TB Notes Newsletter
No. 2, 2008
and Health Systems
Study 26 Reaches Enrollment
Study 26 has reached its patient enrollment goal. As of 2:32 pm
on January 29, 2008, the Tuberculosis Trials Consortium (TBTC) was
very happy to report that 8000 patients were enrolled in USPHS/TBTC
Study 26. TBTC wishes to thank the volunteer patients, community
members, health professionals, and scientists involved in conducting
and supporting this study.
—Reported by Elsa Villarino, MD
Div of TB Elimination
Tuberculosis Diagnosis at Death Among HIV-Infected Persons:
US Metropolitan Statistical Areas, 1998–2003
TB remains a potentially deadly disease for those living with HIV
infection. Deaths can occur during TB treatment, or prior to TB
disease diagnosis, which is referred to as “TB diagnosis at death.”
HIV increases the risk of death during TB treatment, especially for
those having multidrug-resistant (MDR) TB. National TB Surveillance
System data for 1997–1999 show that 40% of HIV-infected persons with
MDR TB died during treatment, compared with 8% of HIV-uninfected
persons with MDR TB (Moore M, unpublished data, January 2003).
Analysis of National TB Surveillance System data excluding
California from 1997 to 2005, adjusted for MDR TB, age, residency in
a longterm care facility, US birth, race/ethnicity, sex, substance
abuse, and other disease characteristics indicates that HIV-infected
persons’ odds of death during TB treatment was five times higher
than that of HIV-uninfected persons (Marks SM and Magee E,
unpublished data, March 2008). The same study also revealed that
HIV-infected patients had a five times greater odds of being
diagnosed at death, adjusted for age, US birth, sex, race/ethnicity,
and other disease characteristics. The current study examines trends
in TB diagnosis at death among HIV-infected persons in the 10 most
HIV-prevalent US metropolitan statistical areas (MSAs) during
1998–2003. Diagnosis of TB at death may reflect lack of ready access
to TB or HIV prevention, diagnosis, and treatment services.
Identifying groups or areas with high rates of TB diagnosis at death
in HIV-infected persons may help pinpoint disparities in access to
HIV care, access and provision of TB care, or both.
To identify those diagnosed at death, we used 1998–2003 data on
“status at diagnosis of TB” from the National TB Surveillance System
for 20 MSAs having the highest reported TB/HIV co-morbidity in
numbers of patients. We report results individually, but
anonymously, for the top 10 MSAs having the highest TB/HIV
co-morbidity. HIV status was classified into positive (from a
documented test, previous HIV/AIDS diagnosis, or self report),
negative (from a documented test within the past year), or unknown
(indeterminate results, refused testing, not offered testing,
results unknown, unknown). We first compared the risk of being
diagnosed with TB at death for HIV-infected versus HIV-uninfected
patients (excluding those from the three California sites since the
state of California only provides CDC access to AIDS case registry
and TB registry matching results; data are limited to reported AIDS
cases, and exclude HIV negativity or unknown status). Then, we
examined those groups of TB patients having unknown HIV status,
comparing those diagnosed at death with those diagnosed while alive.
For known HIV-infected TB patients, we then identified factors
associated with TB diagnosis at death by using multivariate
log-binomial analysis to obtain adjusted prevalence ratios.
In rank order from highest to lowest prevalence, the top 10
TB/HIV-prevalent MSAs were New York City, Miami, Los Angeles,
Houston, Chicago, Atlanta, Ft. Lauderdale, Newark, Dallas, and
Washington, DC. In these 10 MSAs, TB/HIV patients comprised 14% of
all TB patients, or 25% of those with known HIV status (range:
7%–32% of all patients or 17%–37% of patients with known HIV
status). HIV-infected TB patients in the top 10 MSAs had the
following demographics: male (72%), non-Hispanic black (58%),
foreign born (36%), Hispanic (25%), missing race/ethnicity (8%),
non-Hispanic white (7%), Asian/Pacific Islander (2%), age 65 or over
(2%), and American Indian/Alaska Native (0.1%).
Regardless of HIV status, 2.0% of all patients in the top 10 MSAs
were diagnosed with TB at death (range: 1.0%–3.4%). The percentage
of HIV-infected TB patients who were diagnosed at death varied from
year to year by MSA, but appeared to decline over time. The median
and average TB diagnoses at death per year for HIV-infected patients
in the top 10 MSAs was 3.2%, but averaged 6.6% in MSA 10; 5.6% in
MSA 3; and 4.4% in MSA 6.
HIV-infected patients had on average 4.9 times (95% confidence
interval [CI]: 3.8-6.2) greater risk than HIV-uninfected patients to
have been diagnosed at death (vs. while alive) during 1998–2003.
This significantly greater unadjusted risk of TB diagnosis at death
for HIV-infected patients was true at all but one of the MSAs (RR
range 2.7–8.4) (Figure).
