Abstract Body

Background: Although tenofovir (TFV) is commonly used in HIV treatment and prevention, pharmacogenomic studies of this agent are limited. Modeling of genetic factors with other determinants of exposure is important. We performed intensive pharmacokinetic (iPK) studies in a diverse cohort of HIV-infected women and identified single nucleotide polymorphisms (SNPs) in genes associated with tenofovir (TFV) metabolism. Models with genetic and non-genetic factors were fit to fully assess contributors to TFV exposure.

Methods: The Women’s Interagency HIV Study (WIHS) is a multicenter, prospective cohort of representative HIV-infected women. Participants on TFV-based therapy underwent 24-hour iPK sampling after a witnessed dose under steady-state conditions. Factors that affect PK were measured, including race, age, ritonavir (RTV) use, baseline kidney function (by estimated GFR via serum creatinine) and weight. Participant areas under the time concentration curve (AUCs) were calculated using the trapezoidal rule. Genetic data were collected on all participants. SNPs with an allele frequency of >0.05 were screened in three genes identified as important in the PK of TFV, specifically the ATP-binding cassette (ABC) B1, ABCC2 and organic anion transporter 1 (OAT1). Each SNP was evaluated as a single addition to an established model with non-genetic factors.

Results: Of 93 participants, 61% were African-American, 60.2% used RTV, and 9.7% had an eGFR <70 mL/min. Median (range) TDF AUC was 3340 (1026-9356) ng × h/ml and BMI 27 (15-62) kg/m2. None of 88 SNPs met our a priori p<0.001 threshold for association with TDF AUC in multivariate modeling; rs4728707 in ABCB1 (CC versus CA; 8.1% allele frequency, no AA) had the smallest observed p-value (1.29 fold effect, 95% CI 1.02, 1.63; p=0.03). Of non-genetic factors, increasing BMI was associated with lower log TDF AUC, while age, ritonavir use and baseline eGFR were associated with higher log TDF AUC (Table). Due to non-normality of residuals, we repeated the model using robust standard errors and rs472870 had CI 1.09 to 1.54 with p=0.004.

Conclusions: A comprehensive evaluation of 88 SNPs spanning three genes associated with TFV exposure assessed in multivariate models with non-genetic factors identified no strong associations, but ABCB1 rs4728707 had an estimated 29% increase in tenofovir AUC in 24 hour iPK monitoring. Genetic evaluations in real-world populations may help define optimal strategies to maximize efficacy and minimize toxicity for treated individuals.

⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥⇥

⇥⇥

⇥

Factor Effect on log AUC (+95% CI) p-value
Age (per 10 years) 1.16 (1.05, 1.29) 0.006
BMI (per 10 percent increase) 0.96 (0.93, 0.99) 0.01
Ritonavir use vs not 1.26 (1.06, 1.49) 0.01
African-American vs other 1.04 (0.87, 1.25) 0.67
eGFR < 70 vs > 70 1.38 (1.00, 1.91) 0.05
rs4728707 (CC versus CA) 1.29 (1.02, 1.63) 0.03