Global, regional, and national burden of meningitis and its aetiologies, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019

1Meningitis is a disease defined by inflammation of the meninges, the layers of membranes that cover the brain and spinal cord. In addition to mortality, meningitis can result in long-term sequelae such as cognitive impairment, hearing loss, motor weakness or paralysis, incoordination, and epilepsy., Bacterial and viral infections are key causes of acute meningitis. Viral meningitis is more common than bacterial meningitis but is associated with lower rates of mortality and complications. By contrast, bacterial meningitis is more likely to be associated with a poor prognosis and require prompt treatment.,

2A systematic review published in 2018 found that the aetiologies responsible for the highest proportion of bacterial meningitis cases in all regions globally were Streptococcus pneumoniae and Neisseria meningitidis (or meningococcus), with Haemophilus influenzae also posing a high burden in selected regions. Although vaccines can prevent infections due to these three aetiologies, not all countries have fully immunised their populations. For example, global infant third-dose vaccine coverage for H influenzae type b (Hib) is 72%, and third-dose vaccine coverage for S pneumoniae is estimated to be only 51%. Rates of meningitis incidence and mortality are highest in the meningitis belt, which consists of 26 countries in Africa spanning from Senegal to Ethiopia. This region has historically reported epidemics of N meningitidis serogroup A. To reduce the incidence rate of meningitis, 24 of the 26 countries in the meningitis belt had introduced meningococcal serogroup A conjugate vaccine (MACV [MenAfriVac]) campaigns by the end of 2021., Globally, 53 countries have meningococcal conjugate vaccines in their routine immunisation schedule, including 27 countries with quadrivalent meningococcal ACYW vaccines and nine countries with meningococcal serogroup B vaccines.

3Evidence before this study The global burden of meningitis and the burden attributable to a subset of aetiologies have been estimated by multiple groups, including WHO and the Maternal and Child Epidemiology Estimation Group (WHO-MCEE) and the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD). We searched PubMed with the search terms “meningitis” [MeSH] AND (“mortality” OR “incidence”) AND “global”, with no language restrictions, for articles published from database inception to April 26, 2023. We did not identify any studies that evaluated global levels of and trends in meningitis burden attributable to a comprehensive set of aetiologies across countries. Added value of this study This study uses data from GBD 2019 and improves upon previous GBD estimates in several ways. First, we added many new data sources on meningitis morbidity and mortality since GBD 2017. Second, we used a standardised approach to enhance the comparability of non-fatal data sources using a novel Bayesian meta-regression tool. Third, we produced estimates of meningitis burden attributable to ten different aetiologies (Escherichia coli, Haemophilus influenzae, Klebsiella pneumoniae, Listeria monocytogenes, Neisseria meningitidis, Staphylococcus aureus, group B Streptococcus, Streptococcus pneumoniae, viruses, and a residual other pathogen category), six of which are new inclusions to GBD. Finally, we address a key limitation of earlier GBD reports by producing non-fatal and fatal aetiology proportions that are linked to each other, using location-specific, age-specific, year-specific, and aetiology-specific case-fatality ratio estimates. Implications of all the available evidence There has been considerable progress in reducing the burden of infectious meningitis since 1990. However, the reduction has not been equal across locations, and countries in the African meningitis belt and south Asia still have a large burden. Meningitis attributable to S pneumoniae, which has a high case-fatality ratio, is responsible for more deaths than any other bacterial aetiology. There is a need to maintain and increasecoverage of pneumococcal, H influenzae type b, and multivalent meningococcal vaccines. For all aetiologies, improved surveillance systems, early diagnosis (through low-cost, accurate, and rapid tests), and improvements in health-care access and treatment are necessary to reduce the burden of meningitis and track progress. Our analysis highlights the importance of controlling bacterial meningitis not attributable to vaccine-preventable aetiologies. K pneumoniae was associated with more mortality than other aetiologies that do not have an available vaccine and was the third-largest cause of deaths attributable to infectious meningitis overall. New vaccines and vaccine programmes are in development for K pneumoniae, group B Streptococcus, and other forms of bacterial meningitis. The development of, and access to, vaccines are essential in reducing the global meningitis burden and in preventing the over-reliance on antibiotics that worsens antimicrobial resistance.

4Evidence before this study

5The global burden of meningitis and the burden attributable to a subset of aetiologies have been estimated by multiple groups, including WHO and the Maternal and Child Epidemiology Estimation Group (WHO-MCEE) and the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD). We searched PubMed with the search terms “meningitis” [MeSH] AND (“mortality” OR “incidence”) AND “global”, with no language restrictions, for articles published from database inception to April 26, 2023. We did not identify any studies that evaluated global levels of and trends in meningitis burden attributable to a comprehensive set of aetiologies across countries.

6Added value of this study

7This study uses data from GBD 2019 and improves upon previous GBD estimates in several ways. First, we added many new data sources on meningitis morbidity and mortality since GBD 2017. Second, we used a standardised approach to enhance the comparability of non-fatal data sources using a novel Bayesian meta-regression tool. Third, we produced estimates of meningitis burden attributable to ten different aetiologies (Escherichia coli, Haemophilus influenzae, Klebsiella pneumoniae, Listeria monocytogenes, Neisseria meningitidis, Staphylococcus aureus, group B Streptococcus, Streptococcus pneumoniae, viruses, and a residual other pathogen category), six of which are new inclusions to GBD. Finally, we address a key limitation of earlier GBD reports by producing non-fatal and fatal aetiology proportions that are linked to each other, using location-specific, age-specific, year-specific, and aetiology-specific case-fatality ratio estimates.

