Spatiotemporal Analysis of Serogroup C Meningococcal Meningitis Spread in Niger and Nigeria and Implications for Epidemic Response

Abstract

Background: After the re-emergence of serogroup C meningococcal meningitis (MM) in Nigeria and Niger, we aimed to re-evaluate the vaccination policy used to respond to outbreaks of MM in the African meningitis belt by investigating alternative strategies using a lower incidence threshold and information about neighboring districts.

Methods: We used data on suspected and laboratory-confirmed cases in Niger and Nigeria from 2013 to 2017. We calculated global and local Moran's I-statistics to identify spatial clustering of districts with high MM incidence. We used a Pinner model to estimate the impact of vaccination campaigns occurring between 2015 and 2017 and to evaluate the impact of 3 alternative district-level vaccination strategies, compared with that currently used.

Results: We found significant clustering of high incidence districts in every year, with local clusters around Tambuwal, Nigeria in 2013 and 2014, Niamey, Niger in 2016, and in Sokoto and Zamfara States in Nigeria in 2017.We estimate that the vaccination campaigns implemented in 2015, 2016, and 2017 prevented 6% of MM cases. Using the current strategy but with high coverage (85%) and timely distribution (4 weeks), these campaigns could have prevented 10% of cases. This strategy required the fewest doses of vaccine to prevent a case. None of the alternative strategies we evaluated were more efficient, but they would have prevented the occurrence of more cases overall.

Conclusions: Although we observed significant spatial clustering in MM in Nigeria and Niger between 2013 and 2017, there is no strong evidence to support a change in methods for epidemic response in terms of lowering the intervention threshold or targeting neighboring districts for reactive vaccination.

Keywords: Niger; Nigeria; epidemic response; meningitis; vaccine.

Figure 1.

Cumulative annual district-level incidence of suspected cases of meningitis in Niger and Nigeria 2013 to 2017. Centers of significant clusters of high incidence by Anselin’s local Moran’s I outlined in black. High-incidence outlier in 2016 indicated with asterisk. Gray areas represent districts for which no data are reported.

Figure 2.

Observed and modeled weekly case counts, annual cases averted by reactive vaccination, and doses vaccine delivered in Niger and Nigeria 2013 to 2017.

Figure 3.

Number of interventions (individual districts vaccinated), doses, proportion of total cases averted, and number needed to vaccinate to prevent a case over the period 2013–2017 for different reactive vaccination strategies using polysaccharide ACW vaccine. Points show estimates for 4-week delays, lines show 2- and 6-week delays, with arrow heads indicating shorter delays.

Figure 4.

Sensitivity, specificity, positive predictive value, and negative predictive value of different targeting strategies for predicting cumulative annual incidence between 20 and 100 suspected cases per 100 000.