The epidemiology of bacterial meningitis in the United States during 2008–2023: an analysis of active, laboratory, population-based, multistate surveillance data

Summary

Background

Bacterial meningitis is a severe syndrome with dynamic epidemiology, but assessments of current trends are limited. We aimed to describe changing epidemiologic patterns among common bacterial causes of meningitis in the United States.

Methods

We analyzed data on bacterial meningitis cases caused by Streptococcus pneumoniae, group B Streptococcus (GBS), Haemophilus influenzae, Neisseria meningitidis, and Listeria monocytogenes in 10 U.S. surveillance sites. We compared incidence (cases per 100,000) across four epidemiologic periods: 2008–2009, 2010–2019, 2020–2021, and 2022–2023.

Findings

We identified 5,032 bacterial meningitis cases; among those with outcome data, 11% (573/5028) died. S. pneumoniae was the dominant pathogen (59% [2922/5032]) throughout. However, GBS predominated among infants aged 0–2 months (85% [660/775]), the age group with the highest incidence. Between 2008–2009 and 2010–2019, overall bacterial meningitis incidence declined from 1.3 to 1.1, driven by decreases in S. pneumoniae meningitis caused by serotypes contained in the 13-valent pneumococcal conjugate vaccine (PCV13) and N. meningitidis meningitis. Meningitis caused by non-b H. influenzae strains increased during this period. During 2020–2021, incidence declined to 0.7, driven by decreases in S. pneumoniae, H. influenzae, and N. meningitidis meningitis, regardless of organism subtype. During 2022–2023, incidence increased to 1.0, driven by increases in S. pneumoniae and H. influenzae meningitis. Case fatality ratios remained stable throughout.

Interpretation

Bacterial meningitis incidence rates have declined since 2008, with a notable low during 2020–2021, followed by a resurgence during 2022–2023. Case fatality remains high. Strategies that provide effective and broader pneumococcal and H. influenzae serotype protection and prevent infant GBS meningitis could reduce residual meningitis burden.

Funding

U.S. Centers for Disease Control and Prevention.

Research in context.

Evidence before this study

We searched PubMed using the search term “bacterial meningitis” together with “epidemiology”, “incidence”, “trends”, or “rates.” Our search was limited to publications since 2008. We searched for population-based surveillance studies in the United States with data on common causes of bacterial meningitis. The most recent publication was from 2014 and used International Classification of Diseases coding data to estimate the incidence of bacterial meningitis hospitalizations during 1997–2010, due to commonly identified bacterial pathogens: S. pneumoniae, N. meningitidis, H. influenzae, staphylococcus species, and other Gram-negative bacteria. The study reported S. pneumoniae as the leading cause of bacterial meningitis but with a significant decrease in incidence and mortality that was associated with the introduction of a seven-valent pneumococcal conjugate vaccine (PCV7) in 2000 and through use of adjunctive dexamethasone. The study also reported a decrease in N. meningitidis incidence, which was attributed to secular trends and the introduction of the quadrivalent meningococcal conjugate vaccine in 2005.

Added value of this study

This study provides comprehensive estimates of meningitis incidence caused by S. pneumoniae, group B. Streptococcus (GBS), Haemophilus influenzae, Neisseria meningitidis, and Listeria monocytogenes in ten U.S. sites during 2008–2023. Our study showed that the incidence of bacterial meningitis due to these pathogens has declined since 2008. Between 2008–2009 and 2010–2019, the decline in incidence was driven by decreases in S. pneumoniae meningitis caused by serotypes contained in PCV13 and was attributed to PCV13 introduction in 2010. N. meningitidis meningitis also decreased during this period while H. influenzae meningitis caused by non-b H. influenzae strains increased. Consequently, there was a change in the relative incidence of causative pathogens, with H. influenzae accounting for a greater proportion of meningitis cases than N. meningitidis. Between 2010–2019 and 2020–2021, at the onset of the COVID-19 pandemic, the decline in incidence was driven by dramatic decreases in meningitis caused by respiratory pathogens; S. pneumoniae, H. influenzae, and N. meningitidis and was seen across all organism subtypes. During 2022–2023, there was a resurgence in bacterial meningitis incidence, driven by increases in S. pneumoniae and H. influenzae meningitis. By 2023, the incidence of bacterial meningitis due to our pathogens of interest had returned to pre-COVID-19 pandemic levels. There was no significant change in the incidence of meningitis caused by GBS or L. monocytogenes throughout the surveillance period. Our study demonstrates the dynamic epidemiology of bacterial meningitis due to common pathogens over the last two decades. Overall, S. pneumoniae remained the leading cause of bacterial meningitis. However, the incidence of bacterial meningitis was highest among young infants, and in this age group GBS was the most common cause. Finally, there was little change in the case fatality of bacterial meningitis throughout the surveillance period.

Implications of all the available evidence

Current descriptions of bacterial meningitis epidemiology are crucial for informing preventative strategies and clinical guidelines. The persistence of S. pneumoniae as a leading cause of meningitis highlights the potential utility of vaccines providing effective and broader pneumococcal serotype protection. Meningitis due to non-b H. influenzae strains increased during the surveillance period. Consequently, H. influenzae now accounts for more meningitis cases than N meningitidis, which is an important consideration for vaccine development and antimicrobial therapy guidelines. Finally, preventative strategies that prioritize GBS meningitis among infants could be beneficial, given the high and unchanged burden of disease due to this pathogen in this age group.

Introduction

Bacterial meningitis is associated with substantial morbidity and mortality.¹ Important community-acquired causes of bacterial meningitis include Streptococcus pneumoniae, group B Streptococcus (GBS), Haemophilus influenzae, Neisseria meningitidis, and Listeria monocytogenes. In the United States, between 1998 and 2007, the incidence of meningitis caused by these five pathogens declined by 31% from 2.0 to 1.4 per 100,000.² This decline was driven by a decrease in S. pneumoniae meningitis and was attributed to the use of the seven-valent pneumococcal conjugate (PCV7) vaccine, which was introduced in 2000. Since then, several factors may have further influenced the epidemiology of bacterial meningitis, including the replacement of PCV7 with a 13-valent product (PCV13) in 2010, updated pneumococcal and meningococcal vaccine recommendations (Appendix 1), as well as the COVID-19 pandemic. Moreover, new preventative strategies including maternal GBS vaccines,³ higher-valent pneumococcal vaccines,⁴ and vaccines targeting non-typeable H. influenzae⁵ and select non-type b Haemophilus influenzae serotypes⁶ are now in development. Current descriptions of bacterial meningitis epidemiology are necessary for the evaluation and prioritization of meningitis-related interventions.