8Implications of all the available evidence

9There has been considerable progress in reducing the burden of infectious meningitis since 1990. However, the reduction has not been equal across locations, and countries in the African meningitis belt and south Asia still have a large burden. Meningitis attributable to S pneumoniae, which has a high case-fatality ratio, is responsible for more deaths than any other bacterial aetiology. There is a need to maintain and increasecoverage of pneumococcal, H influenzae type b, and multivalent meningococcal vaccines. For all aetiologies, improved surveillance systems, early diagnosis (through low-cost, accurate, and rapid tests), and improvements in health-care access and treatment are necessary to reduce the burden of meningitis and track progress. Our analysis highlights the importance of controlling bacterial meningitis not attributable to vaccine-preventable aetiologies. K pneumoniae was associated with more mortality than other aetiologies that do not have an available vaccine and was the third-largest cause of deaths attributable to infectious meningitis overall. New vaccines and vaccine programmes are in development for K pneumoniae, group B Streptococcus, and other forms of bacterial meningitis. The development of, and access to, vaccines are essential in reducing the global meningitis burden and in preventing the over-reliance on antibiotics that worsens antimicrobial resistance.

10Reducing meningitis incidence and death rates is the focus of WHO's Defeating Meningitis by 2030 global roadmap, a plan developed in a global collaboration between key country stakeholders (ie, representatives from governments, global health organisations, public health bodies, academia, the private sector, and civil society) and international organisations. The global roadmap outlines three main goals: eliminating bacterial meningitis epidemics, reducing cases of and deaths from vaccine-preventable bacterial meningitis, and reducing disability and improving quality of life after recovery from acute meningitis. Robust and comparable estimates of meningitis incidence and mortality are invaluable in tracking progress and identifying gaps towards achieving these goals.

11Here, we present results from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2019, synthesised with results from the global burden of antimicrobial resistance study done by collaborators of the Global Research on Antimicrobial Resistance project, describing the burden and trends of acute infectious meningitis and ten aetiologies for 204 countries and territories from 1990 to 2019. Notable limitations of GBD 2016 and GBD 2017 were the divergent methods used for estimation of fatal and non-fatal aetiology proportions for S pneumoniae, N meningitidis, and Hib, and the absence of estimates for other important pathogens.,, As part of the study assessing the global burden of antimicrobial resistance, we developed a new method of pathogen estimation that addresses the limitations of these earlier GBD publications and provide the first estimates of meningitis burden attributable to a comprehensive set of pathogens. This manuscript was produced as part of the GBD Collaborator Network and in accordance with the GBD Protocol.

12GBD 2019 produces estimates of meningitis mortality and morbidity by age and sex for 204 countries and territories, for the years between 1990 and 2019. The study investigating the global burden of antimicrobial resistance produced estimates of the aetiology-specific fatal and non-fatal burden of selected infectious syndromes, including meningitis, between 1990 and 2019. Modelling was done at the 1000 draw level, and 95% uncertainty intervals (UIs) were computed as the 25th and 975th ranked values of 1000 draws. We used the GBD 2019 global population age standard to calculate age-standardised rates. Below, we summarise key methods from GBD 2019 and the study of the global burden of antimicrobial resistance for estimation of the burden of meningitis and its aetiologies. More details of these methods, including a flowchart for estimation of meningitis mortality and non-fatal burden, are provided in appendix 1 (pp 3, 8). Full descriptions of GBD 2019 and the study estimating the global burden of antimicrobial resistance have been previously published., We present our results stratified by Socio-Demographic Index (SDI) quintiles. SDI is a composite indicator computed based on three variables (income per capita, average years of schooling, and total fertility rate) for each country. All count data are presented to three significant figures and all proportions and rates are presented to one decimal place.

13The data used for estimation of mortality due to meningitis originated from vital registration, verbal autopsy, sample-based vital registration, minimally invasive tissue sampling from the Child Health and Mortality Prevention Surveillance (CHAMPS), and epidemiological surveillance, comprising a total of 24 726 site-years (21 852 site-years from vital registration, 825 site-years from sample-based vital registration, 1432 site-years from verbal autopsy, 611 site-years from surveillance sources, and six site-years from minimally invasive tissue sampling). The data were processed with a set of standard algorithms accounting for incompleteness, misclassification of the underlying cause of death, garbage coding, and stochastic variability. The International Classification of Diseases ninth revision (ICD-9) and ICD-10 codes mapped to meningitis are provided in appendix 1 (p 4).

14We estimated overall meningitis mortality using the Cause of Death Ensemble model (CODEm), which has widely been used to produce global estimates of cause-specific mortality.,, We modelled mortality separately for children aged 0–4 years and those aged 5 years and older because these two populations have very different meningitis mortality distributions and trends. The CODEm strategy evaluates various potential models with different combinations of covariates and model classes (mixed-effects linear models and spatiotemporal Gaussian process regression models). A full list of covariates is provided in appendix 1 (p 4). For this analysis, 26 countries in sub-Saharan Africa listed by WHO as being at risk for meningitis epidemics are considered meningitis belt countries. Models were weighted with out-of-sample predictive validity and combined in one ensemble model. Meningitis mortality estimates are scaled in a process called CoDCorrect so that all-cause mortality and the sum of cause-specific mortality are consistent.,

15The data used for morbidity estimation originated from published studies identified via a systematic review (appendix 1 p 9), surveillance data, cause-specific meningitis mortality estimates (calculated as described above), claims data, and inpatient data. We estimated overall meningitis morbidity using DisMod-MR (version 2.1), a Bayesian meta-regression tool that includes a compartmental model to estimate prevalence, incidence, remission, and mortality. Before modelling in DisMod-MR (version 2.1), we enhanced data comparability by applying a standardised approach to adjust claims and surveillance data to the level of inpatient data. More details of the adjustments are provided in appendix 1 (p 10).

