Invasive Meningococcal Disease Among People Experiencing Homelessness — United States, 2016–2019

Abstract

Background:

Recently, several invasive meningococcal disease (IMD) outbreaks caused by Neisseria meningitidis have occurred among people experiencing homelessness (PEH). However, overall IMD risk among PEH is not well described. We compared incidence and characteristics of IMD among PEH and persons not known to be experiencing homelessness (non-PEH) in the United States.

Methods:

We analyzed 2016–2019 IMD data from the National Notifiable Diseases Surveillance System (NNDSS) and enhanced meningococcal disease surveillance. Incidence was calculated using U.S. census data and Point-in-Time counts from the U.S. Department of Housing and Urban Development.

Results:

Of cases from states participating in enhanced surveillance during 2016–2019 (n=1409), 45 (3.2%) cases occurred among PEH. Annual incidence was higher among PEH (2.12 cases/100,000) than non-PEH (0.11 cases/100,000; relative risk: 19.8, 95% CI: 14.8–26.7). Excluding outbreak-associated cases (PEH n=18, 40%; non-PEH n=98, 7.2%), incidence among PEH remained elevated compared to incidence in non-PEH (relative risk: 12.8, 95% CI: 8.8–18.8). Serogroup C was identified in 68.2% of PEH cases compared to 26.4% in non-PEH (p<0.0001).

Conclusions:

PEH are at increased risk for IMD. Further assessment is needed to determine the feasibility and potential impact of meningococcal vaccination for PEH in the United States.

Background

Invasive meningococcal disease (IMD) is a rare, life-threatening infection caused by the bacterium Neisseria meningitidis. Meningitis and/or meningococcemia are the most common clinical presentations. IMD progresses rapidly with a case-fatality rate of about 15%, even among those who receive appropriate treatment. Up to 20% of survivors experience debilitating permanent sequelae including loss of limbs, cognitive disability, and deafness [1].

Most IMD cases in the United States are sporadic, but IMD outbreaks can also occur particularly in populations living in close quarters, such as students living on a college campus. Recently, several IMD outbreaks have been reported among people experiencing homelessness (PEH) [2,3]. However, IMD among PEH in the United States has not been systematically described since housing status was not historically collected through national meningococcal disease surveillance.

In 2016, a housing status variable was incorporated into national meningococcal surveillance to enable a clearer understanding of meningococcal disease risk among PEH. Here, we analyze 2016–2019 U.S. meningococcal disease surveillance data from 45 states and three large jurisdictions to compare incidence and characteristics of IMD among PEH and people not known to be experiencing homelessness (non-PEH). We also describe outbreaks among PEH reported during this period.

Methods

To determine whether PEH are at increased risk for IMD, data on IMD cases reported to the National Notifiable Diseases Surveillance System (NNDSS) during 2016–2019 among 45 states and three large jurisdictions participating in enhanced meningococcal disease surveillance (EMDS) were analyzed [4]. Health departments assessed patient housing status at the time of meningococcal disease onset using the U.S. Department of Housing and Urban Development’s (HUD) definition for homelessness [5]. Cases in persons known not to be experiencing homelessness or with unknown or missing housing status were classified as non-PEH. Information on outbreaks affecting PEH was compiled from health department records. Meningococcal serogroup was determined by slide agglutination [6]; sequence type (ST) and clonal complex (CC) were determined by whole genome sequencing at CDC [7]. Incidence was calculated using U.S. Census Bureau Bridged-Race Population Estimates and HUD Point-in-Time (PIT) counts of PEH for jurisdictions participating in EMDS [8–11]. PIT counts estimate sheltered and unsheltered PEH on one night in January each year and were used as a proxy for person-years of homelessness among U.S. residents in the respective year [12,13]. Chi-squared, Fisher’s exact, and Wilcoxon signed-rank tests were used to assess differences between IMD cases in PEH and non-PEH. Poisson regression was used to calculate relative risk (RR) estimates and 95% confidence intervals.

Outbreaks among people experiencing homelessness and associated public health responses were characterized based on review of jurisdiction public health records.