We used data from two active, laboratory-and population-based surveillance systems within the Emerging Infections Program (EIP) Network—the Active Bacterial Core surveillance (ABCs) and the Foodborne Diseases Active Surveillance Network (FoodNet)—to describe bacterial meningitis epidemiologic trends in the United States during 2008–2023.

Methods

Surveillance

We identified meningitis cases caused by S. pneumoniae, GBS, H. influenzae, and N. meningitidis through ABCs and meningitis cases caused by L. monocytogenes through FoodNet. A full description of study methods is available in Appendix 2 and surveillance methods for ABCs and FoodNet have been previously described.^7,8 Briefly, in ABCs a case of meningitis is defined as the presence of S. pneumoniae, GBS, H. influenzae, or N. meningitidis in cerebrospinal fluid (CSF) by culture or nucleic acid amplification test (NAAT) or from another normally sterile site in association with a clinical diagnosis of meningitis. However, no meningitis cases were positive on NAAT without an accompanying positive culture. Since FoodNet does not collect information on clinical meningitis, only cases in which L. monocytogenes was isolated from CSF by culture were defined as meningitis cases. We included cases identified between January 1, 2008 and December 31, 2023 from select counties in California, Colorado, Georgia, Maryland, New York, Oregon, and Tennessee and from the entire state of Connecticut, Minnesota, and New Mexico. These sites encompassed an estimated 34,490,3316 persons (10.5% of the total U.S. population), according to the 2023 census. ABCs sites sent available isolates to reference laboratories for organism-specific subtyping. Pneumococcal and GBS isolates also had antimicrobial susceptibility testing data available for analysis.

Study design and statistical analysis

We expected that PCV13 introduction in 2010 and the COVID-19 pandemic during 2020–2021 would have the largest potential impact on overall bacterial meningitis etiology and epidemiologic trends. Thus, in addition to calculating annual incidence, we described epidemiologic changes within the context of these events by dividing our study duration into four analytic periods (2008–2009, 2010–2019, 2020–2021, and 2022–2023).

Additionally, we described demographics and comorbidities among meningitis case patients during 2010–2023 (Appendix 2). This descriptive analysis was limited to ABCs data during this period because comorbidity data is not collected in FoodNet and was only uniformly collected in ABCs from 2010 onwards.

We calculated annual pathogen-specific meningitis incidence, expressed as the number of cases per 100,000 population, using U.S. Census annual population estimates. We stratified S. pneumoniae, H. influenzae, and GBS meningitis incidence by relevant age and organism subtype group. The low number of meningitis cases due to N. meningitidis and L. monocytogenes precluded similar incidence stratifications. To evaluate the impact of PCV13 introduction on S. pneumoniae meningitis trends, as well as to inform the use of recently recommended 15-and-20-valent pneumococcal conjugate vaccines (PCV15 and PCV20),9, 10, 11 we stratified S. pneumoniae meningitis incidence by four categories: PCV13-serotypes (1, 3, 4, 5, 6A, 6B, 6C, 7F, 9V, 14, 18C, 19A, 19F, 23F), PCV15-unique-serotypes (22F, 33F), PCV20-unique-serotypes (8, 10A, 11A, 12F, 15B), and non-vaccine serotypes (NVT, all serotypes not in PCV13, PCV15, or PCV20). While serotype 6C is not included in PCV13, it was included within the PCV13-serotypes category due to reported cross protection from serotype 6A.¹² Likewise, to better understand changes in H. influenzae trends and inform future vaccine use, we stratified H. influenzae meningitis incidence by three categories: vaccine-preventable H. influenzae type b (Hib), nontypeable H. influenzae (NTHi), and serotypes other than Hib and NTHi (non-b Hi serotypes).

Missing race/ethnicity and pathogen subtype data were accounted for in incidence calculations using a single imputation process by assigning values to cases with missing data based on period, age group, and state-specific distribution of cases with known data (Appendix 3). Percent change in average annual incidence and corresponding 95% CIs were estimated using Poisson regression. Clustering by state was not accounted for in the main Poisson regression model, however, was considered in a sensitivity analysis by using a mixed effects Poisson regression model including random effects for state (Appendix 2). Among pathogens with available subtyping and antimicrobial susceptibility data, we assessed changes in proportions across relevant periods. We calculated case fatality ratios and changes over periods, overall and by pathogen, using data from patients with a known outcome (99.9% of patients). Comparisons of proportions across periods were done using two-sided chi-squared tests, assuming statistical significance at p < 0.05.

To estimate the national burden of meningitis due to the five pathogens of interest in the United States, we applied age-and-race-specific observed incidence of meningitis, calculated from the ABCs and FoodNet data, to the total U.S. population. Analyses were performed using R statistical programming language version 4.0.4. This activity was reviewed by the Centers for Disease Control and Prevention (CDC) and was conducted consistently with applicable federal law and CDC policy.

Ethics approval

The CDC determined that this surveillance project was not human subjects’ research; therefore CDC Institutional Review Board approval was not required.

Role of the funding source

This work was supported by the CDC's Emerging Infections Programs. CDC was involved in the study design, data analysis, data interpretation, and writing and submission of the report.

