Early impact of 13-valent pneumococcal conjugate vaccine on pneumococcal meningitis—Burkina Faso, 2014–2015

Highlights • Burkina Faso introduced 13-valent pneumococcal conjugate vaccine (PCV13) in 2013.• By 2015, incidence of PCV13 pneumococcal meningitis decreased by 32%.• Large decreases occurred among children aged <1 year (76%) and 1–4 years (58%).


Introduction
Streptococcus pneumoniae is a leading infectious cause of global morbidity and mortality 1 and primary etiology of bacterial meningitis, along with Neisseria meningitidis and Haemophilus influenzae type b (Hib). 2 In the meningitis belt of sub-Saharan Africa, 3 pneumococcal meningitis has a distinct seasonality similar to that of meningococcal meningitis, high case fatality ratio (CFR), and predominance of serotype 1 disease in persons aged >5 years. [4][5][6][7] Burkina Faso, a West African country located entirely within the meningitis belt, experiences hyper-endemic rates of meningitis. 8 Historically, most meningitis cases were due to N. meningitidis serogroup A. 3 Following the successful introductions of Hib vaccine in 2006 and MenAfriVac TM in 2010, 3 S. pneumoniae became the primary bacterial meningitis pathogen. In 2013, Burkina Faso introduced 13-valent pneumococcal conjugate vaccine (PCV13) into the routine infant immunization program.
To date, 55 Gavi-supported countries have introduced pneumococcal conjugate vaccines (PCVs) into routine immunization programs. 9 To help sustain these programs, evaluating PCV impact on pneumococcal disease and circulating serotypes is key. PCV introduction resulted in substantial decreases in invasive pneumococcal disease (IPD) and pneumonia among children globally, with most dramatic decreases observed among young children targeted for vaccination. [10][11][12][13][14][15][16] In addition to direct effects on disease, widespread use of PCVs can also reduce nasopharyngeal carriage of vaccinetype pneumococci among both vaccinated and unvaccinated individuals. 11 The resulting herd protection decreases PCVtype IPD incidence in unvaccinated children and adults. 17,18 Data on PCV impact in Asia and West Africa are limited. 14 It is important to evaluate whether PCV delivered on the World Health Organization (WHO) Expanded Program on Immunization routine schedule of three early-infancy doses can control disease in high-transmission settings, such as men-ingitis belt countries, as effectively as PCV programs in other settings.
We previously described case-based meningitis surveillance data prior to PCV13 introduction (2011)(2012)(2013) in Burkina Faso -one of the few African countries to successfully implement case-based surveillance nationwide. 19 The highest pneumococcal meningitis incidence and mortality occurred among children aged <1 year, and 71% of cases were due to PCV13 serotypes. Here, we evaluate early PCV13 impact by comparing pneumococcal meningitis surveillance data from before (2011-2013) and after (2014 and 2015) PCV13 introduction in Burkina Faso, thereby helping fill the information gap on PCV impact in West Africa.

PCV13 vaccination
PCV13 was introduced into the routine immunization program nationwide on October 31, 2013, with doses administered to children aged 8, 12, and 16 weeks. No catch-up doses were recommended for older children.

National surveillance system
Burkina Faso has collected high-quality case-based meningitis surveillance data since 2010, and joined MenAfriNet in 2015. 19,20 Case-level demographic and clinical information, as well as cerebrospinal fluid (CSF) specimens, are collected from all suspect meningitis cases in all 63 districts using WHO and MenAfriNet instruments 19,21 and tested at five national reference laboratories.
According to WHO case definitions, 22 a suspected meningitis case has sudden onset of fever ≥38.5°C with neck stiffness, altered consciousness, or other meningeal signs (including flaccid neck, bulging fontanel, or convulsions in young children). A laboratory-confirmed pneumococcal meningitis case is a suspected case with S. pneumoniae isolated from CSF by culture or detected in CSF by real-time polymerase chain reaction (rt-PCR) or latex agglutination using laboratory methods previously described. 19 Single-target rt-PCR was unable to differentiate some genetically similar serotypes (e.g., 12F/12A/12B/44/46).

PCV13 vaccination status among PCV13-eligible children
Vaccination records were sought for children with pneumococcal meningitis born after August 1, 2013 and therefore potentially eligible to receive PCV13. As PCV13 vaccination status is not routinely reported, the reporting districts were asked to retrospectively abstract the case child's local vaccination records, using a standardized procedure and data collection tool.

