85 PDF
85 PDF
Abstract
Introduction
14 JANUARY 2006
bmj.com
Methods
We undertook a case-control study in the city of
Samara, 700 miles southeast of Moscow. All participants were residents of the city. We defined cases as all
adults with culture confirmed pulmonary tuberculosis
newly diagnosed at any of the citys specialist
tuberculosis clinics between 1 January 2003 and
31 December 2003, and recruited to a WHO DOTS
(directly observed treatment short course) programme.
We estimated that 307 cases and an equal number of
controls should be recruited to achieve 80% power to
detect an odds ratio of 2.0 at the 5% significance level if
10% or more of the general population were exposed
to the risk factor. Controls were sampled randomly
from the general population of Samara city; they were
matched for year of birth and sex, and they had no history of tuberculosis. A team of 22 trained interviewers
administered a previously piloted questionnaire to
This article was posted on bmj.com on 8 December 2005:
http://bmj.com/cgi/doi/10.1136/bmj.38684.687940.80
European Centre
on Health of
Societies in
Transition,
Department of
Public Health and
Policy, London
School of Hygiene
and Tropical
Medicine, London
WC1E 7HT
Richard Coker
reader
Martin McKee
professor
Boika Dimitrova
research fellow
Tanaka Business
School, Imperial
College London,
London
Rifat Atun
reader
Samara Social
Research Institute,
Samara, Russia
Ekaterina
Dodonova
research fellow
Samara Oblast
Health Department,
Samara
Sergei Kuznetsov
head of department of
adult healthcare
continued over
BMJ 2006;332:857
85
Research
Mycobacterium
Reference Unit,
Institute of Cell and
Molecular Sciences,
Queen Marys
School of Medicine,
London
Francis
Drobniewski
professor
Correspondence to:
R Coker
richard.coker@
lshtm.ac.uk
Odds ratios (95% confidence intervals) for risk factors for tuberculosis in Russia
Univariate
Adjusted for
employment and
financial security
No
1.00
1.00
NA
NA
Yes
Variable
All variables
together
Diabetes:
1.00
7.83 (2.37 to 25.89)
1.00
Yes
1.00
NA
NA
1.00
2.80 (1.47 to 5.36)
1.00
Yes
1.00
NA
1.00
2.75 (1.80 to 4.20)
Assets:
Most
2
1.00
1.00
NA
NA
1.00
1.08 (0.56 to 2.07)
Least
Living space/person:
Most
2
1.00
1.00
NA
NA
1.00
Least
Employed:
Yes
1.00
1.00
1.00
No
NA
1.00
5.84 (3.79 to 9.01)
Shortage of food:
No
1.00
1.00
Yes
Financial security:
Most
1.00
1.00
1.00
NA
1.00
Least
NA
NA
NA
NA
NA
NA
1.00
1.00
1-2
3-10
11-20
>20
1.00
1.00
Yes
1.00
1.00
1.00
1.00
1.00
Yes
1.00
1.00
1.00
1.00
Yes
1.00
1.00
1.00
1.00
1.00
Yes
NA=not appropriate.
Results
We recruited 334 cases and 334 age and sex matched
controls. Two measures of socioeconomic position
were generated (table). The first was a measure of accumulated wealth, developed from responses to questions about the possession of a range of household
assets. The second was a measure of financial security:
86
14 JANUARY 2006
bmj.com
Research
diabetes (2.66) or who were living with a relative with
tuberculosis (2.94).
Although the figures should be interpreted with
considerable caution because of the small numbers
involved, it is possible to calculate estimates of the
population attributable risk by applying the odds ratios
obtained to the prevalence of exposure in the controls
(clearly the controls are not a random sample of the
population, but we could not obtain population based
figures for the variables in question). Calculations
using the univariate odds ratios yield population
attributable risks of 0.8% and 2% for having been in
prison or a pretrial detention centre, respectively. However, because of the much higher prevalence of
exposure to raw milk, the population attributable risk
for this factor was 18%; similarly, that for being unemployed was 28%.
Discussion
Our findings measure the risks associated with a
variety of social factors and tuberculosis. Poverty,
unemployment, drinking unpasteurised milk, diabetes,
living with a relative with tuberculosis, living in
overcrowded conditions, and a prison or detention history were independently associated with an increased
risk of tuberculosis.
Research has shown that a history of imprisonment
is strongly associated with tuberculosis but did not
examine the role of pretrial detention centres,
unpasteurised milk, or diabetes.1 Others have drawn
attention to the role of the criminal justice system, specifically pretrial detention centres, in the epidemic of
tuberculosis in Russia but did not measure this association.2 3 Our study confirms that incarceration is associated with a substantial increase in the risk of
pulmonary tuberculosis. However, the small size of the
population attributable risks associated with the two
forms of incarceration suggest that, contrary to
common belief, imprisonmentbefore trial or after
sentencingdoes not contribute greatly to the overall
burden of tuberculosis in Russia.
The association found with unpasteurised milk
may be linked to Mycobacterium bovis infection. The
dairy industry has been affected by the political transition, and the consumption of unpasteurised milk has
increased.4 5 If the association between drinking raw
milk and tuberculosis proves to be related to M bovis,
ensuring a safe milk supply would be a public health
priority.
Our study has several limitations. Although living
with a relative who had tuberculosis was associated
with a greater risk, recall bias is possible. We did not
investigate the potential role of HIV because of ethical,
political, and practical considerations. Although HIV
infection may be an important but unexplored risk factor, this is a recent phenomenon, and as yet the degree
of immune suppression in infected individuals is not
marked.6
We thank David Leon of the London School of Hygiene and
Tropical Medicine for methodological advice.
Contributors: RC, MM, and BD designed the study; RC, MM,
RA, BD, ED, SK, and FD supervised data collection and analysis;
RC, MM, RA, and FD wrote the report. RC is guarantor.
14 JANUARY 2006
bmj.com
87