This activity was reviewed by CDC and was conducted consistent with applicable federal law and CDC policy.¹

Results

Of 1,409 IMD cases reported to NNDSS during 2016–2019 from jurisdictions participating in EMDS, 45 (3.2%) cases from 17 states occurred among PEH (Table 1). Forty percent (18/45) of PEH cases were outbreak associated, compared with 7.2% (98/1364) of non-PEH cases (p<0.0001; Table 1). PEH IMD patients were older than non-PEH patients, with only two cases identified among PEH aged less than 18 years old (Table 1). PEH patients were also more often male than non-PEH patients; race, ethnicity, and case-fatality ratio (CFR) were not significantly different between the two groups although the CFR among PEH was less than half of that among non-PEH (Table 1). Among 44 PEH cases with known meningococcal serogroup, 30 (68.2%) were caused by serogroup C (NmC) compared to 329 (26.4%) of 1,244 non-PEH cases (p<0.0001). Among NmC cases with whole genome sequencing data available, 20 of 28 (71.4%) cases in PEH were CC103/ST-2006 compared to 119 of 282 (42.2%) cases in non-PEH. Of the 20 CC103/ST-2006 NmC cases in PEH, 10 (50.0%) were from an outbreak in Boston, MA. Bacteremia was more common among cases in PEH than in non-PEH (p=0.001) (Table 1); this may be related to the difference in serogroup distribution between these groups [14].

Table 1:

Demographics and other characteristics of invasive meningococcal disease cases among people experiencing homelessness (PEH)^* vs. not known to be experiencing homelessness (non-PEH) — United States, 2016–2019.

	PEH	Non-PEH	p-value	Total
N	45	1364		1409
Age – median, years (IQR)	52.0 (26.0)	34.0 (43.5)	p=0.008
Age group – n (%)
<18 years	2 (4.4)	341 (25.0)	p=0.002	343
18 years and older	43 (95.6)	1023 (75.0)		1066
Sex – n (%)
Male	35 (77.8)	694 (50.9)	p=0.0004	729 (51.7)
Female	10 (22.2)	670 (49.1)		680 (48.3)
Race – n (%) ɫ
White	30 (69.8)	898 (72.0)	p=0.53	928 (71.9)
Black	10 (23.3)	219 (17.6)		229 (17.8)
Other	3 (7.0)	130 (10.4)		133 (10.3)
Ethnicity – n (%) ɫ
Hispanic	8 (20.0)	204 (17.1)	p=0.63	212 (17.2)
Not Hispanic	32 (80.0)	992 (82.9)		1024 (82.8)
Outcome – n (%) ɫ
Survived	43 (95.6)	1166 (87.5)	p=0.11	1209 (87.8)
Died	2 (4.4)	166 (12.5)		168 (12.2)
Syndrome – n (%) ɫ
Bacteremia	25 (62.5)	390 (34.9)	p=0.001	415 (35.9)
Meningitis	11 (27.5)	591 (53.0)		602 (52.1)
Other	4 (10.0)	135 (12.1)		139 (12.0)
Serogroup – n (%) ɫ
B	5 (11.4)	472 (37.9)	p<0.0001	477 (37.0)
C	30 (68.2)	329 (26.4)		359 (27.9)
W	3 (6.8)	113 (9.1)		116 (9.0)
Y	4 (9.1)	194 (15.6)		198 (15.4)
NG	2 (4.6)	124 (10.0)		126 (9.8)
Other	0	12 (1.0)		12 (0.9)
Outbreak Associated – n (%)
Yes	18 (40.0)	98 (7.2)	p<0.0001	116 (8.2)
No	23 (51.1)	1171 (85.9)		1194 (84.7)
Unknown	4 (8.9)	95 (7.0)		99 (7.0)
HIV§ Status – n (%)ɫ
Positive	0	34 (4.8)	p=0.62	34 (4.6)
Negative	24 (100)	675 (95.2)		699 (95.4)
MSM¶ – n (%)ɫ
Yes	2 (9.5)	66 (19.4)	p=0.39	68 (18.8)
No	19 (90.5)	317 (80.7)		336 (81.2)

^*Health departments assessed patient housing status using the U.S. Department of Housing and Urban Development’s definition for homelessness.

^ɫPercentages shown in parentheses are calculated based on the total with known information for these characteristics: race, ethnicity, outcome, syndrome, serogroup, HIV status, and MSM.

^§Human immunodeficiency virus.

^¶Men who have sex with men.

