Transcript Diagnosed at a CD4+ T cell count < 200

Increased Mortality in Rural
Patients with HIV in New England
Timothy Lahey1, 2; Michelle Lin1; Bryan Marsh2; Jim Curtin2; Kim Wood2; Betsy Eccles1,2; C. Fordham von Reyn1,2
Dartmouth Medical School1 and Dartmouth-Hitchcock Medical Center2, Lebanon, NH 03756
Background
Results (continued)
The number of HIV infected people in rural areas is
increasing.
Although
patients
with
human
immunodeficiency virus (HIV) infection who live in the
rural United States may have less access to expert care and
antiretroviral treatment, the impact of living in rural areas
on mortality from HIV infection is unstudied.
Methods
We compared mortality rates in 323 rural and 313 urban
patients with HIV infection treated in the Dartmouth
Hitchcock HIV Program in a retrospective cohort study
using a univariate comparisons and students ttests. Variables
that impacted mortality on univariate comparisons, or that
were considered prima facie relevant, were included in a
multivariate logistic regression model. Subjects were also
compared according to socioeconomic strata. A survival
analysis of this data is ongoing.
Results
Results. The characteristics of rural and urban patients with
HIV are displayed in Table 1.
Figure 1. Mortality was higher in rural patients with
HIV when stratified by CD4 count at first
presentation.
Mortality was higher in rural patients (10.4% vs 6.0%,
p=0.028). When stratified by CD4 count, mortality
remained higher in rural patients with HIV (Figure 1).
When stratified by other demographic factors that
themselves correlated with the likelihood of death –
insurance status, sex with men, age, and travel time – rural
patients were still more likely to die than urban patients
(Figure 3).
Table 1. Characteristics of rural and urban patients in
Dartmouth HIV Program, 1995-2005
Rural
Urban p-value
(N=323) (N=313) by t test
Age at end of follow
43.4
41.4
0.002
up, years
Male, %
72.9
70.4
Race, %
White
Black
Asian
Risk behavior, %
Man who has sex with
men
Intravenous drug use
Woman who has sex
with men
Population of town,
mean
Born in United States,
%
Year of diagnosis,
mean
CD4+ T cell count at
diagnosis, cells/mL
Diagnosed at a CD4+
T cell count < 200, %
93.0
6.1
0.3
77.9
20.8
1.3
<0.001
<0.001
0.168
55.5
36.1
<0.001
12.2
12.8
14.8
24.0
0.336
<0.001
9,738
78,721
<0.001
82.4
82.3
0.978
1994
1995
0.001
376
351
0.298
27.8
28.1
0.945
Rural patients with HIV infection were slightly older, more
likely to be white, and a greater proportion were men who
have sex with men. While the mean year of diagnosis was
slightly earlier in rural patients, and the mean CD4 count at
first presentation to our clinic was similar. Rural patients in
our cohort were more likely to receive antiretroviral
medications at any CD4 count (73.7 vs. 62.1%, p=0.002),
and received PCP prophylaxis at comparable rates (23.5%
vs. 25.6%, p=0.555).
The characteristics of rural and urban patients who died in
our cohort between 1995 and 2005 were not different (Table
2).
Table 2. Characteristics of rural and urban
decedents in Dartmouth HIV Program, 1995-2005
Rural
Urban
P-value
(N=35) (N=19) by t
test
Age at the end of follow 45.3
42.1
0.204
up, years
Male, %
73.5
72.2
0.921
Race, %
White 97.1
77.8
0.025
Black 2.9
22.2
0.025
Risk behavior, %
Man who has sex with 39.4
22.2
0.222
men
Intravenous drug use 17.6
11.1
0.544
Woman who has sex with 14.7
22.2
0.505
men
Born in the US
81.5
72.7
0.561
Year of diagnosis, mean
Mean last CD4 count
(cells/mL)
1994
194
1993
178
0.553
0.779
Received antiretroviral
treatment, %
79.4
72.2
0.567
Figure 2. Mortality was higher in rural patients with
HIV when stratified by other demographic
characteristics.
In the simple logistic model, the odds of death was higher
inrural patients (OR 1.85, p=0.044). In the saturated model,
the pattern remained (OR 1.85, p=0.079). In the
parsimonious model, involving insurance status, sex with
men, and age, rural patients were more likely to die (OR
2.02, p=0.033).
The risk of mortality remained higher in rural patients when
adjusting for age, sex, race, HIV risk factors, year of
diagnosis, travel time, lack of insurance, and receipt of
antiretroviral treatment or PCP prophylaxis in a logistic
regression model (OR 2.11, 1.064 to 4.218, p=0.047).
The causes of death were similar in rural and urban patients
with HIV in our cohort, with infection and liver disease
being most common in both groups (Table 2).
Table 3. Cause of death in rural and urban
decedents in Dartmouth HIV Program, 1995-2005
Rural
Urban P-value
(n=34) (n=18) by t test
Any infection
41.2
33.4
0.350
Opportunistic infection 32.4
27.8
0.740
Other infection 8.8
5.6
0.681
Liver disease
23.5
11.1
0.289
Cancer
14.7
11.1
0.724
Medical, not infectious 8.8
27.8
0.074
Trauma and suicide
8.8
11.1
0.795
Unknown
3.0
5.5
0.649
Conclusions
Among patients treated in the same regional HIV program,
rural patients have higher mortality rates than urban patients
even when adjusting for demographic characteristics like
age, race, sex and HIV risk factors. As the number of HIV
infected patients in rural areas increases, it will be important
to understand the factors that contribute to increased
mortality in this population.
Limitations / Future Directions
There is no major metropolitan center in our study area. We
believe, however, that this would lead to underestimation of
the differences in rural vs. urban outcomes. Further, rural
patients in our cohort were more likely to be white men
than their urban peers. Although this is the opposite of what
has been seen in other areas of the country, the differential
outcome in rural and urban patients with HIV remained
after incorporating such demographic characteristics into
the multivariate model. Further, as white men generally
have better health outcomes, we suspect other factors may
contribute to rural HIV mortality more significantly.
Importantly, we are now conducting a survival analysis of
this data.