Results

Overall meningitis trends

During 2008–2023, we identified 5,032 cases of meningitis caused by S. pneumoniae, GBS, H. influenzae, N. meningitidis, and L. monocytogenes. Between 2008–2009 and 2010–2019, following PCV13 introduction, the average annual incidence of meningitis declined by 16.6% (95% CI: −23.5, −9.0) from 1.3 to 1.1 per 100,000. Between 2010–2019 and 2020–2021, at the onset of COVID-19 pandemic, a further 41.4% (95% CI: −46.9, −35.4) decline in incidence was observed to 0.7 per 100,000. During 2022–2023, a resurgence in meningitis was observed, with incidence increasing by 57.7% (95% CI: 40.3, 77.4) to 1.0 per 100,000 (Table 1). Across all periods, incidence remained highest for infants aged 0–2 months, among whom GBS was the predominant (660/775 [85.1%]) cause. In race/ethnicity stratified analysis, incidence was highest among American Indian or Alaska Native and Black, non-Hispanic people across all periods. Of 5,028 cases with outcome data, 573 (11.4%) died. There was little change in case fatality ratios across periods (Table 2). In a sensitivity analysis investigating the impact of clustering by state, we observed little change in incidence rates or change in average annual incidence, overall or when stratified by pathogen (Appendix 2).

Table 1.

Incidence of bacterial meningitis in Emerging Infections Programs Network sites, 2008–2023, stratified by pathogen, age group, and race/ethnicity.^a

	2008–2009		2010–2019		2020–2021		2022–2023		Percentage change in average annual incidence, 2010–2019 vs 2008–2009 (95% CI)	Percentage change in average annual incidence, 2020–2021 vs 2010–2019 (95% CI)	Percentage change in average annual incidence in 2022–2023 vs 2020–2021 (95% CI)
	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)
Overall incidence	610	1.33 (1.23–1.44)	3247	1.11 (1.07–1.15)	454	0.65 (0.6–0.72)	721	1.03 (0.96–1.11)	−16.6 (−23.5, −9.0)	−41.4 (−46.9, −35.4)	57.7 (40.3, 77.4)
Pathogen
Streptococcus pneumoniae	356	0.78 (0.7–0.86)	1900	0.65 (0.62–0.68)	231	0.33 (0.29–0.38)	435	0.62 (0.56–0.68)	−16.4 (−25.2, −6.2)	−49.0 (−55.6, −41.7)	87.0 (59.6, 119.6)
Group B Streptococcus	103	0.23 (0.19–0.27)	685	0.23 (0.22–0.25)	138	0.2 (0.17–0.23)	139	0.2 (0.17–0.23)	4.1 (−14.9, 28.8)	−15.5 (−29.9, 1.1)	0.0 (−21.0, 26.6)
Haemophilus influenzae	39	0.09 (0.06–0.12)	378	0.13 (0.12–0.14)	47	0.07 (0.05–0.09)	96	0.14 (0.11–0.17)	52.3 (11.1, 115.2)	−47.8 (−62.0, −30.2)	102.8 (43.9, 189.7)
Neisseria meningitidis	89	0.19 (0.16–0.24)	191	0.07 (0.06–0.08)	21	0.03 (0.02–0.05)	28	0.04 (0.03–0.06)	−66.5 (−73.9, −56.8)	−53.9 (−71.5, −29.4)	32.4 (−24.5, 135.8)
Listeria monocytogenes	23	0.05 (0.03–0.08)	93	0.03 (0.03–0.04)	17	0.02 (0.02–0.04)	23	0.03 (0.02–0.05)	−36.6 (−59.1, 2.4)	−23.3 (−55.8, 25.1)	34.3 (−27.9, 155.2)
Age group
0–2 months	84	52.2 (42.15–64.64)	518	58.59 (53.76–63.86)	87	46.27 (37.5–57.09)	86	45.72 (37.01–56.48)	12.5 (−10.1, 42.6)	−21 (−37.5, −1.5)	−1.2 (−26.7, 33.2)
3–11 months	55	11.39 (8.75–14.84)	235	8.86 (7.8–10.07)	27	4.79 (3.28–6.98)	34	6.02 (4.3–8.43)	−22.2 (−41.5, 5.3)	−46 (−64.5, −21.1)	25.9 (−23.9, 110.2)
1–4 years	30	1.18 (0.83–1.69)	161	1.1 (0.95–1.29)	25	0.79 (0.53–1.17)	38	1.24 (0.91–1.71)	−7 (−36.1, 39.9)	−28.4 (−54.1, 7.0)	57.4 (−4.4, 163.8)
5–17 years	32	0.4 (0.29–0.57)	175	0.36 (0.31–0.41)	13	0.12 (0.07–0.20)	58	0.52 (0.4–0.67)	−11.5 (−38.4, 31.2)	−67.7 (−82.5, −45.6)	347.8 (153.7, 753.7)
18–49 years	163	0.78 (0.67–0.91)	716	0.56 (0.52–0.6)	87	0.29 (0.24–0.36)	169	0.56 (0.48–0.65)	−28.4 (−39.4, −14.8)	−48 (−58.7, −35.5)	92.8 (49.4, 150.7)
50–64 years	145	1.72 (1.46–2.02)	812	1.42 (1.33–1.53)	116	0.86 (0.72–1.04)	176	1.32 (1.14–1.53)	−17 (−30.2, −0.6)	−39.4 (−50.3, −26.7)	53.1 (21.3, 93.9)
≥65 years	101	1.92 (1.58–2.33)	630	1.58 (1.46–1.7)	99	0.89 (0.73–1.09)	160	1.38 (1.18–1.61)	−17.7 (−33.0, 2.0)	−43.3 (−54.4, −30.3)	54 (20.1, 98.4)
Race/ethnicity
White, non-Hispanic	340	1.12 (1.01–1.25)	1745	0.96 (0.91–1.0)	239	0.58 (0.51–0.66)	375	0.93 (0.84–1.02)	−14.7 (−24.0, −4.1)	−39.1 (−46.9, −30.4)	58.7 (35.0, 86.8)
Black, non-Hispanic	155	2.04 (1.75–2.39)	820	1.79 (1.67–1.92)	122	1.12 (0.94–1.34)	173	1.56 (1.35–1.81)	−12.5 (−26.1, 4.2)	−37.2 (−48.3, −24.4)	39 (10.4, 75.5)
AI/AN, non-Hispanic	15	2.69 (1.62–4.46)	91	3.03 (2.47–3.72)	8	1.25 (0.63–2.5)	8	1.24 (0.62–2.48)	12.8 (−32.6, 102.7)	−58.7 (−81.6, −20.2)	−0.8 (−63.5, 169.6)
Asian, NH, PI, non- Hispanic	9	0.44 (0.23–0.84)	95	0.54 (0.44–0.66)	15	0.28 (0.17–0.47)	27	0.49 (0.34–0.71)	24.6 (−33.4, 165.6)	−48.1 (−71.1, −13.4)	73.8 (−6.3, 234.8)
Multiracial, non-Hispanic	2	0.3 (0.08–1.20)	20	0.32 (0.21–0.5)	8	0.47 (0.23–0.93)	10	0.56 (0.3–1.04)	7.4 (−68.6, 572.8)	43.9 (−40.4, 214.9)	20.1 (−52.6, 214.6)
Hispanic	84	1.82 (1.47–2.26)	457	1.23 (1.12–1.34)	53	0.53 (0.4–0.69)	120	1.14 (0.95–1.36)	−32.7 (−46.4, −14.6)	−57.1 (−68.1, −43.6)	116.2 (57.4, 200.9)

