Transcript Slides - View the full AIDS 2016 programme

When Genes and ART Collide:
Modifying Effect of ARVs on
Genetic Predisposition to Diabetes
Melissa A. Frasco, Ph.D.
Research Epidemiologist
#AIDS2016 | @AIDS_conference
Agenda
Background on diabetes in the HIV population
Interaction between diabetes genetic markers and ART
Application to treatment guidelines
#AIDS2016 | @AIDS_conference
Agenda
Background on diabetes in the HIV population
Interaction between diabetes genetic markers and ART
Application to treatment guidelines
#AIDS2016 | @AIDS_conference
Risk Factors for Diabetes in HIV
• Untreated HIV infection may lead to
– dyslipidemia
– lipodystrophy
– insulin resistance
– metabolic syndrome
• Antiretroviral drugs (ARVs) appear to be
implicated in risk of type 2 diabetes mellitus
(DM)
#AIDS2016 | @AIDS_conference
ART-Associated Diabetes
• Two out of three drugs classes used in combination
ART increase risk of DM
• Protease inhibitors (PI)
– Confer acute metabolic risks
– Metabolic effect varies by individual drug
• Nucleoside reverse transcriptase inhibitors (NRTI)
– Confer cumulative risk of diabetes
– Stauvudine, zidovudine and didanoside
#AIDS2016 | @AIDS_conference
Genetic Predisposition to Diabetes
• Genome-Wide Association Studies (GWAS)
have identified genetic markers for DM
– European, Asian, and Hispanic populations
• Replication of genetic markers have been
challenging in African ancestral populations
– DM incidence higher in African ancestry
#AIDS2016 | @AIDS_conference
Agenda
Background on diabetes in the HIV population
Interaction between diabetes genetic markers and ART
Application to treatment guidelines
#AIDS2016 | @AIDS_conference
Objective
To evaluate the association of confirmed DM
genetic markers in HIV-infected women with
consideration for the effect of ART regimen
#AIDS2016 | @AIDS_conference
Methods
• Study population: Women’s Interagency HIV Study
(WIHS) participants with fasting glucose & HgbA1C
measurements
• Study design
– prospective cohort study with visits every 6 months
– followed for DM incidence or end of study (50 DM / 2,957 PY)
– maximum follow-up was 10.4 years
• Exposure: 17 DM genetic markers
• Modifying factor: type of ART regimen
#AIDS2016 | @AIDS_conference
Confirmed DM Genetic Markers
Gene
Polymorphism
CDKAL1
rs7754840
C
Meta-OR
European
GWAS*
1.16
CDKN2A/B
rs564398
T
1.18
5.37 x 10-21
CDKN2A/B
rs10811661
T
1.19
4.83 x 10-16
FTO
rs8050136
A
1.06
1.76 x 10-4
HHEX
rs7923837
G
1.12
2.68 x 10-14
IGF2BP2
rs1470579
C
1.14
9.19 x 10-16
IGF2BP2
rs4402960
T
--
JAZF1
rs864745
T
1.11
KCNQ1
rs2237895
C
--
Risk
Allele
p-value
4.58 x 10-19
2.22 x 10-9
*Saxena, 2012
#AIDS2016 | @AIDS_conference
Results in Non-African Americans
Hazard Ratios for DM Genetic Associations
by Regimen
3.5
3
Hazard ratio
2.5
2
1.5
1
0.5
0
rs7754840
rs564398
rs10811661
No treatment
rs8050136
2 NRTI + NNRTI
rs7923837
rs1470579
≥3 NRTI (non-PI)
rs4402960
rs864745
rs2237895
≥2 NRTI + ≥1 PI
Cox regression adjusted for time-dependent BMI category, age, and genetic ancestry principal components,
using calendar year as the time scale to control for drug utilization trends
#AIDS2016 | @AIDS_conference
Results in non-African Americans
ART regimens
All Subjects
50 DM / 2,957
PY
SNP
HR
95%CI
No Treatment
2 NRTI + NNRTI† ≥3 NRTI (non-PI) ≥2 NRTI + ≥1 PI
†
11 DM / 779 PY
6 DM / 845 PY
HR
HR
95% CI
95% CI
7 DM / 250 PY
HR
95% CI
26 DM / 1,083 PY
HR
95% CI
rs7754840
1.21 (0.80-1.84) 0.94 (0.46-1.93)
0.36
(0.09-1.40) 1.29
(0.44-3.79)
1.77 (1.12-2.80)
rs564398
1.25 (0.77-2.04) 0.88 (0.47-1.65)
0.76
(0.37-1.53) 1.39
(0.70-2.76)
1.69 (1.02-2.79)
rs10811661 1.72 (0.89-3.34) 1.35 (0.65-2.80)
1.11
(0.50-2.49) 2.07
(0.95-4.50)
2.09 (1.07-4.09)
rs8050136
1.24 (0.83-1.84) 0.88 (0.45-1.75)
0.68
(0.29-1.61) 2.14 (1.04-4.39)
1.50
(0.96-2.32)
rs7923837
1.21 (0.78-1.89) 1.03 (0.59-1.81)
0.76
(0.38-1.52) 1.70
1.50
(0.94-2.40)
rs1470579
1.84 (1.19-2.86) 1.00 (0.45-2.20)
1.02
(0.45-2.32) 3.05 (1.47-6.35) 2.66 (1.65-4.31)
rs4402960
1.85 (1.21-2.85) 1.04 (0.47-2.34)
1.11
(0.49-2.48) 2.94 (1.44-6.01) 2.63 (1.61-4.30)
rs864745
1.09 (0.70-1.69) 0.92 (0.51-1.66)
0.61
(0.28-1.31) 1.37
(0.66-2.82)
1.39
(0.87-2.24)
rs2237895
0.93 (0.62-1.39) 0.80 (0.41-1.57)
0.37
(0.13-1.07) 1.04
(0.43-2.52)
1.25
(0.80-1.94)
(0.88-3.29)
Cox regression adjusted for time-dependent BMI category, age, and genetic ancestry principal components,
using calendar year as the time scale to control for drug utilization trends
#AIDS2016 | @AIDS_conference
Results Summary
• The risk of DM is substantial in genetically predisposed individuals who
are taking PI-based regimens or NRTI-heavy regimens
• 10 of the 17 genetic markers tested showed statistically significant
heterogeneity in genetic-DM associations across ART regimens
• No associations were observed in African American women with these
DM genetic markers
• IGF2BP2 rs1470579 was the top association

HR per allele = 2.66 (95% CI 1.65-4.31)
in women taking ≥2 NRTIs + ≥1 PI regimens

HR per allele = 3.05 (95% CI 1.47-6.35)
in women taking ≥3 NRTIs (no PI) regimens

phet=2.92 x 10-3 across ART regimens
#AIDS2016 | @AIDS_conference
Conclusion
• Preferential treatment with non-PI regimens
for HIV-infected persons carrying DM genetic
markers
• Studies needed to replicate the interaction
between ART and DM genetic markers found
in African ancestral populations
#AIDS2016 | @AIDS_conference
Agenda
Background on diabetes in the HIV population
Interaction between diabetes genetic markers and ART
Application to treatment guidelines
#AIDS2016 | @AIDS_conference
Application to Treatment Guidelines
 1st line ART = 2 NRTIs + NNRTI
• 2nd and 3rd line include PI + ≥2 NRTI
– risk for development of DM is amplified in
populations predisposed to diabetes
– implement routine screening for DM
#AIDS2016 | @AIDS_conference