16Data used for aetiology estimation originated from multiple cause of death vital registration data, hospital discharge data, microbial laboratory data, and published studies from the literature. We calculated mortality and morbidity estimates for each of ten aetiologies of meningitis: Escherichia coli, H influenzae, Klebsiella pneumoniae, Listeria monocytogenes, N meningitidis, Staphylococcus aureus, group B Streptococcus, S pneumoniae, viruses, and a residual other pathogen category. The ICD-9 and ICD-10 codes mapped to each aetiology are listed in appendix 1 (p 11). A key methodological improvement compared with GBD 2016 and GBD 2017 that we made in this study was to estimate the aetiologies using consistent methods for non-fatal and fatal proportions (ie, fatal aetiology proportions were derived from non-fatal aetiology proportions in combination with case-fatality ratio (CRF) estimates, as described below).

17First, CFRs for each pathogen were modelled with the Bayesian meta-regression tool MR-BRT (meta-regression Bayesian, regularised, trimmed). CFRs were calculated as a function of age group, pathogen, and the Healthcare Access and Quality (HAQ) Index, with random effects on data source.,, The HAQ Index is a composite indicator of 32 causes of amenable mortality that measures health-care access and quality over time. We additionally controlled for data provided only from intensive care units (which would be biased towards higher CFRs) and modelled the effects of the proportion of pneumococcal conjugate vaccine (PCV) and Hib vaccinations among individuals aged 15 years or younger on the CFRs for S pneumoniae and H influenzae, respectively. The modelled CFRs were used to back-calculate case quantities from data sources that reported only deaths. Next, incidence proportions were estimated with multinomial estimation as part of a network analysis model, which allows for the inclusion of data sources that are considered partial observations (ie, which do not contain all ten pathogen groups modelled in the study). More details of this approach are provided in the online appendix of the published study on the global burden of antimicrobial resistance (pp 39–44).

18The funder of the study had no role in study design, data collection, data analysis, data interpretation, or the writing of the report.

19In 2019, 2·51 million (95% UI 2·11–2·99) new cases of meningitis occurred globally among all age groups (table 1), of which 1·28 million (0·947–1·71) occurred in children younger than 5 years (table 2). Age-standardised incidence rates per 100 000 population were highest in western sub-Saharan Africa, with an incidence rate of 105·8 (90·1–124·1) per 100 000 population, followed by eastern sub-Saharan Africa with an incidence rate of 97·3 (83·4–112·3) per 100 000 population and central sub-Saharan Africa with an incidence rate of 87·2 (75·6–99·6) per 100 000 population. Overall, the age-standardised global incidence rate was 35·4 (29·6–42·5) cases per 100 000 population, varying from 4·3 (3·6–5·1) per 100 000 population in the USA to 257·1 (221·2–297·7) per 100 000 population in South Sudan (appendix 2 pp 2–19). Globally, in 2019, the incidence rate in children younger than 5 years (hereafter referred to as the under-5 incidence rate) was 192·4 (142·8–258·6) per 100 000 population, down from 312·0 (231·6–418·6) per 100 000 population in 1990, representing a 38·3% (37·0–39·5) decrease (table 2). By SDI quintile, under-5 incidence rates in 2019 ranged from 34·4 (23·5–49·9) per 100 000 population in high SDI quintiles to 400·2 (302·8–529·4) per 100 000 population in low SDI quintiles (table 2). Globally, in 2019, the incidence rate of children aged 0–27 days (neonates) was 854·8 (609·2–1183·3) per 100 000 population, down from 1301·2 per (932·4–1767·2) per 100 000 population in 1990, representing a 34·3% (32·7–35·8) decrease (table 3).

20In 2019, 236 000 deaths (95% UI 204 000–277 000) were attributable to meningitis globally (table 1), of which 112 000 (87 400–145 000) were in children younger than 5 years (table 2). Western sub-Saharan Africa had the highest age-standardised mortality rate (18·4 [15·0–22·1] per 100 000 population), followed by eastern sub-Saharan Africa (13·2 [11·5–15·1] per 100 000 population) and central sub-Saharan Africa (8·9 [6·8–11·2] per 100 000 population; table 1). Overall, the age-standardised mortality rate globally was 3·3 (2·8–3·9) per 100 000 population, down from 7·5 (6·6–8·4) per 100 000 population in 1990, representing a 56·0% (48·3–62·5) decrease. In 2019, rates varied from 0·1 (0·1–0·1) deaths per 100 000 population in Singapore to 26·3 (18·4–38·7) deaths per 100 000 population in Somalia (figure 1; appendix 2 pp 2–19). Worldwide in 2019, the mortality rate in children younger than 5 years (hereafter referred to as the under-5 mortality rate) due to meningitis was 16·9 (13·2–21·9) per 100 000 population, down from 45·0 (37·4–53·6) per 100 000 population in 1990, representing a 62·3% (52·1–70·6) decrease (table 2). Worldwide in 2019, the neonatal mortality rate due to meningitis was 137·2 (109·5–177·8) per 100 000 population, down from 296·7 (257·2–350·9) per 100 000 population in 1990, representing a 53·8% (41·8–62·8) decrease (table 3). More detailed results on meningitis incidence and mortality for all age groups by sex, country, and year are available online via the GBD Results Tool.