Abbreviations: AI/AN; American Indian or Alaska Native, NH; Native Hawaiian, PI; Pacific Islander.

Incidence is reported as average annual incidence per 100,000 during 2008–2009, 2010–2019, 2020–2021, and 2022–2023. Percent changes in incidence with 95% CI were estimated using Poisson regression.

^aRace and ethnicity were obtained from medical records. Missing race and ethnicity data were accounted for in incidence calculations by assigning values to cases with missing data based on period, age group, and state-specific distribution of cases with known data.

Table 2.

Case fatality ratio among patients with bacterial meningitis identified by the Emerging Infections Programs Network, stratified by time period and pathogen, 2008–2023.

Case fatality rate	Streptococcus pneumoniae		Group B Streptococcus		Haemophilus influenzae		Neisseria meningitidis		Listeria monocytogenes		Overall
Overall	396/2523	(13.6%)	109/956	(10.2%)	24/535	(4.3%)	32/297	(9.7%)	12/144	(7.7%)	573/5028	(11.4%)
Time period 1 (2008–2009)	54/356	(15.2%)	11/103	(10.7%)	2/39	(5.1%)	9/89	(10.1%)	2/23	(8.7%)	78/610	(12.8%)
Time period 2 (2010–2019)	243/1899	(12.8%)	66/685	(9.6%)	14/378	(3.7%)	20/191	(10.5%)	5/93	(5.4%)	349/3247	(10.7%)
Time period 3 (2020–2021)	37/229	(16.2%)	14/138	(10.1%)	4/47	(8.5%)	2/21	(9.5%)	2/17	(11.8%)	59/452	(13.1%)
Time period 4 (2022–2023)	61/434	(14.1%)	18/139	(12.9%)	4/95	(4.2%)	1/28	(3.6%)	3/23	(13.0%)	87/719	(12.1%)
Chi-squared test Time period 1 vs Time period 2	0.23		0.71		0.73		0.93		0.55		0.13
Chi-squared test Time period 2 vs Time period 3	0.18		0.94		0.14		0.84		0.32		0.18
Chi-squared test Time period 3 vs Time period 4	0.54		0.59		0.51		0.80		1.00		0.70

Case fatality ratios were calculated using data from patients with a known outcome, which was 99.9% of patients.

Meningitis etiology-specific trends

S. pneumoniae was identified in 2,922 (58.1%) cases of meningitis. Serotyping data were available for 2,623 (89.8%) cases (Appendix 4). Between 2008–2009 and 2010–2019, the incidence of S. pneumoniae meningitis declined by 16.4% (95% CI: −25.2, −6.2) from 0.8 to 0.7 per 100,000 (Fig. 1, Table 1). This decline in incidence was only among PCV13 serotypes (Table 3). Between 2010–2019 and 2020–2021, S. pneumoniae meningitis incidence declined by 49.0% (95% CI, −55.6, −41.7) to 0.3 per 100,000 and was observed across all serotype groups. During 2022–2023, S. pneumoniae meningitis incidence increased by 87.0% (95% CI: 59.6, 119.6) to 0.6 per 100,000. This resurgence in incidence was significant across all serotype groups. There were changes in the distribution of pneumococcal serotypes over time.The proportion of NVT serotypes among serotyped isolates increased from 34.1% (112/328) during 2008–2009 to 46.7% (1073/2295) during 2010–2023. In terms of specific serotypes, vaccine serotypes 3, 19F, 22F and NVT serotypes 15C, 23B, 35B accounted for an increasing proportion of cases. Of the 364 serotyped isolates during 2022–2023, the most common serotypes were 3, 19F, 23B, and 22F accounting for 48 (13.2%), 37 (10.2%), 40 (11.0%), and 35 (9.6%) of isolates, respectively (Appendix 4.1). Of 2,606 isolates available for antimicrobial susceptibility testing against penicillin, cefotaxime, erythromycin, levofloxacin, and tetracycline, 1,099 (42.2%) exhibited non-susceptibility to at least one agent, with no significant change in non-susceptibility patterns over time (Appendix 4). S. pneumoniae remained the predominant causative pathogen across all periods, accounting for 58.4% (356/610) of cases in 2008–2009, 58.5% (1900/3247) in 2010–2019, 50.9% (231/454) in 2020–2021, and 60.3% (435/721) in 2022–2023 (Fig. 2).

Fig. 1.

Annual incidence of bacterial meningitis in Emerging Infections Programs Network sites, 2008–2023, stratified by pathogen.

Table 3.

Incidence of S. pneumoniae, group B Streptococcus, and H. influenzae Meningitis in Emerging Infections Programs Network sites, 2008–2022, stratified by relevant age and pathogen serotype groups.

