Title: Antiretroviral Therapy Exposure and Insulin Resistance in the Women
1Antiretroviral Therapy Exposure and Insulin
Resistance in the Womens Interagency HIV Study
Phyllis C. Tien MD1, Michael F. Schneider MS2,
Stephen R. Cole PhD2, Alexandra M. Levine MD3,
Mardge Cohen MD4, Jack DeHovitz MD5, Mary Young
MD6, Jessica E. Justman MD7
Corresponding Author Phyllis Tien, MD SFVAMC and
UCSF 4150 Clement St San Francisco, CA
94121 415-221-4810 x2577 ptien_at_ucsf.edu
933
1Department of Medicine, University of
California, San Francisco, and San Francisco
Veterans Affairs Medical Center, San Francisco,
CA 2Department of Epidemiology, Johns Hopkins
Bloomberg School of Public Health, Baltimore, MD
3Department of Medicine, University of Southern
California, Los Angeles, CA 4CORE Center/Stroger
Hospital of Cook County, Chicago, IL 5Department
of Medicine, State University of New
York-Downstate Medical Center, Brooklyn, NY
6Department of Medicine, Georgetown University
Medical Center, Washington, DC 7Departments of
Epidemiology and Medicine, Columbia University,
New York, NY
Results
Abstract
Table 1. At the visit immediately prior to the
index visit, HIV-infected women were older, more
likely to be postmenopausal, and more likely to
be anti-HCV than the HIV-uninfected women.
HIV-infected women had lower BMI and hip
circumference.
Table 3. Longer cumulative exposure to NRTIs was
associated with greater IR in HIV-infected women.
Background Insulin resistance (IR) has been of
increased concern since the introduction of
effective antiretroviral therapy (ART), but the
associated factors remain ill-defined. We
determine the relationship of the type and
duration of ART exposure with IR in a nationally
representative sample of HIV and at risk HIV-
women. Methods Prospective cohort study between
October 2000 and March 2007 of 1614 HIV and 604
HIV- participants from the Womens Interagency
HIV Study. The Homeostasis Model Assessment
(HOMA) defined as fasting insulin (µIU/ml) x
fasting glucose (FG) (mg/dL)/405 estimated IR at
11,019 semiannual visits. The first visit with
both FG and insulin data available was considered
the index visit. Results HIV women had higher
median HOMA (2.19 vs 1.83, plt0.001) at index.
After adjustment, HIV women had larger median
HOMA than HIV- women 1.20(95CI 1.11,1.30)
times larger for those reporting protease
inhibitor (PI)-containing highly active
antiretroviral therapy (HAART) since the prior
visit 1.10(95CI1.01,1.20) times larger for
those reporting non-PI-containing HAART since the
prior visit. Those reporting mono/combination
ART and no ART since the prior visit had larger
HOMA than HIV- women, but the association did not
reach statistical significance 1.05(95CI0.93,1.
19) and 1.05(95CI0.96,1.15), respectively.
Other factors associated with larger HOMA in
multivariable analysis included BMI (per 5 unit
increase) 1.18(95CI1.16,1.20), HCV
seropositivity 1.17(95CI1.09,1.24), Hispanic
ethnicity compared to African-American
1.13(95CI1.06,1.19), menopause
1.11(95CI1.02,1.21), family history of DM
1.05(95CI1.00,1.11), and CD4 count (per 100
cell increase) 1.01(95CI1.00,1.02). Among
HIV women, cumulative exposure to nucleoside
reverse transcriptase inhibitors (NRTI) of gt0 to
3 years or gt3 years was associated with median
HOMA 1.06(95CI 0.98, 1.16) and 1.13(95CI
1.02, 1.25) times larger, respectively than the
HOMA without any cumulative NRTI exposure.
Cumulative exposure to stavudine of gt0 to 1 year
or gt1 year was associated with HOMA 1.07(95CI
0.98, 1.16) and 1.15(95 CI 1.05, 1.27) times
larger, respectively, than the HOMA without any
cumulative stavudine use. Conclusion Longer
cumulative exposure to NRTI in particular
stavudine was associated with greater insulin
resistance in HIV women. The role of insulin
resistance and cardiovascular disease in HIV
individuals remains a paramount clinical concern.