The PIT counts of PEH ranged from 522,567 to 544,472 per year, resulting in an estimated total of 2,122,003 person-years at risk among PEH during 2016–2019. IMD incidence was 19.8 (95% CI: 14.8–26.7) times higher among PEH (2.1 per 100,000 person-years at risk) than non-PEH (0.11 per 100,000) (Table 2). Excluding outbreak-associated cases (n=18), incidence among PEH (1.3 per 100,000) remained 12.8 (95% CI: 8.8–18.8) times higher than in non-PEH. When limiting to adults aged 18 years or older, the relative risk among PEH compared with non-PEH was slightly higher both for total (RR: 24.6, 95% CI: 18.1–33.3) and sporadic (RR: 16.8, 95% CI: 11.5–24.7) meningococcal disease (Table 2).

Table 2:

Incidence and relative risk of invasive meningococcal disease in people experiencing homelessness (PEH) and people not known to be experiencing homelessness (non-PEH)^* (per 100,000) – United States, 2016–2019.

Year	Total Number of Cases	Cases in PEH	Estimated PEH population	Estimated Incidence among PEH	Cases in non-PEH**	Estimated non-PEH population	Estimated Incidence among non-PEH	Relative risk (PEH / non-PEH) (95% confidence interval)
Total	1409	45	2,122,003	2.12	1364	1,276,827,013	0.11	19.8 (14.8–26.7)
2016	372	12	522,567	2.30	360	316,691,173	0.11	20.2 (11.4–35.9)
2017	345	8	526,026	1.52	337	318,372,999	0.11	14.4 (7.1–29.0)
2018	324	16	528,938	3.02	308	320,334,199	0.10	31.5 (19.0–52.0)
2019	368	9	544,472	1.65	359	321,428,642	0.11	14.8 (7.6–28.7)
Sporadic only	1293	27	2,122,003	1.27	1266	1,276,827,013	0.10	12.8 (8.8–18.8)
Adults aged 18 and older	1066	43	1,690,399	2.54	1023	988,466,231	0.10	24.6 (18.1–33.3)
Sporadic cases among adults aged 18 and older	966	27	1,690,399	1.60	939	988,466,231	0.09	16.8 (11.5–24.7)

^*Health departments assessed patient PEH status using the U.S. Department of Housing and Urban Development’s definition for homelessness.

^**Includes 187 cases with unknown housing status: 136 in 2016, 17 in 2017, 16 in 2018, and 18 in 2019

Outbreaks among PEH and public health response

Of the 18 outbreak cases among PEH, 17 were part of three outbreaks in Massachusetts, Oregon, and Colorado. (The final outbreak case was part of an outbreak affecting an otherwise housed population.) The Massachusetts Department of Public Health reported 13 outbreak cases, including 10 NmC and three NmY, during 2016–2019 among individuals from three separate adult homeless shelters and unsheltered individuals; no direct epidemiologic links among cases were identified. Three additional outbreak-related cases were reported in non-PEH associated with facilities that provide services to people experiencing homelessness. Most of the 13 cases in PEH occurred in males (n=11) with a median age of 46 years. All 10 NmC cases were caused by CC103/ST-2006 and the two NmY cases with WGS results were caused by CC23/ST-1655. The size of the population affected by the outbreak was around 3,384 individuals experiencing homelessness in the Boston area; families experiencing homelessness were excluded because there were no cases within the family shelter system, which operates separately from the adult shelter system. For all cases, close contact investigations were performed within the shelter settings and chemoprophylaxis (single dose of oral ciprofloxacin; oral rifampin for fluoroquinolone-allergic individuals) was offered and accepted by 286 out of 307 close contacts.

A MenACWY vaccination campaign among PEH and front-line shelter staff in the Boston area launched on February 17, 2016. Vaccination began after the first four NmC cases and one NmY case; seven additional cases occurred after the vaccination campaign began but none in individuals who received meningococcal vaccine. Vaccination efforts were sustained across several years given the protracted nature of the outbreak. By March 16, 2018, approximately 6000 individuals in the shelter community were vaccinated with one or more doses of MenACWY vaccine.

The Oregon Health Authority reported an outbreak of two NmC cases among 100 residents of a homeless shelter in 2018. One case occurred in a 13-year-old male and the other in a 7-year-old female; isolates from both cases were CC11/ST-11. A vaccination campaign targeted the 24 shelter residents aged <30 years; 17 (71%) were vaccinated with one or more doses of MenACWY vaccine. Expanded chemoprophylaxis was administered to 81 out of 108 targeted individuals, which included residents and the directors of the shelter. No additional cases occurred among shelter residents after vaccination and administration of chemoprophylaxis.