	2008–2009		2010–2019		2020–2021		2022–2023		Percentage change in average annual incidence, 2010–2019 vs 2008–2009 (95% CI)	Percentage change in average annual incidence, 2020–2021 vs 2010–2019 (95% CI)	Percentage change in average annual incidence in 2022–2023 vs 2020–2021 (95% CI)
	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)	Cases	Incidence (95% CI)
Streptococcus pneumoniae	356	0.78 (0.7–0.86)	1900	0.65 (0.62–0.68)	231	0.33 (0.29–0.38)	435	0.62 (0.56–0.68)	−16.4 (−25.2, −6.2)	−49 (−55.6, −41.7)	87.0 (59.6, 119.6)
Age group
0–2 months	14	8.7 (5.15–14.69)	34	3.85 (2.75–5.38)	2	1.06 (0.27–4.25)	4	2.13 (0.8–5.67)	−55.8 (−75.8, −14.9)	−72.3 (−95.5, −9.3)	99.9 (−61.0, 1342.2)
3–11 months	23	4.76 (3.17–7.17)	95	3.58 (2.93–4.38)	7	1.24 (0.59–2.6)	17	3.01 (1.87–4.85)	−24.8 (−51.4, 21.4)	−65.4 (−85.4, −30.7)	142.7 (4.7, 527.8)
1–4 years	16	0.63 (0.39–1.03)	80	0.54 (0.44–0.68)	11	0.35 (0.19–0.63)	28	0.92 (0.63–1.33)	−13.4 (−47.9, 53.6)	−36.2 (−67.9, 14.7)	163.6 (35.0, 452.9)
5–17 years	20	0.25 (0.16–0.39)	132	0.27 (0.23–0.32)	10	0.09 (0.05–0.17)	45	0.4 (0.3–0.54)	6.8 (−31.6, 76)	−67.1 (−83.8, −40.7)	351.7 (137.6, 849.8)
18–49 years	105	0.5 (0.41–0.61)	517	0.4 (0.37–0.44)	58	0.19 (0.15–0.25)	112	0.37 (0.31–0.45)	−19.4 (−34.3, −0.1)	−52 (−63.8, −37.6)	91.7 (40.2, 164.8)
50–64 years	102	1.21 (0.99–1.47)	615	1.08 (1–1.17)	85	0.63 (0.51–0.78)	128	0.96 (0.81–1.14)	−10.6 (−27.2, 10.8)	−41.3 (−53.6, −26.9)	51.9 (15.7, 100.5)
≥65 years	76	1.44 (1.15–1.81)	427	1.07 (0.97–1.17)	58	0.52 (0.4–0.68)	101	0.87 (0.71–1.06)	−26 (−41.7, −4.9)	−51 (−63.1, −36.1)	65.9 (20.6, 130.4)
Serotypea
PCV13	152	0.33 (0.28–0.39)	511	0.18 (0.16–0.19)	54	0.08 (0.06–0.10)	136	0.19 (0.16–0.23)	−47.3 (−55.9, −36.6)	−55.7 (−66.9, −41.9)	150 (83.6, 245.4)
PCV15-unique serotypes	29	0.06 (0.04–0.09)	208	0.07 (0.06–0.08)	28	0.04 (0.03–0.06)	58	0.08 (0.06–0.11)	12.4 (−22.4, 69.3)	−43.5 (−62.7, −17.7)	105.6 (32.3, 227.4)
PCV20-unique serotype	54	0.12 (0.09–0.15)	289	0.1 (0.09–0.11)	28	0.04 (0.03–0.06)	48	0.07 (0.05–0.09)	−16.1 (−36.7, 13.3)	−59.4 (−73.0, −41.2)	70.2 (7.6, 174.4)
Non-vaccine serotypes	121	0.26 (0.22–0.32)	883	0.3 (0.28–0.32)	116	0.17 (0.14–0.20)	188	0.27 (0.23–0.31)	14.4 (−5, 39)	−44.9 (−54.8, −33.4)	60.9 (27.9, 103.3)
Group B Streptococcus	103	0.23 (0.19–0.27)	685	0.23 (0.22–0.25)	138	0.2 (0.17–0.23)	139	0.2 (0.17–0.23)	4.1 (−14.9, 28.8)	−15.5 (−29.9, 1.1)	0.0 (−21, 26.6)
Age group
Early-onsetb (per 1000 live births)	12	0.02 (0.01–0.03)	93	0.03 (0.02–0.03)	16	0.02 (0.01–0.03)	11	0.01 (0.01–0.03)	38 (−21.2, 165.2)	−19.5 (−54.3, 32.9)	−30.8 (−68.8, 47.8)
Late-onsetb (per 1000 live births)	51	0.08 (0.06–0.11)	350	0.1 (0.09–0.11)	66	0.09 (0.07–0.11)	61	0.08 (0.06–0.1)	22.2 (−8, 65.8)	−11.7 (−32.7, 14.0)	−6.9 (−34.4, 31.8)
0–2 monthsb	63	39.15 (30.58–50.11)	443	50.11 (45.65–55.0)	82	43.61 (35.12–54.15)	72	38.28 (30.38–48.22)	28 (−0.9, 68.2)	−13 (−31.7, 9.5)	−12.2 (−36.2, 20.4)
3–11 monthsb	6	1.24 (0.56–2.77)	33	1.24 (0.88–1.75)	10	1.77 (0.95–3.29)	6	1.06 (0.48–2.37)	0.1 (−54.9, 165.3)	42.5 (−33.4, 178.6)	−40 (−79.6, 61.5)
50–64 years	16	0.19 (0.12–0.31)	82	0.14 (0.12–0.18)	19	0.14 (0.09–0.22)	12	0.09 (0.05–0.16)	−24 (−54.2, 34.6)	−1.7 (−42.0, 58.3)	−36.3 (−69.9, 29.7)
≥65 years	3	0.06 (0.02–0.18)	65	0.16 (0.13–0.21)	16	0.14 (0.09–0.24)	23	0.2 (0.13–0.3)	185.3 (6.1, 1066.9)	−11.1 (−50.3, 49.5)	36.9 (−27.2, 163.8)
Haemophilus influenzae	39	0.09 (0.06–0.12)	378	0.13 (0.12–0.14)	47	0.07 (0.05–0.09)	96	0.14 (0.11–0.17)	52.3 (11.1, 115.2)	−47.8 (−62, −30.2)	102.8 (43.9, 189.7)
Age group
<5 years	14	0.44 (0.26–0.74)	155	0.86 (0.73–1.01)	21	0.54 (0.35–0.82)	24	0.63 (0.42–0.94)	95.6 (17.6, 254)	−37.6 (−61.6, −3.9)	17.5 (−34.6, 112.8)
5–64 years	19	0.05 (0.03–0.08)	146	0.06 (0.05–0.07)	15	0.03 (0.02–0.05)	47	0.09 (0.06–0.11)	22.6 (−21.9, 104.2)	−56 (−75.2, −27.7)	213 (79.4, 478.9)
≥65 years	6	0.11 (0.05–0.25)	77	0.19 (0.15–0.24)	11	0.1 (0.06–0.18)	25	0.22 (0.15–0.32)	69 (−19.8, 335)	−48.4 (−74.1, −7.2)	116.5 (9.2, 358.3)
Serotypec
Hib	4	0.01 (0.00, 0.02)	20	0.01 (0.00, 0.01)	5	0.01 (0.00–0.02)	0	–	−21.6 (−70.3, 169.4)	4.9 (−65.1, 159.0)	–
Non-b H. influenzae strains	36	0.08 (0.05–0.11)	358	0.12 (0.11–0.14)	41	0.06 (0.04–0.08)	96	0.14 (0.11–0.17)	60.3 (15.1, 130.9)	−52 (−65.7, −34.5)	132.5 (62.5, 238.5)
NTHi	19	0.04 (0.03–0.07)	194	0.07 (0.06–0.08)	22	0.03 (0.02–0.05)	61	0.09 (0.07–0.11)	60.1 (2.8, 165.1)	−52.4 (−70.2, −27.8)	175.3 (71.8, 357.8)
Non-b Hi serotypes	17	0.04 (0.02–0.06)	164	0.06 (0.05–0.07)	19	0.03 (0.02–0.04)	35	0.05 (0.04–0.07)	51.2 (−5.3, 158.6)	−51.4 (−70.8, −24.0)	82.9 (5.9, 225.9)

