Estimating global and regional morbidity from acute bacterial meningitis in children: assessment of the evidence

Aim

To estimate global morbidity from acute bacterial meningitis in children.

Methods

We conducted a systematic review of the PubMed and Scopus databases to identify both community-based and hospital registry-based studies that could be useful in estimation of the global morbidity from bacterial meningitis in children. We were primarily interested in the availability and quality of the information on incidence rates and case-fatality rates. We assessed the impact of the year of study, study design, study setting, the duration of study, and sample size on reported incidence values, and also any association between incidence and case-fatality rate. We also categorized the studies by 6 World Health Organization regions and analyzed the plausibility of estimates derived from the current evidence using median and inter-quartile range of the available reports in each region.

Results

We found 71 studies that met the inclusion criteria. The only two significant associations between the reported incidence and studied covariates were the negative correlation between the incidence and sample size (P < 0.001) and positive correlation between incidence and case-fatality rate (P < 0.001). The median incidence per 100 000 child-years was highest in the African region – 143.6 (interquartile range [IQR] 115.6-174.6), followed by Western Pacific region with 42.9 (12.4-83.4), the Eastern Mediterranean region with 34.3 (9.9-42.0), South East Asia with 26.8 (21.0-60.3), Europe with 20.8 (16.2-29.7), and American region with 16.6 (10.3-33.7). The median case-fatality rate was also highest in the African region (31.3%). Globally, the median incidence for all 71 studies was 34.0 (16.0-88.0) per 100 000 child-years, with a median case-fatality rate of 14.4% (5.3%-26.2%).

Conclusions

Our study showed that there was now sufficient evidence to generate improved and internally consistent estimates of the global burden of acute bacterial meningitis in children. Although some of our region-specific estimates are very uncertain due to scarcity of data from the corresponding regions, the estimates of morbidity and case-fatality from childhood bacterial meningitis derived from this study are consistent with mortality estimates derived from multi-cause mortality studies. Both lines of evidence imply that bacterial meningitis is a cause of 2% of all child deaths.Meningitis is an infectious disease affecting the brain membrane and spinal cord (1). Globally, bacterial meningitis is the most severe type of meningitis, mainly caused by a triad of species Neisseria meningitidis, Streptocccus pneumonia, and Haemophilus influenzae (2). While viral meningitis is usually a self-limiting disease with good prognosis, bacterial meningitis is potentially fatal, requiring urgent medical assistance and management with antibiotics treatment (3). Various estimates of the burden of bacterial meningitis have been proposed to date, but they have mainly focused on mortality (4), long-term sequels (5), or etiology-specific morbidity and mortality (6-8).Interestingly, there have been no comprehensive attempts to estimate the overall global burden of bacterial meningitis in children. This is not surprising, because such attempt would face almost insurmountable methodological challenges. First, there is a problem with case definition of “bacterial meningitis” (9). In low resource settings, where the problem is most common, many children may present with “purulent meningitis,” whose cause is highly likely bacterial, but laboratory capacity may not be sufficient to isolate the causal agent and confirm the diagnosis. This leads to a discrepancy between morbidity burden estimates based on “all purulent meningitis” and “laboratory confirmed meningitis” – the latter always being lower than the former, but to a different extent in different contexts (10). The second large methodological obstacle is the problem of “meningitis belt.” The meningitis belt is the band of countries extending from Senegal to Ethiopia, characterized by semi-arid climate, dry seasons, and dusty winds, with seasonal outbreaks of meningococcal meningitis being recorded since the beginning of the 20th century (11). The problem with these epidemics is that they can last for several years and dramatically change the importance of meningococcus in comparison to the other two bacterial agents (S. pneumoniae and H. influenzae) both regionally and globally (11). This makes it difficult to express the “burden of disease” for any given year, because it will be very different in intra-epidemic and inter-epidemic years. Moreover, the extent of vaccine coverage against N. meningitidis, S. pneumoniae and H. influenzae is changing the burden rapidly and rather dramatically in many places, rendering the scarce evidence from before the period of vaccination rather useless and indicates a need of revision (12). Finally, the emergence of HIV/AIDS pandemic led to a substantial number of infected children, whose resistance to other infections is impaired and they present a specific category of population in which the rates of incidence and case-fatality rates may be very different from those in other children (13).It is apparent that meningitis continues to contribute significantly to global mortality and morbidity, but the impact of the efforts to control it is difficult to estimate given that we do not have comprehensive estimates of global morbidity patterns. Understanding the global morbidity from bacterial meningitis would be useful because it would also help to validate the existing mortality estimates through application of appropriate case-fatality rates. The purpose of the present study is to provide a comprehensive assessment of the evidence that is available for estimating the global morbidity from acute bacterial meningitis in children globally. We will also propose initial, robust estimates of the burden, with suggestions on the possible ways to address the methodological challenges in future studies.