Statistical methods
We analyzed meningitis cases diagnosed from January 1, 2011 to December 31, 2015. A pre-PCV13 period (2011-2013) and a post-PCV13 period (2014 and 2015) were defined. Cases in non-residents of Burkina Faso were excluded from analyses.
Administrative PCV13 coverage was calculated by dividing the reported number of administered doses by population estimates for children aged 0-11 months projected from the 2006 national census. As PCV13 is sometimes given to children outside the eligible age range and population denominators are estimated, administrative coverage can exceed 100%. In 2015, Burkina Faso had a total population of 18,450,494 and an estimated birth cohort of 732,675.
Pneumococcal meningitis cases were categorized as due to PCV13 serotypes (1, 3, 4, 5, 6A/6B, 7F/7A, 9V/9A, 14, 18C/18F/18B/18A, 19A, 19F, or 23F), non-PCV13 serotypes, or non-typeable strains. Annual incidences (cases per 100,000 persons) were calculated using projected age-stratified population census estimates. Within each age stratum (<1 years, 1-4 years, 5-9 years, 10-14 years, and ≥15 years), the number of cases confirmed by culture or rt-PCR as S. pneumoniae was divided by the number of cases with CSF tested via culture or rt-PCR at a national laboratory. This proportion was then applied to the number of suspected meningitis cases within that age stratum for which no diagnostic testing was performed; this number was then added to confirmed cases to calculate the adjusted incidence. Comparisons of the characteristics of suspected meningitis cases that were tested vs. not tested at a national laboratory in Burkina Faso from 2011-2015 have been previously described. 20 To calculate the incidence of PCV13, non-PCV13, and non-typeable pneumococcal meningitis, the adjusted pneumococcal meningitis incidence in each age group was multiplied by the proportion of serotyped cases in each category. Percentage change ([relative risk − 1] × 100) in incidence with 95% confidence intervals (95%CI) was calculated using the mean incidence pre-PCV13 (2011-2013) and each year post-PCV13 (2014 and 2015) and the Poisson distribution for incidence rates. CFRs were calculated by dividing the number of reported deaths by the total number of cases.

Study approval
This analysis was approved by Burkina Faso Ministry of Health ethical committee and was determined by the Centers for Disease Control and Prevention's Human Research Protection Office to be public health non-research.

Role of the funding source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Data completeness and quality
From 2011-2015, 18,538 suspected meningitis cases were reported; nearly all (97%) had CSF collected (Supplementary  Table S1). Ninety percent of CSF specimens were tested by Gram stain, 22% by latex, 22% by culture, and 46% by rt-PCR.
In total, 61% were tested by latex, culture, or rt-PCR. The percentage of all CSFs tested via culture or rt-PCR at a national laboratory increased from 45% in 2011 to 74% in 2015.
S. pneumoniae detection varied by diagnostic test (Supplementary Table S2). In 2014-2015, 318 laboratoryconfirmed pneumococcal meningitis cases were tested using latex, of which 275 (86.5%) were positive for S. pneumoniae; comparatively, 46 (22.6%) of 204 tested by culture grew S. pneumoniae. The majority (984/1053; 93.4%) of pneumococcal meningitis cases were confirmed using rt-PCR, with or without another positive test. Although testing practices varied by year, sensitivity of each of the three methods remained consistent.