Incidence is reported as average annual incidence per 100,000 during 2008–2009, 2010–2019, 2020–2021, and 2022–2023. Percent changes in incidence with 95% CI were estimated using Poisson regression. The low number of meningitis cases due to N. meningitidis, and L. monocytogenes limited our ability to assess changes in incidence by more detailed demographic and subtype strata.

Missing serotype data for S. pneumoniae and H. influenzae were accounted for in incidence calculations by assigning serotype values to cases with missing data based on period, age group, and state-specific distribution of cases with known data, and thus resulted in small discrepancies between total cases and sum of cases by serotype strata.

^aPneumococcal serotype stratified analysis included: PCV13-serotypes (1, 3, 4, 5, 6A, 6B, 6C, 7F, 9V, 14, 18C, 19A, 19F, 23F), PCV15-unique-serotypes (22F, 33F), PCV20-unique-serotype (8, 10A, 11A, 12F, 15B), and non-vaccine serotypes (all serotypes not in PCV13, PCV15, or PCV20).

^bAs invasive neonatal GBS infections are further classified by age of onset as early-onset disease (<7 days) and late-onset disease (7–89 days), we calculated incidence of GBS meningitis in these age categories per 1000 live births. Additionally, to enable comparisons to previously published disease estimates, GBS meningitis incidence rates in infants aged 0–2 months and 3–11 months were calculated per 1000 live births. Incidence rates in infants aged 0–2 months were 0.10, 0.13, 0.11, and 0.10 per 1000 live births in 2008–2009, 2010–2019, 2020–2021, and 2022–2023, respectively. Incidence rates in infants aged 3–11 months were 0.010, 0.009, 0.013, and 0.012 per 1000 live births in 2008–2009, 2010–2019, 2020–2021, and 2022–2023, respectively.

^cH. influenzae serotype stratified analysis included: vaccine-preventable H. influenzae type b (Hib), nontypeable H. influenzae (NTHi), and serotypes other than Hib and NTHi (non-b Hi).

Fig. 2.

Proportion of all bacterial meningitis cases caused by each pathogen, stratified by epidemiologic period and age group — Emerging Infections Programs Network sites, 2008–2023.

GBS was identified in 1,065 (21.2%) cases of meningitis. GBS meningitis incidence did not change throughout the surveillance period and was highest among children aged 0–2 months. When looking specifically at neonatal meningitis, the incidence of late-onset disease was higher than early-onset disease across all periods (Table 3). There were changes in GBS etiological proportions over time; GBS accounted for 16.9% (103/610) of meningitis cases in 2008–2009, 21.1% (685/3247) in 2010–2019, 30.4% (138/454) in 2020–2021, and 19.3% (139/721) in 2022–2023. Of 738 serotyped GBS isolates, 342 (46.3%) were serotype III, with no changes in GBS serotype proportions over time. Of GBS meningitis isolates tested for antimicrobial susceptibility, isolates remained susceptible to common treatment antibiotics (penicillin, cefotaxime, vancomycin). An increasing proportion of isolates exhibited non-susceptibility to clindamycin and erythromycin, with 44.6% (41/92) and 58.2% (53/91) of isolates identified as non-susceptible during 2022–2023, respectively (Appendix 5).

H. influenzae was identified in 560 (11.1%) cases of meningitis. Subtyping data were available for 506 (90.4%) cases. Between 2008–2009 and 2010–2019, the incidence of H. influenzae meningitis increased by 52.3% (95% CI: 11.1, 115.2) from 0.09 to 0.13 per 100,000. This increase was only significant among children aged <5 years and NTHi, though a non-significant increase in meningitis caused by non-b Hi was also observed (Table 3). Between 2010–2019 and 2020–2021, incidence declined by 47.8% (95% CI −62.0, −30.2) to 0.07 per 100,000. This decline was significant among all age groups, and among NTHi and non-b Hi. During 2022–2023, there was a resurgence in H. influenzae meningitis incidence to 0.14 per 100,000 and was significant among the 5–64 and ≥ 65 years age groups as well as among NTHi and non-b Hi. Among 506 subtyped isolates, NTHi, and H. influenzae serotypes a (Hia), f, b, and e accounted for 266 (52.6%), 121 (23.9%), 75 (14.8%), 27 (5.3%), and 17 (3.4%) of cases, respectively. The proportion of Hia isolates increased over time from 12.1% (4/33) during 2008–2009 to 22.7% (17/75) in 2022–2023, though this was not statistically significant (Appendix 6).

