina Faso NmX ST-181 carriage prevalence increased during the epidemic season of 2009, especially in the eastern district of Kaya [
26] (Figure
1). In the following epidemic season (2010) NmX was found to cause disease and towards the end of 2010, NmX ST-181 carriage was significantly higher in all 3 carriage study sites, and particularly in Kaya [
23],[
24] (Figure
1). The same pattern was observed for NmW ST-11, first seen among carriers in the western district of Dandé in early 2011, then spreading throughout the country [
24]. The following epidemic season NmW ST-11 was the main disease-causing genotype and by the end of 2012, this clone dominated among carriage strains and with the highest prevalence found in Dandé (Figure
1). The introduction of virulent clones in an immunologically naive population seems to be one of the important factors for meningococcal outbreaks [
49], together with the inherent properties of the various Nm serogroups and clones, and contact patterns between individuals within the society.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
PAK, IS, RO, LS, FD, DK, TAC, MPP and DAC participated in the design of the study. AKB, ASO, IS, RO, LS were responsible for collecting carriage isolates. RO and DK provided the clinical isolates. IMS conducted laboratory confirmation and molecular characterization of isolates. PAK and FD were responsible for coordination and supervision of the carriage study. PAK analyzed the data and drafted the manuscript. DAC conceived the study and was responsible for the molecular analysis. All the authors revised the manuscript and approved the final version.