Tuberculosis Outbreak Associated with a Homeless Shelter — Kane County, Illinois, 2007–2011

Despite the overall decline in tuberculosis (TB) incidence in the United States to a record low (1), outbreaks of TB among homeless persons continue to challenge TB control efforts. In January 2010, public health officials recognized an outbreak of TB after three overnight guests at a homeless shelter in Illinois received diagnoses of TB disease caused by Mycobacterium tuberculosis isolates with matching genotype patterns. As of September 2011, a total of 28 outbreak-associated cases involving shelter guests, dating back to 2007, had been recognized, indicating ongoing M. tuberculosis transmission. The subsequent investigation found that all patients were homeless and had been overnight shelter guests. Excess alcohol use was common (82%), and two bars emerged as additional sites of potential transmission. Patients with outbreak-associated TB were treated successfully for TB disease. To prevent future cases of TB, public health officials are implementing a program to offer 12 once-weekly doses of isoniazid and rifapentine under direct observation for treatment of latent tuberculosis infection (LTBI) (2) in this high-risk population. Although the United States has made progress toward TB elimination, this outbreak demonstrates the vulnerability of homeless persons to outbreaks of TB, highlighting the need for aggressive and sustained TB control efforts.

Initial Investigations

In April 2007, a man aged 55 years received a diagnosis of sputum smear–positive TB disease caused by an M. tuberculosis isolate with a genotype pattern* not documented previously in Kane County, Illinois. The man had been a frequent overnight guest at a Kane County facility that provided short-term shelter each night for approximately 180 persons whose housing situation was unstable. Subsequent case finding among other guests and staff members at the shelter identified no additional cases. In October 2009 and January 2010, two additional cases with the index patient's TB genotype pattern were identified among overnight shelter guests, alerting public health officials to a potential outbreak.

By March 2010, three additional cases with the outbreak genotype pattern had been identified among overnight shelter guests, leading county and state officials to request on-site epidemiologic assistance from CDC. Because all patients had been guests at the shelter, CDC recommended on-site case finding among guests and staff members at the shelter. The average length of stay at the shelter for guests was 2 weeks. During contact investigations and four mass screenings at the shelter during May 2010–June 2011, public health officials evaluated 386 persons recently exposed to a person with an infectious outbreak case, finding six (2%) additional TB cases.

During April 2007–July 2011, a total of 25 cases with the outbreak genotype pattern were identified (Figure). All patients had stayed overnight at the shelter, raising concern about ongoing transmission. The local health department concurrently identified approximately 10 TB cases each year unrelated to the outbreak, and the increased load during 2010 and 2011 led officials to request on-site assistance from CDC again in September 2011.

Subsequent Investigation

For the September 2011 investigation, a confirmed outbreak case was defined as TB disease having the outbreak genotype pattern diagnosed since April 2007 in a county resident. A suspected outbreak case was TB disease without an M. tuberculosis isolate available for genotyping (i.e., clinical disease), diagnosed since April 2007 in a county resident who had an epidemiologic link to a patient with a confirmed outbreak case. Investigators reviewed each eligible case to estimate infectious periods (3), identify potential sites of transmission, and determine epidemiologic linkages. Sources included medical records and interviews with patients or proxies, health department staff members, and shelter staff members.

As of September 23, 2011, a total of 28 outbreak cases had been identified (Table 1). Nearly one third of cases (29%) were detected through investigation-related activities (Figure, Table 1). Excluding one child, the median age was 49 years (range: 19–64 years) (Table 1). The one patient who had not slept in the men's sleeping area had known social connections (e.g., through alcohol consumption) to a patient who had slept in the men's sleeping area. Overall, 24 (86%) patients had connections through shared activities at the shelter or through shared behaviors (e.g., alcohol use at bar A). Of 25 with infectious pulmonary TB, 20 (80%) patients were present overnight at a location other than the shelter during their infectious periods, and the other five (20%) spent time at sites other than the shelter during the daytime.