21In 2019, viral meningitis comprised 29·9% (95% UI 28·9–30·8) of total all-age meningitis cases, ranging regionally from 27·9% (26·9–29·0) in central sub-Saharan Africa to 37·4% (36·0–38·7) in the high-income Asia Pacific (appendix 2 pp 21–175). N meningitidis comprised 17·3% (16·5–18·0) of total cases, ranging from 8·5% (7·9–9·0) in Australasia to 21·4% (20·4–22·5) in central sub-Saharan Africa. S pneumoniae comprised 13·0% (12·4–13·6) of total cases, ranging from 10·4% (10·0–10·7) in eastern Europe to 14·7% (14·1–15·3) in Tropical Latin America. In the sub-Saharan African super-region, which has the highest incidence, viruses were estimated to be responsible for 28·6% (27·5–29·8) of cases, N meningitidis for 17·0% (15·8–18·1) of cases, and S pneumoniae for 13·6% (12·8–14·4) of cases. In 2019, the highest case burden in children younger than 5 years was attributable to viruses, comprising 25·1% (24·0–26·3) of total cases, followed by N meningitidis (16·9% [15·9–18·0]) and S pneumoniae (13·3% [12·5–14·0]). The highest case burden in neonates was attributable to viruses, comprising 37·1% (34·1–40·3) of total cases, followed by group B Streptococcus (20·4% [17·9–23·3]) and N meningitidis (9·7% [8·3–11·4]).

22The largest proportion of total all-age meningitis deaths was attributable to S pneumoniae, comprising 18·1% (95% UI 17·1–19·2) of total all-age meningitis deaths, followed by N meningitidis (13·6% [12·7–14·4]) and K pneumoniae (12·2% [10·2–14·3]; figure 2; appendix 2 pp 21–175). In children younger than 5 years, the proportions of deaths due to S pneumoniae, N meningitidis, and K pneumoniae were similar to those for adults, but group B Streptococcus was responsible for a much larger fraction of deaths due to its high incidence proportion in neonates (appendix 2 pp 21–175). The largest proportion of deaths due to meningitis in 2019 in children younger than 5 years was attributable to S pneumoniae, comprising 17·3% (16·0–18·6) of total deaths due to meningitis in children younger than 5 years, followed by N meningitidis (12·9% [11·7–13·9]) and K pneumoniae (12·0% [9·7–14·8]). In 2019, the highest proportion of deaths due to meningitis in neonates was attributable to group B Streptococcus (22·8% [19·9–25·9]), followed by K pneumoniae (17·1% [13·6–21·1]) and viruses (15·3% [13·5–17·2]). Overall, 3200 (2420–4230) neonatal meningitis deaths were attributable to group B Streptococcus in 2019, down from 5920 (4870–7290) in 1990. Between 1990 and 2019, the largest reduction in deaths was seen from H influenzae among children younger than 5 years (76·5% [69·5–81·8]), followed by N meningitidis (72·3% [64·4–78·5]) and viruses (58·2% [47·1–67·3]; appendix 2 pp 176–77).

23Between 1990 and 2019, the largest reduction in both mortality and incidence was for meningitis attributable to H influenzae. In 1990, 435 000 (95% UI 340 000–561 000) global all-age cases and 38 000 (31 500–45 400) global all-age deaths were attributable to H influenzae meningitis; of these, 354 000 cases (260 000–475 000) and 32 200 deaths (26 100–39 200) occurred in children younger than 5 years (appendix 2 pp 176–77). In 2019, there were 187 000 (151 000–234 000) all-age cases attributable to H influenzae, corresponding to a reduction of 56·9% (54·1–59·5), and 11 100 (9080–13 500) total deaths attributable to H influenzae, corresponding to a reduction of 70·8% (64·2–76·2; appendix 2 pp 176–77).

24Between 1990 and 2019, the second largest reduction in mortality and incidence was for meningitis attributable to N meningitidis. In 1990, 744 000 (95% UI 608 000–902 000) all-age cases and 80 900 (69 300–95 100) all-age deaths were attributable to N meningitidis meningitis globally; of these, 442 000 (329 000–588 000) cases and 52 300 (42 200–64 300) deaths occurred in children younger than 5 years (appendix 2 pp 176–77). In 2019, there were 433 000 (361 000–518 000) all-age N meningitidis cases, corresponding to a reduction of 41·7% (38·4–44·8), and 32 100 (27 600–38 200) all-age deaths, corresponding to a reduction of 60·2% (52·7–66·4; appendix 2 pp 176–77).

25A high burden of mortality and incidence is concentrated in the African meningitis belt, especially for children younger than 5 years. In 2019, there were 70 600 (95% UI 52 700–93 200) under-5 deaths attributed to meningitis in the meningitis belt, corresponding to 62·9% (57·5–67·8) of under-5 meningitis deaths globally; there were 637 000 (480 000–843 000) under-5 incident cases in the meningitis belt, corresponding to 50·0% (48·1–51·7) of under-5 meningitis cases globally (appendix 2 pp 178–79).