Estimated Relative Difference in Medians (95 CI) Estimated Relative Difference in Medians (95 CI)
No. person-visits Median HOMA Unadjusted Analyses Adjusted Analysis
Cumulative number of NRTI-years Cumulative number of NRTI-years
0 1,910 2.07 1 (reference) 1 (reference)
gt0 to 3.0 3,039 2.25 1.06 (0.99, 1.13) 1.06 (0.98, 1.16)
gt3.0 3,032 2.22 1.09 (1.01, 1.19) 1.13 (1.02, 1.25)
Cumulative number of PI-years Cumulative number of PI-years
0 4,309 2.15 1 (reference) 1 (reference)
gt0 to 1.5 1,837 2.22 1.07 (1.01, 1.14) 1.03 (0.96, 1.10)
gt1.5 1,835 2.25 1.06 (0.98, 1.13) 1.01 (0.93, 1.10)
Cumulative number of NNRTI-years Cumulative number of NNRTI-years
0 4,586 2.21 1 (reference) 1 (reference)
gt0 to 1.0 1,699 2.17 0.99 (0.93, 1.06) 0.96 (0.89, 1.03)
gt1.0 1,696 2.20 1.00 (0.93, 1.08) 0.95 (0.86, 1.03)
Characteristic HIV-infected women N 1,614 HIV-uninfected women N 604 P-value
Age, years 38.8 (33.0, 44.8) 33.9 (27.0, 41.7) lt0.001
Race, (n) 0.69
African-American 56.2 (907) 58.1 (351)
Hispanic 27.9 (451) 27.2 (164)
Caucasian 15.9 (256) 14.7 (89)
Current smoker, (n) 47.6 (766) 51.8 (313) 0.07
Body Mass Index, kg/m2 26.6 (23.3, 31.1) 27.7 (23.4, 32.9) lt0.01
Waist size, cm 87.3 (79.2, 98.2) 86.3 (77.6, 99.0) 0.39
Hip size, cm 99.0 (91.3, 107.8) 102.0 (94.9, 112.2) lt0.001
Menopausal, (n) 15.2 (245) 8.1 (49) lt0.001
HCV antibody positive, (n) 29.4 (465) 15.6 (93) lt0.001
Family history of diabetes, (n) 29.5 (454) 27.0 (156) 0.27
Clinical AIDS, (n) 31.9 (515) NA NA
log10 HIV RNA, copies/ml 2.93 (1.90, 4.04) NA NA
CD4 count, cells/mm3 431 (270, 633) NA NA
Nadir CD4 count, cells/mm3 257 (130, 389) NA NA
Antiretroviral therapy naïve, (n) 15.6 (252) NA NA
1,614 HIV-infected women contributed a total of
7,981 person-visits of follow-up including the
index visit anti-log of coefficient of exposure
group from regression model adjusted for
cumulative number of NRTI, PI, or NNRTI reported
from WIHS baseline through the visit prior to
index, WIHS cohort status (1994-1995 or
2001-2002), race, baseline HCV antibody status,
family history of diabetes, and age, smoking
status, BMI, CD4 count, menopause assessed at
visit immediately prior to index if data for a
particular variable was missing at visit
immediately prior to index then data concurrent
with index visit was used 1,482 women
contributing 7,529 person-visits were included in
analysis
Table 4. Longer cumulative exposure to stavudine
was associated with greater IR in HIV-infected
women
Estimated Relative Difference in Medians (95 CI) Estimated Relative Difference in Medians (95 CI)
No. person-visits Median HOMA Unadjusted Analyses Adjusted Analysis
Cumulative number of person-years reporting Lamivudine Cumulative number of person-years reporting Lamivudine Cumulative number of person-years reporting Lamivudine
0 2,915 2.11 1 (reference) 1 (reference)
gt0 to 1.1 2,539 2.20 1.04 (0.98, 1.11) 1.04 (0.96, 1.12)
gt1.1 2,527 2.34 1.10 (1.02, 1.18) 1.06 (0.97, 1.17)
Cumulative number of person-years reporting Zidovudine Cumulative number of person-years reporting Zidovudine Cumulative number of person-years reporting Zidovudine
0 4,508 2.17 1 (reference) 1 (reference)
gt0 to 1.0 1,717 2.15 1.00 (0.94, 1.07) 1.03 (0.95, 1.11)
gt1.0 1,756 2.34 1.06 (0.99, 1.14) 1.08 (0.98, 1.19)
Cumulative number of person-years reporting Stavudine Cumulative number of person-years reporting Stavudine Cumulative number of person-years reporting Stavudine
0 5,622 2.15 1 (reference) 1 (reference)
gt0 to 1.0 1,179 2.25 1.06 (0.98, 1.14) 1.07 (0.98, 1.16)
gt1.0 1,180 2.41 1.15 (1.05, 1.25) 1.15 (1.05, 1.27)
Cumulative number of person-years reporting Abacavir Cumulative number of person-years reporting Abacavir Cumulative number of person-years reporting Abacavir
0 5,772 2.20 1 (reference) 1 (reference)
gt0 to 1.0 1,099 2.15 0.97 (0.91, 1.05) 0.96 (0.89, 1.03)
gt1.0 1,110 2.19 1.00 (0.92, 1.09) 0.99 (0.90, 1.07)
Cumulative number of person-years reporting Tenofovir Cumulative number of person-years reporting Tenofovir Cumulative number of person-years reporting Tenofovir
0 5,904 2.23 1 (reference) 1 (reference)
gt0 to 1.0 1,043 2.17 0.97 (0.91, 1.04) 1.00 (0.94, 1.07)
gt1.0 1,034 2.05 0.93 (0.86, 1.01) 0.99 (0.91, 1.07)
Introduction
all values are median (inter-quartile range)
unless otherwise noted if data were missing at
visit immediately prior to index then data
concurrent with index visit was used