To better understand the transmission dynamics, investigators conducted a case-control study. Because all outbreak patients had been overnight guests of the homeless shelter who had, with one exception, slept in the men's sleeping area, eligible case-patients were defined as men confirmed to be part of the outbreak (i.e., TB with the outbreak genotype) who had stayed overnight at the shelter at least once during August 2006 (i.e., the beginning of the index patient's infectious period) through July 2011 (i.e., the end of the last infectious period among men with confirmed outbreak TB). Controls were men who had stayed overnight at the shelter at least once during the same period but who had completed evaluations to exclude TB disease and LTBI (i.e., had a negative test for infection) and were asymptomatic at the time of interview.

Of the 25 patients eligible as case-patients, 17 (68%) enrolled in the case-control study. Of 72 men eligible as controls, 24 (35%) were located, and 23 (96%) met the inclusion criteria; all 23 enrolled. Although the small sample size limited the ability to detect statistically significant associations, longer duration of stay at the shelter, excess alcohol use, and occasional or frequent attendance at certain bars (A or B) had nonstatistically significant associations with being a case-patient (odds ratio ≥1.9) (Table 2). Because only 35% of eligible men could be located, selection bias of controls might have affected the outcome of this case-control study.

Public Health Interventions

In close collaboration with shelter staff members, public health officials have provided housing support, food, transportation, and treatment for TB disease by directly observed therapy to 24 of the 28 patients (i.e., excluding two patients who received care from other health jurisdictions, one who died, and one who was never located); all of these 24 patients with TB disease had completed or were continuing treatment as of December 2011. Supportive resources alone (i.e., excluding costs of health-care services) to provide successful treatment for these 24 patients with TB disease cost $204,500. Programmatic resources were not available to permit extension of these services to the 146 persons who had been exposed at the shelter and did not have TB disease but did have LTBI; 10 (7%) had completed LTBI treatment as of September 2011. Based on the subsequent investigation and case-control study, future case finding and LTBI treatment efforts will prioritize persons who slept in the men's area at the shelter and who socialized together at certain sites in the community. County and state officials have been working with the shelter to implement administrative control measures to reduce transmission at the shelter, including TB symptom screening upon admission to the shelter for overnight guests and evaluation for TB disease and infection for guests within 10 days of initial stay and annually. Although three additional outbreak cases were identified after the subsequent investigation, as of March 5, 2012, no further cases had been identified since December 2011.

Reported by

Claire Dobbins, MS, Kate Marishta, MPH, Paul Kuehnert, MS, Kane County Health Dept; Michael Arbisi, MS, Elaine Darnall, Craig Conover, MD, Illinois Dept of Public Health. Julia Howland, MPH, CDC/CSTE Applied Epidemiology Fellow; Krista Powell, MD, Sandy Althomsons, MPH, Sapna Bamrah, MD, Denise Garrett, MD, Maryam Haddad, MSN, Div of TB Elimination, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC. Corresponding author: Krista Powell, duf8@cdc.gov, 404-639-8120.

Editorial Note

Despite progress toward TB elimination (1), this outbreak demonstrates the vulnerability of persons affected by homelessness to outbreaks of TB, highlighting the need for aggressive and sustained TB control efforts. Outbreaks among persons experiencing homelessness are difficult to control, in part because of the challenges in finding and locating contacts and providing treatment for LTBI (4,5), as illustrated in this outbreak. Excess alcohol use and congregation in crowded shelters, which frequently are associated with homeless persons, increase their risk for TB (6–8). Of patients in this outbreak, 80% spent time at sites other than the shelter during their infectious periods, and attendance at certain bars had a nonstatistically significant association with being a case-patient, suggesting transmission was not limited to the shelter. Therefore, outbreaks of TB among homeless populations can pose a risk to entire communities.