26In 2019, age-standardised mortality rates in the meningitis belt ranged from 1·9 (95% UI 1·3–2·5) per 100 000 population in Sudan to 24·5 (17·5–32·9) per 100 000 population in Niger and 25·4 (19·0–33·3) per 100 000 population in Mali. Incidence rates ranged from 14·0 (11·9–16·4) per 100 000 population in Sudan to 257·1 (221·2–297·7) per 100 000 population in South Sudan (appendix 2 pp 2–19).

27Even in the meningitis belt, which is at high risk of epidemics of meningococcal meningitis, S pneumoniae was responsible for the largest proportion of meningitis-related mortality in 2019, comprising 18·5% (95% UI 17·2–19·8) of total all-age meningitis deaths (121 000 [98 600–149 000]; appendix 2 pp 178, 180). The second largest proportion was attributed to N meningitidis, comprising 12·6% (11·4–13·8) of total all-age meningitis deaths. The all-age mortality rate due to S pneumoniae meningitis dropped in the meningitis belt, from 6·5 (5·3–8·0) per 100 000 population in 1990 to 2·6 (2·1–3·2) per 100 000 population in 2019, representing a 60·0% (50·6–68·0) decrease. The all-age mortality rate due to N meningitidis decreased from 7·4 (6·0–9·0) per 100 000 population in 1990 to 1·7 (1·4–2·2) per 100 000 population in 2019, representing a 76·2% (69·7–81·1) decrease (appendix 2 p 180).

28Among children younger than 5 years, the largest proportion of meningitis-related mortality was attributed to S pneumoniae in the meningitis belt in 2019, comprising 17·7% (95% UI 16·2–19·2) of total under-5 meningitis deaths (70 600 [52 700–93 200]; appendix 2 pp 178, 180). The second largest proportion was attributed to K pneumoniae, comprising 12·0% (9·5–14·9) of total under-5 meningitis deaths. The under-5 meningitis mortality rate due to S pneumoniae dropped from 23·4 (17·9–30·4) per 100 000 population in 1990 to 9·2 (6·7–12·5) per 100 000 population in 2019, representing a 60·5% (47·6–70·8) decrease. Mortality rates due to K pneumoniae among children younger than 5 years decreased from 14·4 (10·0–20·2) per 100 000 population in 1990 to 6·2 (4·2–8·7) per 100 000 population in 2019, representing a 56·6% (42·4–67·7) decrease (appendix 2 p 180).

29This study presents what are to our knowledge the most comprehensive estimates of the meningitis burden attributable to a comprehensive set of pathogens by age group across countries, contributing to our understanding of the pathogen-specific burden of meningitis. In 2019, the highest proportion of meningitis deaths both in all age groups and in children younger than 5 years was attributable to S pneumoniae, followed by N meningitidis and K pneumoniae, and the highest proportion of deaths in neonates was attributable to group B Streptococcus. S pneumoniae, N meningitidis, and H influenzae are the three pathogens responsible for the most meningitis-related deaths among children younger than 5 years, and there have been large reductions in the number of deaths attributable to all three pathogens since 1990.

30A reduction in the burden of these pathogens could be attributed to successful vaccination rollouts over the past 30 years, with the highly effective Hib, pneumococcal, and meningococcal conjugate vaccines playing a key role. To achieve the goals of the WHO global roadmap to reduce cases of vaccine-preventable bacterial meningitis by 50% and deaths by 70% by 2030, continued vaccination against these pathogens is essential. Progress towards the roadmap goal of eliminating meningitis epidemics, which most commonly occur in the meningitis belt, is largely attributable to the rollout of vaccination programmes: specifically, the highly successful MenAfriVac campaign. An analysis of enhanced surveillance data from nine countries in the meningitis belt that completed MenAfriVac campaigns from 2005 to 2015 estimated that confirmed cases of N meningitidis serogroup A had declined by more than 99%. Our modelled results for meningococcus, which include serogroup A and other serogroups not covered by the vaccine, estimate a 41·7% decline in cases. A key factor in the successful rollout of MenAfriVac is its much lower cost than other comparable conjugate vaccines, improving the feasibility of mass rollout and sustainable vaccination.,,

31Although MenAfriVac campaigns successfully reduced the incidence rate of N meningitidis serogroup A in the meningitis belt, both sporadic and epidemic cases due to serogroups W, X, and C have increased., Niger, in particular, began reporting an increase in both the number and proportion of serogroup W cases in the same year as the introduction of MenAfriVac. This observation underscores the need for an affordable multivalent meningococcal vaccine. Serotype replacement is also an issue for S pneumoniae: serotypes not covered by the 13-valent pneumococcal conjugate vaccine (PCV13), such as 12F and 23B, have resulted in a relative increase in the incidence of invasive disease in the post-vaccine era. The fact that meningitis is multipathogenic and that only some of the pathogens are currently vaccine preventable might have contributed to a slower decline in under-5 mortality rates for meningitis between 1990 and 2019 (62·3% [95% UI 52·1–70·6]) compared with diseases that are caused by a vaccine-preventable single pathogen, such as measles (90·5% [87·6–93·0]) and tetanus (91·9% [86·3–94·2]).