antiretroviral therapy naiveté was determined
from WIHS baseline through the index visit
Includes 3 (n54) and 4 (n25) Asian, Pacific
Islander, Native American, Alaskan, other among
HIV-infected and HIV-uninfected women
respectively Current smoking was missing for 3
HIV-infected women body mass index was missing
for 12 HIV-infected and 3 HIV-uninfected waist
size was missing 215 HIV-infected and 79
HIV-uninfected hip size was missing for 216
HIV-infected and 80 HIV-uninfected Menopausal
status was missing for 4 HIV-infected and 1
HIV-uninfected HCV infection status was missing
for 33 HIV-infected and 9 HIV-uninfected family
history of diabetes was missing for 73
HIV-infected and 27 HIV-uninfected (women who
reported not being familiar with health history
of family members were classified as not having a
family history of diabetes) HIV RNA was missing
for 6 HIV-infected CD4 was missing for 11
HIV-infected Nadir CD4 was missing for 92
HIV-infected women P-value from Pearson ?2 test
of overall association when percentages are
compared, P-value from Wilcoxon rank sum test
when medians are compared
- Insulin resistance (IR) has been increasingly
recognized in HIV-infected individuals since the
introduction of effective antiretroviral therapy
(ART). - The protease inhibitor (PI) class of ART has most
often been studied, because early reports
suggested an association between PIs and
disorders of glucose metabolism. - Some more recent studies show an association of
the nucleoside reverse transcriptase inhibitor
(NRTI) class with IR. - We examine the association of both the type and
duration of ART exposure, as well as
non-HIV-related factors with IR using the
Homeostasis Model Assessment (HOMA) from October
2000 to March 2007 among participants from the
Womens Interagency HIV Study (WIHS).
Table 2. HIV-infected women reporting any use of
HAART at their last visit had greater IR than
HIV-infected women as expected, family history
of DM, higher BMI, and HCV seropositivity was
associated with greater IR, as was a higher CD4
count .
Estimated Relative Difference in Medians (95 CI) Estimated Relative Difference in Medians (95 CI)
No. person-years No. person-years Median HOMA Unadjusted Analyses Adjusted Analysis
HIV-uninfected women HIV-uninfected women 1,598 1.91 1 (reference) 1 (reference)
HIV-infected women No therapy HIV-infected women No therapy 1,325 2.11 1.11 (1.03, 1.19) 1.05 (0.96, 1.15)
Mono / combination Mono / combination 199 2.07 1.07 (0.94, 1.21) 1.05 (0.93, 1.19)
HAART with PI HAART with PI 1,491 2.30 1.21 (1.13, 1.30) 1.20 (1.11, 1.30)
HAART no PI HAART no PI 1,215 2.18 1.13 (1.05, 1.22) 1.10 (1.01, 1.20)
Race Race
African-American African-American 3,268 2.11 1 (reference) 1 (reference)
Hispanic Hispanic 1,710 2.25 1.08 (1.02, 1.15) 1.13 (1.06, 1.19)
Caucasian Caucasian 850 1.91 0.95 (0.87, 1.02) 1.00 (0.93, 1.08)
Hepatitis C Hepatitis C
Negative Negative 4,296 2.05 1 (reference) 1 (reference)
Positive Positive 1,423 2.39 1.16 (1.10, 1.24) 1.17 (1.09, 1.24)
Family History of Diabetes Family History of Diabetes
No No 3,934 2.05 1 (reference) 1 (reference)
Yes Yes 1,647 2.28 1.11 (1.04, 1.17) 1.05 (1.00, 1.11)
Age (per 10 years) Age (per 10 years) -- -- 1.06 (1.03, 1.09) 1.01 (0.98, 1.05)
Smoking Smoking
No No 3,051 2.17 1 (reference) 1 (reference)
Yes Yes 2,773 2.05 0.96 (0.91, 1.02) 0.98 (0.93, 1.03)
BMI (per 5 units) BMI (per 5 units) -- -- 1.17 (1.15, 1.19) 1.18 (1.16, 1.20)
CD4 count (per 100 cells) CD4 count (per 100 cells) -- -- 1.01 (1.00, 1.02) 1.01 (1.00, 1.02)
Reported Menopause Reported Menopause
No No 5,112 2.09 1 (reference) 1 (reference)
Yes Yes 699 2.46 1.19 (1.10, 1.28) 1.11 (1.02, 1.21)
Methods
- The WIHS is a multicenter prospective cohort
study established in 1994 to investigate the
progression of HIV in women with and at risk for
HIV. - A total of 3,766 women (2,791 HIV-infected and
975 HIV-uninfected) were enrolled in either
1994-95 (n2,623) or 2001-02 (n1,143) from 6 US
cities (Bronx, Brooklyn, Chicago, Los Angeles,
San Francisco and Washington DC) (1,2). - Beginning in October 2000, fasting (gt8 hours)
glucose (FG) and insulin levels were measured - Of the 2,870 women with a visit between October
2000 and March 2007, 2,583 had at least one FG
and insulin measurement at the same visit the
first visit with both FG and insulin data
available will be referred to as the index visit.