Discussion
Our findings suggest an early direct effect of Burkina Faso's infant PCV13 program, as PCV13-serotype meningitis inci- dence among children aged <1 year was 76% lower in 2015 compared to pre-PCV13, equivalent to an absolute decline of 13.1 cases per 100,000. PCV13 impact appeared to be greater for PCV13 serotypes besides serotype 1 (79% decrease) than for serotype 1 (59% decrease) among this vaccineeligible age group. While children aged 1-4 years experienced a smaller percentage decrease in PCV13-serotype incidence than in non-PCV13 serotype incidence (58% vs. 90%, respectively), they experienced a greater absolute decrease in PCV13-serotype incidence than in non-PCV13 serotype incidence (2.1 cases per 100,000 vs. 0.5 cases per 100,000, respectively). Whether the PCV13 program has indirect benefits -reducing disease among those too old to be vaccinated -was not yet clear from this early data. The reason for the decline in non-PCV13 serotype incidence is unclear, though an increase in non-PCV13 serotype disease from 2011 to 2012 (2.5 to 8.3 cases per 100,000) drives this decline. A large meningococcal meningitis epidemic occurred in 2012, 23 which doubled the annual number of suspected meningitis cases and may have increased healthcare-seeking behavior or collection and testing of CSF specimens. Environmental conditions driving the meningo-coccal epidemic could have also increased individual susceptibility to pneumococcal meningitis. Therefore, 2011 and 2013 are considered as more stable baseline years; comparing 2014-2015 to 2011/2013 reveals significant declines in PCV13 serotypes but not in non-PCV13 serotypes, while suspected, probable, meningococcal, and H. influenzae meningitis incidence remained stable. 20 Alternatively, the years used as a pre-PCV13 baseline (2011-2013) could represent a natural high point in the cyclical nature of pneumococcal meningitis incidence. In the absence of more years of quality pre-PCV13 surveillance data, it is difficult to establish a true baseline.
Serotype 1 continues to dominate as a cause of pneumococcal meningitis following PCV13 introduction, causing 56% of pneumococcal meningitis in all age groups and 64% of cases among those aged ≥5 years. Similar to meningococcal serogroups, serotype 1 has natural variations in incidence over time and can cause outbreaks when population immunity wanes. 24 It is difficult to know what the natural trend of serotype 1 in 2014-2015 would have been in the absence of PCV13 introduction. Also, serotype 1 transmission may differ from other serotypes, as it is not commonly found in  pneumococcal carriage studies. 25 As in The Gambia, 14 these factors make it difficult to assess PCV13 impact on serotype 1, although our data clearly show that PCV13 is not yet controlling serotype 1 disease. PCV13 given on a compressed schedule (3 doses given at 1-month intervals) early in life may not be as effective against serotype 1 as it is against other PCV13 serotypes circulating in Burkina Faso. More birth cohorts may need to receive PCV13 or a booster dose may be needed 26 before serotype 1 is better controlled.
In 2014-2015, 25% of pneumococcal meningitis infections among children aged <1 year were due to non-PCV13 serotypes. In particular, 65% of non-PCV13 serotypes were 12F/ 12A/12B/44/46, serotypes previously reported to have a high prevalence in the region and which may be considered for inclusion in future PCV formulations. Serotype 12F is emerging as an important cause of IPD with the propensity to cause outbreaks. 27,28 We also found a high proportion (16%) of nontypeable strains in 2014-2015, despite using both rt-PCR and conventional PCR methods. This is consistent with 2011-2013 19 and with previous studies in the region, 7 and may reflect circulating serotypes not covered by existing detection methods or low antigen levels in the tested specimens. Due to the low proportion of culture confirmation, the majority of serotyping was performed on culture-negative or culture-not-performed specimens, which may have limited the ability to identify specific serotypes.
Because surveillance only captured meningitis cases, we were unable to estimate PCV13 impact on other clinical syndromes such as pneumonia and bacteremia, which may have a different serotype distribution. However, decreases in pneumococcal meningitis incidence among children aged <1 year observed in 2014 (49%) and 2015 (68%) were comparable to the decreases in IPD incidence reported in South Africa (31% among children aged <1 year by one year postintroduction) and The Gambia (55% among children aged <2 years by three years post-introduction). 14,29 Additionally, focusing exclusively on meningitis may preclude a full understanding of potential herd protection, as PCVs have been shown to decrease pneumonia among older age groups. 13 Burkina Faso introduced PCV13 into the routine immunization program without a catch-up campaign, so full herd effects may not be seen until more birth cohorts are vaccinated or a booster dose is added to the schedule. 30 Monitoring changes in pneumococcal carriage and additional years of surveillance will aid our understanding of PCV13 impact and pneumococcal transmission dynamics. 25 As this is an ecologic study, we cannot attribute all changes in pneumococcal meningitis incidence directly to PCV13. However, ecologic studies remain important for evaluation of vaccination programs with high coverage and potential herd protection. While meningitis incidence varied considerably from year to year, children aged <1 year were the only age group to experience a lower pneumococcal meningitis incidence in both years post-introduction as compared to the three prior years. Despite surveillance quality improvement over time, challenges regarding specimen transport and laboratory confirmation in Burkina Faso remain; however, reported incidence rates adjust for changes in culture and rt-PCR testing capacity over time and likely reflect true trends in incidence.
This analysis adds to the literature regarding PCV13 impact in West Africa 14 using nationwide population-based data and  routine pneumococcal serotyping. Continued vigilant surveillance and serotyping in Burkina Faso is crucial to monitor medium-and long-term impacts of PCV13, including changes in incidence among young children, herd protection among older age groups, impact on serotype 1 disease, and potential serotype replacement. Surveillance is also needed to rapidly detect any pneumococcal meningitis outbreaks, such as the serotype 1 outbreak experienced in Ghana in 2015-2016, 24 which occurred among children aged ≥5 years despite two years of routine infant PCV13 immunization. This early evaluation of PCV13 impact shows encouraging results, and if trends continue, PCV13 may build on the success of Hib vaccine and MenAfriVac TM in reducing the burden of bacterial meningitis in Burkina Faso.