32For the first time, we have quantified the global burden of meningitis attributable to K pneumoniae and have identified this pathogen as the third leading cause of meningitis mortality after S pneumoniae and N meningitidis. Unlike S pneumoniae and N meningitidis, which are the most common causes of community-acquired meningitis, K pneumoniae meningitis is usually hospital-acquired and has higher mortality rates, especially in patients older than 60 years with comorbidities.,,, A systematic review found that K pneumoniae is the leading cause of Gram-negative meningitis and bacteraemia in low-income and middle-income countries. Despite its high fatal burden, K pneumoniae is not included as a priority pathogen in WHO's meningitis global roadmap, possibly because of a scarcity of global data on meningitis due to K pneumonaie before this study. K pneumoniae is also not included in most commercially available PCR or rapid diagnostic tests. Developing rapid diagnostic tests to detect this pathogen would improve tracking of this disease and its high burden.

33The emerging shift in the cause of bacterial meningitis mortality to K pneumoniae emphasises the need for better control of hospital-acquired infections and antimicrobial stewardship. Many of these K pneumoniae strains have been reported to possess broad and threatening antimicrobial resistance genes, including extended-spectrum beta-lactamase with carbapenemase genes,, and these carbapenem-resistant strains are classified as critical on the WHO global priority antimicrobial resistance pathogen list. Many vaccines for K pneumoniae are in development and in clinical trials, with one of the most promising being multicomponent conjugate vaccines. These new vaccines might be able to protect against the multitude of infectious syndromes and antimicrobial resistance that K pneumoniae is known to produce. The rollout of Hib and S pneumoniae conjugate vaccines has also been shown to reduce the need for antibiotic use and helped slow the development of antimicrobial resistance.

34Group B Streptococcus comprises the highest proportion of neonatal meningitis deaths, although the absolute number of meningitis deaths due to group B Streptococcus in neonates has decreased from 5920 (95% UI 4870–7290) in 1990 to 3200 (2420–4230) in 2019. Unlike the key pathogens responsible for post-neonatal childhood meningitis, group B Streptococcus is not vaccine preventable. Instead, prevention of neonatal group B Streptococcus currently relies on prenatal testing of the mother and intrapartum antibiotic prophylaxis. A global meta-analysis found that the risk of early-onset group B Streptococcus in neonates was 1·1% in settings without an intrapartum antibiotic prophylaxis policy, and 0·3% in settings with an intrapartum antibiotic prophylaxis policy. An analysis of global intrapartum antibiotic prophylaxis policies in 90 countries found that 40 of 44 high-income countries had a policy, while only three of 20 countries in sub-Saharan Africa and one of three in east Asia reported having such a policy. Intrapartum antibiotic prophylaxis does not protect against late-onset group B Streptococcus disease occurring in infants aged 7–89 days,, which is responsible for about a third of the total neonatal group B Streptococcus cases. A maternal vaccine against group B Streptococcus might therefore be a solution, protecting against both early-onset and late-onset group B Streptococcus, and preventing frequent antibiotic use that can be associated with the development of antimicrobial resistance. Such a vaccine is currently a WHO priority and is predicted to be a cost-effective, financially sustainable, and feasible intervention. In addition to vaccination, reducing rates of short gestation and low birthweight might play an important role in preventing early-onset group B Streptococcus.

35Improvements to laboratory systems for both patient treatment and population surveillance are other key pillars in the WHO meningitis roadmap., Specific microbial diagnosis guides appropriate treatment, including antibiotic selection. In the meningitis belt, population-level microbial surveillance enables the timely rollout of reactive vaccination campaigns against meningococcal strains other than serogroup A.,, One such strategy to improve laboratory systems is to invest in infrastructure that reduces the transportation time of CSF samples and to reinforce safe-handling procedures between collection facilities and laboratories., Another strategy is the development and rollout of next-generation rapid diagnostic tests (RDTs). Their low cost and ease of operation accelerate the diagnosis of meningitis in low-resource settings, substantially improving patient care, surveillance, and outbreak response; development of next-generation RDTs is therefore a WHO priority. However, downsides of existing RDTs include the absence of many pathogens in the test, the inability to distinguish between serotypes, and the inability to assess for antimicrobial resistance., As the prevalence of antimicrobial resistance rises, especially in non-vaccine-preventable pathogens such as K pneumoniae, the need for global antimicrobial resistance surveillance systems grows. The WHO Global Antimicrobial Resistance and Use Surveillance System (GLASS), which provides a standardised approach to collection, analysis, and sharing of clinical antimicrobial resistance data for surveillance, is expanding. In 2020, 251 surveillance sites in the WHO African region, where meningitis poses the highest burden, submitted data to GLASS, up from 35 surveillance sites in 2017. The concurrent expansion of next-generation RDTs for point-of-care decision making, and enhanced microbial surveillance systems, including antimicrobial resistance surveillance to support evidence-based antibiotic use, can help improve outcomes for patients with infectious meningitis globally.