Organizations that provide shelter and other types of emergency housing for homeless persons should develop institutional TB control plans (9). Other strategies to reduce TB transmission in shelters have included ventilation system improvements (9). In May 2010, the National Institute for Occupational Health and Safety conducted an on-site assessment of the heating, ventilation, and air-conditioning (HVAC) systems of the shelter associated with this outbreak, and along with appropriate administrative controls, recommended HVAC renovations to reduce TB transmission at the shelter. As of March 5, 2012, shelter and public health officials had secured funding for this renovation project, scheduled to begin in June 2012.

The first priority in TB control is to find and treat persons with active TB, but the second is to find and treat persons with LTBI to avert active cases of TB (9). The standard treatment for LTBI in the United States has been 9 months of isoniazid, but adherence rates have been low (approximately 60%), even in the absence of factors such as homelessness or substance use. CDC recently published guidelines for a shorter course LTBI treatment alternative, 12 doses of once-weekly isoniazid and rifapentine administered under direct observation (2), a regimen that public health officials in Illinois plan to offer persons exposed in this outbreak who have LTBI. Although large populations of homeless persons were not included in treatment trials (2), the practical advantages of this shorter regimen suggest the potential to transform the public health approach to LTBI.

TB outbreaks among homeless persons are resource-intensive, requiring provision of housing and other supportive services to patients (as in this outbreak), ongoing outreach, and TB case finding (7). Because this outbreak occurred during an economic downturn, available public health resources were constrained. Local policymakers had reorganized the health department in November 2010, transferring some health services to other health entities, reducing the health department's workforce by 50% (10). The dynamics of constrained resources have required close collaboration among local, state, and federal officials and the shelter to implement interventions. The extent to which M. tuberculosis was transmitted among persons experiencing homelessness in this outbreak provides a warning about the potential for loss of progress toward TB elimination if resources are shifted from TB control, particularly among vulnerable populations.

Acknowledgments

Shelter staff members; Sara Boline, MPH, Rita Bednarz, Marcia Huston, MD, Annette Julien, Mari Pina, Arlene Ryndak, MPH, Kathy Swedberg, Priya Verma, MD, Jeannie Walsh, Jeanette Zawacki, Judy Zwart, Kane County Health Dept, Illinois. Regina Gore, Dan Ruggiero, Div of TB Elimination, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, CDC.

References

1. CDC. Trends in tuberculosis—United States, 2011. MMWR 2012;61:181–5.
2. CDC. Recommendations for use of an isoniazid-rifapentine regimen with direct observation to treat latent Mycobacterium tuberculosis infection. MMWR 2011;60:1650–3.
3. CDC. Guidelines for the investigation of contacts of persons with infectious tuberculosis: recommendations from the National Tuberculosis Controllers Association and CDC. MMWR 2005;54(No. RR-15):1–49.
4. Reichler M, Reves RR, Bur S, et al. Evaluation of contact investigations conducted to detect and prevent transmission of tuberculosis. JAMA 2002;287:991–6.
5. Yun LWH, Reves RR, Reichler MR, et al. Outcomes of contact investigation among homeless persons with infectious tuberculosis. Int J Tuberc Lung Dis 2003;7(Suppl 3):S405–11.
6. Oeltmann J, Kammerer JS, Pevzner ES, Moonan PK. Tuberculosis and substance abuse in the United States, 1997–2006. Arch Intern Med 2009;169:189–97.
7. Haddad MB, Wilson TW, Ijaz K, Marks SM, Moore S. Tuberculosis and homelessness in the United States, 1994–2003. JAMA 2005;22:2762–6.
8. Lofy KH, McElroy PD, Lake L, et al. Outbreak of tuberculosis in a homeless population involving multiple sites of transmission. Int J Tuberc Lung Dis 2006;10:683–9.
9. CDC. Controlling tuberculosis in the United States: recommendations from the American Thoracic Society, CDC, and the Infectious Disease Society of America. MMWR 2005;54(No. RR-12).
10. Kuehnert PL, McConnaughay KS. Tough choices in tough times: enhancing public health value in an era of declining resources. J Public Health Manag Pract 2012;18:118–25.