Those diagnosed at death were 1.75 times (95% CI: 1.64-1.86) more
likely to be missing HIV status information. This association held
true for seven of the nine highest non-California TB/HIV-prevalent
MSAs, with RRs ranging from 1.6 to 11.5. We found that Asians
(RR=1.6, 1.5–1.6), non-Hispanic whites (RR=1.4, 1.3–1.4), and
foreign-born persons (RR=1.1, 1.05–1.12) were significantly more
likely, and non-Hispanic blacks (RR=0.6, 0.6–0.7), males (RR=0.8,
0.78–0.83), and Hispanics (RR=0.9, 0.86–0.92) less likely to have
unknown HIV status.
From multivariate analysis, we found (as expected) that HIV-infected
TB patients aged 65 or over had 3.7 times greater risk of diagnosis
with TB at death than younger HIV-infected patients. HIV-infected
Hispanics had half the risk as non-Hispanics and HIV-infected
foreign-born patients had nearly half the risk as US-born patients
to be diagnosed at death. In three MSAs, there was a 2.1-2.7 times
greater risk of being diagnosed at death than in the remaining MSAs.
In one MSA, there was a 70% lower risk of diagnosis at death than in
the other MSAs. MDR TB was not significantly associated with being
diagnosed at death, nor was TB disease that was strictly
The following limitations apply to this study. Deaths during TB
treatment or TB diagnoses at death may not be deaths caused by TB,
which requires verification through autopsies (which have declined
from 41% of hospital deaths in 1961 to 5%–10% in the mid-1990s in
the United States)1, from specimens obtained just prior to death, or
from lab results received after death. In Italy, an autopsy study of
350 AIDS patients found 20 (6%) patients with TB were undiagnosed
while alive.2 In the absence of US autopsy studies, one US study
found that 22% of a 1997 TB/HIV cohort died, and 44% of the deaths
were from TB; the study used sputum smears at death to diagnose
pulmonary TB and tissue cultures for extrapulmonary TB.3 Some
sociodemographic factors of known HIV-infected patients could not be
examined because some data were missing for those diagnosed with TB
at death: data on injecting and noninjecting drug use for 28%-29%,
and data on alcohol abuse for 29% (compared with 5%-6% missing data
for those factors of those diagnosed while alive). Similarly, we
found that homelessness data were missing for 9% of those dead at
diagnosis (vs. 3% of those alive), residence in a longterm care
facility data for 2% (vs. 0.2% of those alive), and residence in a
correctional facility data for 1% (vs. 0.2% of those alive).
While TB mortality in HIV-infected persons has declined over time
because of reductions in US TB incidence and prevalence,
implementation of highly active antiretroviral therapy (HAART),4,5
and improvements in TB and HIV case management,6 HIV-infected TB
patients had nearly five times the unadjusted risk of being
diagnosed with TB at death as HIV-uninfected persons in the highest
TB/HIV-prevalent MSAs from 1998–2003. Among HIV-infected persons, TB
patients aged ≥ 65 years had three times the risk of younger
patients, and those residing in three high TB/HIV-prevalent areas
had risks of diagnosis at death two to three times higher than those
in other areas.
TB diagnosis at death reflects both the risk of death with
undiagnosed TB and the risk of a diagnosis of TB after death.
HIV-infected patients aged 65 or older might have a greater risk of
death with undiagnosed TB because of missed diagnoses due to
age-related immunosuppression,7 which reduces the sensitivity of the
tuberculin skin test (TST) and results in atypical clinical and
radiographic presentation. On the other hand, HIV-infected Hispanics
were half as likely to be diagnosed with TB at death as other
race/ethnic groups, and foreign-born patients 40% as likely. To
possibly explain this, first we note that diagnosis of TB after
death generally requires either a specimen taken for examination
just prior to death or an autopsy (half of which are performed when
patients die in or upon arrival at a hospital).8 Second, uninsured
TB patients stay fewer days in a hospital9 and are less likely than
those insured to have usual diagnostic services and access to care.
Thus, it is possible that Hispanics, who are less likely to have
insurance than other groups,10 were not hospitalized or examined
prior to death or autopsied after death. Alternatively, it is
possible that Hispanics are more likely to be diagnosed with TB
while alive because of frequent screening efforts taking place in
Hispanic or foreign-born communities because of immigration
requirements. We might expect both of the above-mentioned
possibilities to be occurring.
Additional data collection is needed for those diagnosed with TB at
death. The extent to which autopsies include HIV testing is unknown.
Post-mortem medical record reviews may help achieve completion of
data collection on alcohol abuse, injecting drug use, or
noninjecting drug use. Efforts are needed to obtain missing
information on homelessness, residence in a long-term–care facility,
or residence in a correctional facility for known HIV-infected TB
patients diagnosed at death.