36Our estimates of meningitis mortality in children younger than 5 years differ from those produced by other groups, both for the overall burden and for the pathogen-specific burden. The 2017 WHO and Maternal and Child Epidemiology Estimation Group (WHO-MCEE) estimate for global under-5 deaths due to meningitis and encephalitis was 142 841, which is lower than, but within the uncertainty interval of, the GBD 2019 estimate published previously for meningitis and encephalitis combined in the same year (143 000 [95% UI 115 000–179 000]; WHO-MCEE combines meningitis and encephalitis, whereas GBD models them as mutually exclusive causes). A 2018 global systematic review by Oordt-Speets and colleagues of 56 published studies found that S pneumoniae and N meningitidis were the most common all-age bacterial meningitis pathogens in all regions, with mean case proportions ranging from 25·1% to 41·2% for S pneumoniae and from 9·1% to 36·2% N meningitidis, across regions. Here, we estimate that N meningitidis comprised 17·3% (16·5–18·0) of global all-age meningitis cases in 2019, ranging from 8·5% in Australasia to 21·4% in central sub-Saharan Africa, while S pneumoniae comprised 13·0% (12·4–13·6) of global all-age meningitis cases in the same year, ranging from 10·4% in eastern Europe to 14·7% in Tropical Latin America (appendix 2 pp 21–175). Additionally, the systematic review by Oordt-Speets and colleagues estimated that, in the African region, which had the most all-age data sources, S pneumoniae was responsible for 41·2% (34·1–48·4) of bacterial meningitis cases and N meningitidis was responsible for 36·2% (26·6–46·4) of bacterial meningitis cases. In our study, in the sub-Saharan African region, S pneumoniae was estimated to be responsible for 13·6% (12·8–14·4) of cases and N meningitidis was estimated to be responsible for 17·0% (15·8–18·1) of cases. Part of the reason for this discrepancy is that the previous systematic review only included 56 articles published between April 25, 2012, and April 25, 2017, whereas we include scientific literature published between 1990 and 2019, in addition to non-literature sources, such as multiple cause of death data and laboratory data. Owing to the historical importance and vaccine-preventable nature of S pneumoniae, N meningitidis, and H influenzae, many literature studies look specifically for these pathogens through rapid tests and assays that are not used for other pathogens, thus leading to a potential over-representation of these pathogens in literature.

37This study has some limitations. First, the estimates of the overall meningitis and aetiology-specific burden are limited by data availability and quality, especially in areas of the meningitis belt where the burden is the greatest in settings with low health-care access. Even when data are available, heterogeneous data sources (eg, surveillance versus inpatient data) might not be directly comparable. We applied a standardised approach to adjust for systematic bias among different data sources before modelling. Locations with sparse or no data must rely on covariates and regional trends to predict estimates. Additionally, many of these locations rely on verbal autopsy data. A multisite validation study by Lozano and colleagues shows that verbal autopsy of meningitis has only modest performance. Strong, population-based surveillance systems are preferred and needed to drive response, inform case management, assess vaccine impact, and track progress. Second, we directly apply meningitis aetiology proportions from the global burden of antimicrobial resistance study to GBD estimates of meningitis cases and deaths, even though the two studies have slightly different definitions of meningitis. More specifically, for fatal cases, the GBD definition of meningitis includes only instances in which meningitis was the underlying cause of death, whereas the antimicrobial resistance study definition includes any instance where meningitis was present in the causal chain, regardless of the underlying cause of death. Third, meningitis cases due to Mycobacterium tuberculosis or HIV-associated opportunistic infections, including Cryptococcus neoformans, were not included in the present study, as GBD classifies them with the underlying diseases tuberculosis and HIV, respectively. Fourth, due to the poor specificity of data documenting cases of viral meningitis, we modelled all viral aetiologies collectively, rather than distinguishing individual viruses of scientific interest, such as enteroviruses and herpes viruses. Fifth, we did not explicitly search for data sources that used molecular methods such as genome sequencing to identify viral pathogens, but such sources could be emphasised in a future systematic review. Sixth, we assumed the distribution of cases of meningitis with unknown aetiology (ie, those not identified through physician diagnosis or by microbiological means as either viral meningitis or a specified bacteria) was the same as the distribution of meningitis cases in which the aetiology was defined. This assumption could be violated if certain pathogens are more difficult to detect than others, or in cases where a pathogen is irregularly tested for within a laboratory. Seventh, age-standardised estimates allow for comparisons between populations with potentially different age distributions, but they should not be interpreted as actual rates. Meningitis incidence and mortality estimates that are not age-standardised are available online via the GBD Results Tool described above. Eighth, we did not assess risk factors for meningitis, but these could be the subject of future research. Finally, presenting K pneumoniae results for all children younger than 5 years might mask its higher burden in children younger than 1 year. We plan to report results for more detailed age groups in future GBD rounds.

38Our study presents the burden of meningitis and its aetiologies before the COVID-19 pandemic. Evidence suggests that social distancing associated with COVID-19 has resulted in a lower incidence of meningitis and other invasive infections attributable to selected pathogens transmitted via the respiratory route, including S pneumoniae, N meningitidis, and H influenzae, during the pandemic, while the incidence of group B Streptococcus remained unchanged., However, data for other pathogens remain unavailable. At the same time, the COVID-19 pandemic disrupted delivery of vaccines in 2020. Although this disruption is unlikely to affect meningitis rates due to persistent herd immunity, growing vaccine hesitancy fuelled by the pandemic will pose a challenge to preventing and controlling meningitis in the years ahead. As data become available for more countries and more pathogens, we will be able to comprehensively quantify the indirect effects of the pandemic on the incidence of meningitis and its aetiologies in future rounds of GBD.

39Although meningitis incidence and mortality rates have decreased globally since 1990, progress lags behind that for other vaccine-preventable diseases. Moreover, although increased vaccine coverage in low-income and middle-income countries might have driven reductions in the meningitis burden, the reduction is not equal across locations. There is a continued need for low-cost multivalent vaccines as a preventive measure and for epidemic control in the meningitis belt. Further strengthening laboratory capacity to diagnose meningitis accurately and rapidly will also assist in the control of epidemics. Countries outside the African meningitis belt with high meningitis burdens, such as those in south Asia, are also affected by these policy needs. Additional enhanced surveillance data will improve country-specific burden estimates, which will help track progress towards reducing the global burden of meningitis by 2030.