N. meningitidis was identified in 329 (6.5%) cases of meningitis. Between 2008–2009 and 2010–2019, the overall incidence of N. meningitidis meningitis declined by 66.5% (95% CI: −73.9, −56.8) from 0.19 to 0.07 per 100,000 (Table 1). Between 2010–2019 and 2020–2021, incidence declined by 53.9% (95% CI: −71.5, −29.4) to 0.03 per 100,000. There was no change in incidence between 2020–2021 and 2022–2023. Overall, 299 (90.9%) isolates were serogrouped. Declines in N. meningitidis meningitis case counts were observed across all serogroups over time. However, there were changes in serogroup proportions; notably, the proportion of N. meningitidis serogroup Y isolates increased from 7.2% (5/69) in 2015–2019 to 36.8% (7/19) in 2020–2021, and 39.1% (9/23) in 2022–2023 (Appendix 7).

L. monocytogenes was identified in 156 (3.1%) cases of meningitis, of which 67 (42.9%) were among adults aged ≥65 years (Appendix 2). There was no change in the incidence of L. monocytogenes meningitis throughout the surveillance period.

National burden of meningitis

We estimate that approximately 3,944 cases of meningitis due to S. pneumoniae, GBS, H. influenzae, N. meningitidis, and L. monocytogenes, including 505 that were fatal, occurred annually in the United States during 2008–2009. This burden decreased to 3,511 cases and 376 fatalities annually during 2010–2019, and 2,132 cases and 273 fatalities annually during 2020–2021. During 2022–2023, there was an increase to 3,408 cases and 426 fatalities, annually.

Discussion

The epidemiology of bacterial meningitis in the United States continues to evolve as interventions targeting specific causes are implemented and sustained. Our study showed that the incidence of bacterial meningitis has declined since 2008. Prior to 2020, this decline in incidence was driven by decreases in S. pneumoniae and N. meningitidis meningitis, although there was a small increase in H. influenzae meningitis caused by NTHi and non-b Hi serotypes during this period. At the onset of the COVID-19 pandemic, there were dramatic decreases in meningitis caused by all respiratory bacterial pathogens. In 2022, a resurgence in meningitis incidence was observed, driven by an increase in S. pneumoniae and H. influenzae meningitis. There was no change in the incidence of meningitis caused by GBS or L. monocytogenes throughout the surveillance period. Additionally, there was little change in the case fatality of bacterial meningitis, overall or by pathogen. Finally, important differences remain in incidence by age, with young infants experiencing the highest burden of bacterial meningitis. Additionally, long-standing systemic health and social inequities have placed certain race and ethnic groups at at a disproportionately higher risk for poor health outcomes. These inequities likely contributed to observed differences in bacterial meningitis incidence across racial and ethnic populations. Targeted prevention strategies among such groups may be beneficial.

The magnitude and timing of the decline in meningitis incidence between 2008–2009 and 2010–2019, together with the observation that decreases in S. pneumoniae meningitis during this period were exclusively among PCV13-serotypes, suggest that PCV13 use contributed to this change. This contrasts with declines in incidence between 2010–2019 and 2020–2021, which were driven by decreases in meningitis caused by all respiratory bacteria, regardless of organism subtype, and support previously reported findings that COVID-19-associated non-pharmaceutical interventions reduced respiratory transmission of bacteria.^13,14 Moreover, as non-pharmaceutical interventions were eased from 2022, a resurgence in S. pneumoniae and H. influenzae meningitis was observed, with incidence rates returning to pre-COVID-19 pandemic levels in 2023.

Despite declines in incidence, S. pneumoniae remains the leading cause of bacterial meningitis in the United States. The introduction and high coverage of PCV13 vaccination (∼80%)¹⁵ has effectively reduced the incidence of S. pneumoniae meningitis due to certain vaccine serotypes. However, these declines in incidence are less than what was observed following PCV7 introduction.² We observed no increase in the incidence of pneumococcal meningitis due to non-vaccine serotypes following PCV13 introduction, consistent with overall invasive pneumococcal disease trends in the United States¹⁶ but in contrast to pneumococcal meningitis trends observed in other countries.^17,18 While there was no increase in NVT incidence, NVT serotypes now account for the majority of S. pneumoniae meningitis cases. Moreover, we observed an increase in the proportion of certain non-susceptible NVT serotypes such as 23B and 35B in later periods. This increase may explain why there was no change in non-susceptibility patterns among S. pneumoniae meningitis isolates, despite the decline in PCV13 serotypes over time. These findings underscore the potential value of recently recommended 15-, 20-, and 21-valent PCVs,9, 10, 11, 19 as well as higher valent PCVs in development,⁴ in reducing the remaning burden of pneumococcal meningitis and associated antimicrobial non-susceptibility in the United States.

The incidence of GBS meningitis did not change throughout the surveillance period. However, as meningitis incidence caused by other pathogens has declined, the etiological contribution of GBS to overall bacterial meningitis cases has increased. Additionally, GBS remains the most common etiology among infants aged 0–2 months, the age group with the greatest risk of meningitis. While universal screening for GBS among pregnant women and intrapartum antibiotic prophylaxis has reduced the incidence of early-onset invasive GBS disease, it has had no effect on late-onset disease,²⁰ which presents more commonly as meningitis. These findings highlight the value of maternal GBS vaccines, about to enter phase 3 trials, in reducing the residual burden of meningitis in the United States—particularly if they prove effective against late-onset and serotype III-associated GBS disease.

