761
Diabetes Mellitus in HIV-infected Persons: The Role of CD4 Counts
Adeel Butt*1,2,3,4, D Kelley1, K McGinnis2,3,5, M Skanderson2,3, S Crystal3,6, M Simberkoff3,7, M Rodriguez-Barradas3,8,9, D Leaf3,10,11, M Goetz3,10,11, A Justice3,12,13, and Veterans Aging Cohort Study
1Univ of Pittsburgh Med Ctr, PA, US; 2VA Pittsburgh Hlthcare System, PA, US; 3Veterans Aging Cohort Study; 4Ctr for Hlth Equity Res and Promotion, Pittsburgh, PA, US; 5Univ Ctr for Social and Urban Res, Univ of Pittsburgh, PA, US; 6Rutgers Univ, New Brunswick, NJ, US; 7VA NY Harbor Hlthcare System, US; 8Baylor Coll of Med, Houston, TX, US; 9Michael E DeBakey VAMC, Houston, TX, US; 10David Geffen Sch of Med, Univ of California, Los Angeles, US; 11VA Greater Los Angeles Hlthcare System, CA, US; 12Yale Univ, New Haven, CT, US; and 13VA Hlthcare System, West Haven, CT, US
Background: The influence of HIV upon the risk of
diabetes is controversial, although ART have been associated with increased
risk. We undertook this study to determine and quantify the role of HIV
infection upon the risk of diabetes.
Methods: Subjects were part of the ongoing Veterans
Aging Cohort Study (VACS), which is a prospective study of HIV-infected
veterans and controls at 8 sites in the United States. Data gathered in
VACS includes demographic, clinical, and laboratory data, and patient and provider
surveys. For our study, diabetes was defined as having 1 inpatient or 2
outpatient codes for diabetes, or fasting glucose >126 mg/dL, or a random
blood glucose >200 mg/dL. The prevalence of diabetes at study entry was the
main outcome measure. c2 test or t-test
were used to compare variables between HIV-infected and -uninfected subjects.
Univariable and multivariable logistic regression analyses were used to predict
factors associated with the presence of diabetes.
Results: We studied 2265 HIV-infected subjects and
2162 controls. HIV-infected subjects were younger (mean age 49.2 vs 50.2 years,
p = 0.001), more likely to be male
(97% vs 92%, <0.0001) and black (66% vs 61%, p = 0.002), and to have hepatitis C virus (HCV) infection (30% vs
12.8%, p <0.0001), drug use diagnosis
(34% vs 16%, p <0.0001), and lower
body mass index (25.4 vs 29.1, p <0.0001).
The prevalence of diabetes was 13.6% in the HIV-infected and 22.5% in the HIV-uninfected
subjects (p <0.0001). In
univariable logistic regression analysis, HIV was associated with a lower odds
of having diabetes (OR 0.54, 95%CI 0.46 to 0.53). In multivariable analysis in
the HIV-infected persons, body mass index (OR 1.12 for each unit increase, 95%CI
1.08 to 1.16) and protease inhibitor use ever (OR 2.00, 95%CI 1.22 to 2.23)
were associated with a higher risk, while a CD4 count <150 cells/mm3
(OR 0.44, 95%CI 0.22 to 0.87) was associated with a lower risk of diabetes
after adjusting for age, race, HCV infection, drug and alcohol use, and
undetectable HIV RNA at baseline.
Conclusions: Among
patients receiving medical care, HIV infection is associated with a
lower risk of diabetes, even after adjusting for traditional and novel risk
factors. Lower CD4 cell counts are associated
with a lower risk, while at higher CD4 counts the risk is not different from that
of HIV-uninfected persons. The role of immune status needs to be studied
further to understand this association.
|