Of the 2,583 women with an index visit, 68 were
excluded due to either a positive or missing
report of pregnancy at the index visit 284 due
to prevalent DM (a FG gt 126 mg/dL or
self-reported DM or self-reported anti-diabetic
medication use all at or prior to the index
visit) and 13 HIV-uninfected at WIHS study entry
due to seroconversion during follow-up.
Therefore, 2,218 women (1,614 HIV-infected and
604 HIV-uninfected), who contributed a total of
11,019 visits with FG and insulin data were
included in the analysis. - IR was quantified using the HOMA, defined as
(insulin (µU/mL) ? glucose (mg/dL)) / 405. (3) - At each visit from index through the end of
follow-up, participants were categorized into one
of five groups based on HIV status and
self-reported ART since the previous visit (a)
HIV-uninfected (b) HIV-infected with no ART (c)
HIV-infected with monotherapy or combination
therapy (d) HIV-infected with PI-based HAART or
(e) HIV-infected with non-PI based HAART. - Drug-years of exposure to the NRTI, PI, and NNRTI
was also determined. - Comparisons of continuous and categorical
characteristics among HIV-infected and
HIV-uninfected women at the visit immediately
prior to the index visit were made by Wilcoxon
rank sum tests and Pearson ?2 tests,
respectively. Linear regression models with
generalized estimating equations were used to
quantify the association of HIV infection and ART
exposures with IR while accounting for the
statistical dependence incurred by repeated HOMA
assessments over time on the same individual.
1,614 HIV-infected women contributed a total of
7,981 person-visits of follow-up including the
index visit anti-log of coefficient of exposure
group from regression model
Conclusions
- We conclude that recent use of any HAART regimen
was associated with higher IR when compared to
HIV-uninfected women. - Among HIV-infected women, cumulative exposure to
NRTI (in particular stavudine), but not PI or
NNRTI, was independently associated with higher
IR. - Study of the biologic mechanisms by which
exposure to individual ART and classes of ART
might induce disorders in glucose metabolism is
needed. - In addition to antiretroviral drug effects,
non-HIV-related factors such as higher BMI,
family history of DM, and HCV seropositivity must
also be considered in the development of IR among
HIV-infected persons.
References
2,128 women (1,614 HIV-infected and 604
HIV-uninfected) contributed a total of 11,017
person-visits (7,981 HIV-infected and 3,036
HIV-uninfected) of follow-up including the index
visit data on demographic and clinical factors
assessed at visit immediately prior to index if
data for a particular variable was missing at
visit immediately prior to index then data
concurrent with index visit was used anti-log
of coefficient of exposure group from regression
model 2,047 women (1,482 HIV-infected and 565
HIV-uninfected) contributing 10,436 person-visits
(7,529 HIV-infected and 2,907 HIV-uninfected)
were included in adjusted analysis Includes
Asian, Pacific Islander, Native American,
Alaskan, and other only assessed among
HIV-infected women in unadjusted analysis and
assessed via an interaction with HIV status in
adjusted analysis (i.e., set to zero for
HIV-uninfected women)
1. Bacon MC, von Wyl V, Alden C, Sharp G, Robison
E, Hessol N, et al. The Women's Interagency HIV
Study an observational cohort brings clinical
sciences to the bench. Clin Diagn Lab Immunol.
200512(9)1013-9. 2. Barkan SE, Melnick SL,
Preston-Martin S, Weber K, Kalish LA, Miotti P,
et al. The Women's Interagency HIV Study. WIHS
Collaborative Study Group. Epidemiology.
19989(2)117-25. 3. Matthews DR, Hosker JP,
Rudenski AS, Naylor BA, Treacher DF, Turner RC.
Homeostasis model assessment insulin resistance
and beta-cell function from fasting plasma
glucose and insulin concentrations in man.
Diabetologia. 198528(7)412-9.