40To download the data used in these analyses, please visit the Global Health Data Exchange GBD 2019 website.

41K Akinosoglou reports payment or honoraria for lectures, presentations, speakers’ bureau fees, manuscript writing or educational events from Pfizer Hellas, Gilead Sciences, Merck Sharp and Dohme, Glaxosmithkline Greece, as payments to the University of Patras; and support for attending meetings or travel, or both, to Pfizer Hellas, Gilead Sciences, Merck Sharp and Dohme, Glaxosmithkline Greece, and Norma Hellas; all outside the submitted work. R Ancuceanu reports payment or honoraria for lectures, presentations, speakers’ bureau fees, manuscript writing or educational events from AbbVie, Sandoz, B. Braun, and Laropharm, all outside the submitted work. S Bhaskar reports leadership or fiduciary roles in other board, society, committee or advocacy groups, paid or unpaid, with Rotary Club of Sydney, Australia as Board Director, with Rotary District 9675, Australia as Chair of Diversity Equity & Inclusion, and with Global Health Hub, Berlin as Founding Member/Chair & Co-Manager, Global Health and Migration Hub Community, all outside the submitted work. D Buonsenso reports payment or honoraria for lectures, presentations, speakers bureau fees, manuscript writing or educational events and participation on an advisory board from Pfizer for the Pneumococcal 22 Vaccine Advisory Board 2022, outside the submitted work. A Demetriades reports payment or honoraria for speakers’ bureau fees from Integra, Stryker, and Safe Orthopaedics; leadership or fiduciary roles, unpaid, as a board member (non-stipendiary) with the European Association of Neurosurgical Societies, Global Neuro Foundation and on the steering committee (non-stipendiary) of AO Spine Knowledge Forum Degenerative; all outside the submitted work. B D Gessner reports support for the present manuscript from Pfizer through salary payments; and stock or stock options through their employment with Pfizer; all outside the submitted work. N E Ismail reports a leadership or fiduciary role as an unpaid council member and bursar of the Malaysian Academy of Pharmacy, outside the submitted work. J J Jozwiak reports payment or honoraria for lectures, presentations, speakers’ bureau fees, manuscript writing or educational events from Novartis and Adamed as personal payments, outside the submitted work. K Krishan reports non-financial support from the UGC Centre of Advanced Study, CAS II, Department of Anthropology, Panjab University, Chandigarh, India, outside the submitted work. A-F A Mentis reports grants or contracts from “MilkSafe: A novel pipeline to enrich formula milk using omics technologies”, a research co financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH-CREATE - -INNOVATE (project code: T2EDK-02222), as well as from ELIDEK (Hellenic Foundation for Research and Innovation, MIMS-860); payment for expert testimony as a peer-reviewer for Fondazione Cariplo, Italy; leadership or fiduciary roles in other boards, societies, committees or advocacy groups, paid or unpaid, by serving as Editorial Board Member for the journals Systematic Reviews and Annals of Epidemiology, and as Associate Editor for Translational Psychiatry; stocks in a family winery; and other financial or non-financial interests as a scientific officer with the BGI Group; all outside the submitted work. M J Postma reports stock or stock options from Health-Ecore (Zeist NL, 25%) and PAG BV (Groningen, NL, 100%) outside the submitted work. S Shrestha reports other financial interests in the School of Pharmacy at Monash University Malaysia by receiving the Graduate Research Merit Scholarship to pursue his PhD, outside the submitted work. L Silva reports grants or contracts from CENTRO-04-3559-FSE-000162, Fundo Social Europeu, outside the submitted work. J A Singh reports consulting fees from Crealta/Horizon, Medisys, Fidia, PK Med, Two Labs, Adept Field Solutions, Clinical Care Options, Clearview Healthcare Partners, Putnam Associates, Focus Forward, Navigant Consulting, Spherix, MedIQ, Jupiter Life Science, UBM, Trio Health, Medscape, WebMD, and Practice Point Communications, the National Institutes of Health, and the American College of Rheumatology; payment or honoraria for speakers’ bureau fees from Simply Speaking; support for attending meetings or travel from the steering committee of OMERACT; participation on a Data Safety Monitoring Board or Advisory Board with the US Food and Drug Administration Arthritis Advisory Committee; a leadership or fiduciary role in board, society, committee or advocacy group, paid or unpaid, with OMERACT as a steering committee member, with the Veterans Affairs Rheumatology Field Advisory Committee as Chair (unpaid), and with the UAB Cochrane Musculoskeletal Group Satellite Center on Network Meta-analysis and as editor and director (unpaid); stock or stock options in TPT Global Tech, Vaxart Pharmaceuticals, Aytu BioPharma, Adaptimmune Therapeutics, GeoVax Labs, Pieris Pharmaceuticals, Enzolytics, Seres Therapeutics, Tonix Pharmaceuticals, and Charlotte's Web Holdings, and previously owned stock options in Amarin, Viking, and Moderna Pharmaceuticals; all outside the submitted work. C Wright is an employee of Meningitis Research Foundation, which receives grants in support of their charitable objectives from GlaxoSmithKline, Pfizer, Sanofi Pasteur, Serum Institute, and the Tableau Foundation, outside the submitted work.