In 2019, the Colorado Department of Public Health and Environment (CDPHE) investigated four cases of CC11/ST-11 NmW associated with the population experiencing homelessness in Denver, CO. Two of the cases occurred in PEH while two occurred in non-PEH who had close contact with PEH. Only one of the PEH patients reported recent shelter use and none of the four cases were epidemiologically linked. An outbreak of hepatitis A was simultaneously occurring in PEH, which initially limited resources for the IMD outbreak response, but plans to deliver MenACWY vaccine at hepatitis A vaccine clinics for PEH began after detection of the third IMD case in October 2019. CDPHE and Denver Public Health implemented the first multi-vaccine immunization clinic on December 2, 2019. While the total size of the Denver PEH population was estimated to be 5,700, vaccination efforts focused on users of overnight shelters and transitional housing, estimated to be approximately 3,000 people. MenACWY vaccination clinics continued through February 2020 with a total of 1,748 doses of MenACWY vaccine administered. Vaccination administration sites included the Denver County Jail, “backpack” clinics at an encampment, and outreach in HIV/STI and substance use disorder clinics.

Discussion

This analysis indicates that IMD incidence in PEH was 19.8 times higher than in non-PEH in the United States during 2016–2019, or 24.6 times higher among adults aged 18 years and older. After excluding outbreak cases, which accounted for 40% of PEH cases, the incidence of IMD in PEH remained 12.8 times higher than in non-PEH, or 16.8 times higher among adults. The increased IMD risk among PEH is similar to the increased risk for other infectious diseases in this population [12, 13, 14]. Environmental conditions, such as crowded shelters, and a high prevalence of underlying conditions among PEH, may contribute to the increased risk for IMD and other infectious diseases [12].

Currently, the only vaccine recommended specifically for PEH is hepatitis A [15]; there are no Advisory Committee on Immunization Practices (ACIP) recommendations for routine meningococcal vaccination for PEH. Although meningococcal vaccination is recommended in response to outbreaks, vaccination campaigns among PEH can be challenging due to difficulties defining the target population for vaccination and obtaining high vaccine uptake. Lessons learned from recent experiences vaccinating PEH during outbreaks of hepatitis A or in response to the COVID-19 pandemic may be leveraged to improve vaccination coverage during future outbreak responses in this population. However, our analysis shows that even outside of outbreaks, PEH face an increased IMD risk comparable to that found among people living with HIV (PLWH), a population for which the ACIP recommended routine meningococcal vaccination in 2016 [16]. The similar increased IMD risk among PEH suggests it may be valuable to explore the feasibility and potential benefits of meningococcal vaccination for PEH beyond the context of outbreaks.

This analysis has several limitations. Housing status was missing for 40% of cases for 2016. While completeness of housing status improved to 95% for 2017–2019, missing data during 2016 may have led to misclassification of PEH as non-PEH since, as described in the methods, all cases with unknown housing status were classified as non-PEH. This potential misclassification may have resulted in underestimation of the differences between the PEH and non-PEH populations. However, excluding 2016 data did not substantially change the findings. Data on risk factors potentially contributing to the increased IMD incidence among PEH, such as exposure to elements, underlying conditions, and substance abuse, were not available for this analysis. Data were also not systematically collected on whether PEH with cases of meningococcal disease were experiencing sheltered or unsheltered homelessness, preventing us from assessing whether meningococcal disease risk may differ between sheltered and unsheltered PEH. Finally, the use of PIT counts captures PEH on a single night and may underestimate the total number of person-years at risk among the PEH population [17,18]. However, even if we assume the true denominator of person-years at risk is three times larger than the estimate from the PIT counts, the overall relative risk for IMD in PEH during 2016–2019 would still be 7-fold greater than the risk in non-PEH.

Despite these limitations, this analysis demonstrates a clear disparity in IMD risk among PEH compared with non-PEH. While this finding suggests that PEH could benefit from routine MenACWY vaccination, challenges exist for vaccinating PEH including defining and reaching the population. The cost-benefit ratio is also unclear, in part because the PEH population fluctuates, and so a proportion of vaccinated individuals may only continue to be at increased risk for a short time after vaccination. Further assessment of the feasibility and potential impact of routine meningococcal vaccination for PEH in the United States would be beneficial to inform optimal prevention strategies for this population.