Since a diagnosis of TB at death is a late diagnosis, it should be
considered a sentinel event for evaluation of missed prevention
opportunities. Alternatively, favorable practices should be
identified at locations where there are lesser risks of diagnoses at
death. Among TB patients, routine voluntary HIV testing (including
opt-out testing in clinical settings) and referral to HIV care are
needed, along with routine screening of known HIV-infected persons
for symptoms of TB disease, testing for latent TB infection (LTBI),
and LTBI treatment completion to prevent deaths in HIV-infected
persons at risk for TB. It may also be helpful to track the effect
of HAART on reductions in TB diagnosed at death among HIV-infected
—Reported by Suzanne M. Marks, MPH, MA;
Todd Wilson, MS; and Valerie Robison, DDS, PhD
- CDC, National Center for Health Statistics. The autopsy,
medicine, and mortality statistics. Vital and Health Statistics
- D’Arminio Monforte A, Vago L, Gori A, Antinori S, Franzetti F,
Antonacci CM, Sala E, Catozzi L, Testa L, Esposito R, Nebuloni M,
Moroni M. Clinical diagnosis of mycobacterial diseases versus
autopsy findings in 350 patients with AIDS. European Journal of
Clinical and Microbiological Infectious Diseases 1996; 15(6):453-8.
- Leonard MK, Larsen N, Drechsler H, Blumberg H, Lennox JL,
Arrellana M, Filip J, and Horsburgh CR. Increased survival of
persons with tuberculosis and human immunodeficiency virus
infection, 1991-2000. Clinical and Infectious Disease
- Jones JL, Hanson DL, Dworkin MS, DeCock KM, and the
Adult/Adolescent Spectrum of HIV Disease Group. HIV-associated
tuberculosis in the era of highly active antiretroviral therapy. The
adult/adolescent spectrum of HIV disease group. International
Journal of TB and Lung Disease 2000;4(11):1026-1031.
- The Antiretroviral Therapy Cohort Collaboration. Incidence of
tuberculosis among HIV-infected patients receiving highly active
antiretroviral therapy in Europe and North America. Clinical
Infectious Diseases 2005;41:1772-1782.
- Munsiff SS, Ahuja SD, Driver CR. Public-private collaboration for
multidrug-resistant tuberculosis control in New York City.
International Journal of Tuberculosis and Lung Disease
- Nijhuis EW, Nagelkerken L. Age-related changes in immune
reactivity: the influence of intrinsic defects and of a changed
composition of the CD4+ T cell. Exp Clin Immunogenetics
- Hoyert DL, Kung HC, Xu J. Autopsy patterns in 2003. National
Center for Health Statistics. Vital Health Statistics 2006;20(32).
- Marks SM, Taylor Z, Ríos Burrows N, Qayad MG, Miller B.
Hospitalization of homeless persons with tuberculosis in the United
States. American Journal of Public Health 2000;90(3):435-438.
- CDC. Health disparities experienced by Hispanics–United States.
New Publications Available Soon!
Look for a new series expected to be released over the next few
months, entitled Promoting Cultural Sensitivity: A Practical Guide
for Tuberculosis Programs Providing Services to Foreign-born
Persons. The series comprises five modules, each focused on a
distinct cultural group: Chinese, Hmong, Mexican, Somali, and
Vietnamese. Each guide contains chapters on the selected group’s
history and immigration; culture; health issues; and common
perceptions, attitudes, and beliefs about TB. A product of DTBE’s
2003 ethnographic study of foreign-born persons in the United
States, the series aims to help TB program staff provide culturally
competent TB care to some of our highest priority foreign-born
Intended for health care providers, community-based workers, program
planners, administrators, health educators, and resettlement
agencies that work with the five selected foreign-born populations,
the guides are designed to increase the knowledge and cultural
sensitivity of providers serving these populations. The ultimate aim
is to foster culturally competent TB care and services for
foreign-born populations in the United States.
About the Guides
Each guide in the series includes the following:
- A 2-page summary of program tips
- Chapters on history and immigration; culture; health issues; and
common perceptions, attitudes, and beliefs about TB
- A concluding summary
- Appendices, including additional resources for working with TB
patients and interpreters and a glossary of terms
- Useful resources
Some of the information in the guides, such as the practical tips,
can be applied directly, while other sections are more informative
and will help providers better understand the background and
sociocultural context of the population. A deeper understanding of
pertinent issues will heighten the cultural sensitivity of TB care
providers, enhance communication, and improve the overall
effectiveness of organizations and staff in cross-cultural settings.
The content of these guides was gathered in two ways. First, an
in-depth review of TB-related epidemiologic, behavioral, and
ethnographic literature on the cultural group was performed.
Secondly, in 2003, DTBE undertook a qualitative study to describe
ethnographic aspects of the increasing burden of TB among five
foreign-born populations. Selected major findings from the study are
presented in each of the guides.
The Hmong guide will be the first to be released, followed by the
Somali guide. All guides are expected to be in print by late spring
or early summer. Limited hard copies and CD-ROMS will be available.
PDF versions of the guides will be accessible on the DTBE website,
starting with the
—Reported by Robin Shrestha-Kuwahara, MPH
Div of TB Elimination