While meningitis due to H. influenzae type b (Hib) remained rare due to the high coverage of Hib vaccination (∼80%),¹⁵ we observed an increase in meningitis caused by NTHi and non-b Hi serotypes, though this increase in incidence was small in terms of absolute disease burden. We also observed an increase in the proportion of Hia disease over time. These trends are consistent with national and global increases in the incidence of invasive NTHi and non-b Hi serotype disease,21, 22, 23 including the specific increase in invasive Hia disease among American Indian and Alaska Native children.²⁴ Continued surveillance of H. influenzae is critical given the resurgence of NTHi and non-b Hi serotype-associated meningitis during 2022–2023, as well as due to advancements in vaccine development against these strains.^5,6

N. meningitidis meningitis incidence declined considerably through 2021 and now accounts for less meningitis cases than H. influenzae, which is an important consideration for empirical antimicrobial therapy guidelines. While we did not detect a change in N. meningitidis meningitis incidence between 2020–2021 and 2022–2023, national data reported through a related surveillance system, the Enhanced Meningococcal Disease Surveillance (EMDS), have shown a sharp increase in overall meningococcal disease cases beginning in late 2022. Additionally, EMDS data from 2023 report 422 cases of meningococcal disease in the United States, the highest number since 2014. This national increase in meningococcal disease has been driven primarily by serogroup Y,²⁵ which was also noted to account for an increasing proportion of N. meningitidis meningitis cases in this study. Continued surveillance of N. meningitidis is critical given these recent increases in serogroup Y disease, as well as due to the risk of N. meningitidis-associated outbreaks and emergence of multi-drug-resistant isolates.²⁶

While the incidence of L. monocytogenes meningitis was low, it did not decline significantly over the surveillance period. This is despite updates in guidance to prevent L. monocytogenes contamination, such as the U.S. Food and Drug Administration's guidance for control of Listeria contamination in ready-to-eat foods in 2017.²⁷ Enhanced efforts to prevent L. monocytogenes meningitis are warranted. Moreover, considering the high proportion of L. monocytogenes meningitis cases among adults aged ≥65 years, targeted interventions in this age group may be beneficial.

This study has limitations. First, our surveillance areas may not be representative of the United States as a whole, resulting in imprecision of national burden estimates due to socio-demographic or regional differences in rates of disease. Second, EIP sites only report cases with laboratory-confirmed meningitis, and therefore may not be detecting all bacterial meningitis cases, leading to an underestimation of incidence rates. Third, ABCs and FoodNet do not cover all possible causal pathogens of bacterial meningitis. Notably, Escherichia coli, Mycobacterium tuberculosis, and Group A Streptococcus were not included in this study despite being recognized as important or emerging causes of meningitis.28, 29, 30 Fourth, there were missing data on race, ethnicity, and organism subtype. These missing data were accounted for within incidence calculations by assigning values based on the distributions of cases for which these data were known. Fifth, clustering by state was not accounted for in the main analysis and may have resulted in an underestimation of uncertainty. Sixth, data on antimicrobial susceptibility for H. influenzae and N. meningitidis were unavailable for analysis in this study. Seventh, the low number of meningitis cases due to H. influenzae, N. meningitidis, and L. monocytogenes limited our ability to detect statistically significant differences or assess changes in incidence by more detailed demographic and organism subtype strata. Finally, our surveillance systems are unable to follow-up cases to assess the burden of bacterial meningitis-associated sequelae.

While meningitis represents only a small proportion of the total disease burden caused by the bacteria included in this study, it accounts for considerable morbidity and mortality. The Global Road Map for Defeating Meningitis by 2030 has the goal of reducing cases of vaccine-preventable bacterial meningitis since 2015 by 50% and deaths by 70%, through pillars such as prevention, diagnosis and treatment, and surveillance.³¹ Our findings indicate that while the incidence of bacterial meningitis in the United States has declined since 2008, largely through the sustained use of bacterial conjugate vaccines and through non-pharmaceutical interventions during the COVID-19 pandemic period, case fatality remains high. S. pneumoniae remains the predominant pathogen despite PCV13 introduction and high vaccination coverage; therefore, higher-valent PCVs could be beneficial. Preventative strategies that target GBS meningitis among young infants could also be beneficial given the high and unchanged disease incidence in this age group. The incidence of meningitis due to NTHi and non-b Hi serotypes has increased over time and highlights the value of vaccines targeting these strains. Finally, while N. meningitidis meningitis incidence declined considerably through 2021, and now accounts for less meningitis cases than H. influenzae, serogrouping and national surveillance data indicate an increase in increase in serogroup Y-associated meningococcal disease. Overall, our findings have implications for vaccine development and evaluation, as well as empirical antimicrobial therapy guidelines. Finally, we highlight the importance of continued meningitis surveillance including strain characterization.

Contributors

All authors contributed to study conception, design, and interpretation of the data. MA, AR, MB, SP, MMF, LHH, RL, JH, BJA, AT, KHT, and WS contributed to data collection. NP, GD did the statistical analyses. NP and MA had access to and verified the underlying data and drafted the initial manuscript. All authors provided critical revisions and final approval for the decision to submit for publication.

Data sharing statement

The data that support the findings of this study are available on request from the authors, NP and MA. The data are not publicly available due to their containing information that could compromise the privacy of research participants.

Disclaimer

The findings and conclusions in the report are those of the authors and do not necessarily represent the official position of the US Centers for Disease Control and Prevention.

Declaration of interests

LH reported support for attending meetings and/or travel from Pfizer and GSK and participation on an Advisory Board for Merck. WS reported payment for a lecture to Abbott Diagnostics. RL reported support for attending meetings of CSTE, NFID, COID, and ID Week; Leadership or fiduciary roles as Executive Officer, CSTE; Executive Officer NFID, Associate Editor AAP Red Book (Report of the Committee on Infectious Diseases-COID), and ID Week Program Committee; and support for work as an Associate Editor on the AAP Red Book. KT and BA reported grant support from US CDC. MF reported grant support unrelated to the manuscript from NIH. No other authors reported a conflict of interest.