Efavirenz Pharmacogenetics and Weight Gain Following Switch to Integrase Inhibitor-Containing Regimens

Abstract

Background: Unwanted weight gain affects some people living with human immunodeficiency virus (HIV) who are prescribed integrase strand transfer inhibitors (INSTIs). Mechanisms and risk factors are incompletely understood.

Methods: We utilized 2 cohorts to study pharmacogenetics of weight gain following switch from efavirenz- to INSTI-based regimens. In an observational cohort, we studied weight gain at 48 weeks following switch from efavirenz- to INSTI-based regimens among patients who had been virologically suppressed for at least 2 years at a clinic in the United States. Associations were characterized with CYP2B6 and UGT1A1 genotypes that affect efavirenz and INSTI metabolism, respectively. In a clinical trials cohort, we studied weight gain at 48 weeks among treatment-naive participants who were randomized to receive efavirenz-containing regimens in AIDS Clinical Trials Group studies A5095, A5142, and A5202 and did not receive INSTIs.

Results: In the observational cohort (n = 61), CYP2B6 slow metabolizers had greater weight gain after switch (P = .01). This was seen following switch to elvitegravir or raltegravir, but not dolutegravir. UGT1A1 genotype was not associated with weight gain. In the clinical trials cohort (n = 462), CYP2B6 slow metabolizers had lesser weight gain at week 48 among participants receiving efavirenz with tenofovir disoproxil fumarate (P = .001), but not those receiving efavirenz with abacavir (P = .65). Findings were consistent when stratified by race/ethnicity and by sex.

Conclusions: Among patients who switched from efavirenz- to INSTI-based therapy, CYP2B6 genotype was associated with weight gain, possibly reflecting withdrawal of the inhibitory effect of higher efavirenz concentrations on weight gain. The difference by concomitant nucleoside analogue is unexplained.

Trial registration: ClinicalTrials.gov NCT00013520 NCT00050895 NCT00118898.

Keywords: HIV-1; efavirenz; integrase strand transfer inhibitors; pharmacogenetics; weight gain.

Figure 1.

Derivation of the 2 study cohorts. A, Derivation of the observational cohort. B, Derivation of the clinical trials cohort. Abbreviations: 3TC, lamivudine; ABC, abacavir; ACTG, AIDS Clinical Trials Group; ART, antiretroviral therapy; EFV, efavirenz; FTC, emtricitabine; INSTI, integrase strand transfer inhibitor; TDF, tenofovir disoproxil fumarate; VCCC, Vanderbilt Comprehensive Care Clinic; VL, viral load.

Figure 2.

Associations between genotype and change in weight at week 48 in an observational cohort of human immunodeficiency virus–positive individuals who switched from efavirenz-containing antiretroviral therapy to integrase strand transfer inhibitor–containing regimens, stratified by race. A, Associations between CYP2B6 metabolizer genotype (stratified into 3 genotype levels) and weight change at week 48. B, Associations between UGT1A1 genotype and weight change at week 48. The C/C, C/T, and T/T genotypes define UGT1A1 normal, intermediate, and slow metabolizers, respectively. C, Associations between CYP2B6 metabolizer genotype and weight change at week 48 among white participants. D, Associations between CYP2B6 metabolizer genotype and weight change at week 48 among black participants. Error bars indicate the median and interquartile range. Linear regression model P values are shown. Numbers above the x-axis represent numbers of evaluable participants for each genotype. Normal and slow metabolizers may also be referred to as extensive and poor metabolizers, respectively.

Figure 3.

Associations between genotype and change in weight at week 48 in an observational cohort of human immunodeficiency virus (HIV)–positive individuals who switched from efavirenz-containing antiretroviral therapy (ART) to integrase strand transfer inhibitor (INSTI)–containing regimens, stratified by INSTI. A, Associations between CYP2B6 metabolizer genotype and weight change at week 48 among participants who switched to dolutegravir-containing ART. B, Associations between CYP2B6 metabolizer genotype and weight change at week 48 among participants who switched to elvitegravir-containing ART. C, Associations between CYP2B6 metabolizer genotype and weight change at week 48 among participants who switched to raltegravir-containing ART. Error bars indicate median and interquartile range. Linear regression model P values are shown. Numbers above the x-axis represent numbers of evaluable participants for each metabolizer level. Normal and slow metabolizers may also be referred to as extensive and poor metabolizers, respectively.

Figure 4.

Associations between CYP2B6 metabolizer genotype and change in weight at week 48 in a clinical trials cohort of human immunodeficiency virus–positive individuals who were randomized to initiate efavirenz-containing regimens in prospective AIDS Clinical Trials Group (ACTG) studies. Left 2 panels: Associations among all evaluable participants from ACTG protocols A5095, A5142, and A5202, stratified by concomitant nucleos(t)ide reverse transcriptase inhibitor (NRTI). Right 2 panels: Associations between CYP2B6 metabolizer genotype and weight change at week 48 just among evaluable participants from A5202, stratified by concomitant NRTI. Blue markers indicate normal metabolizers; green markers indicate intermediate metabolizers, and red markers indicate slow metabolizers. Error bars indicate median and interquartile range. P values are from 2-sided univariate Jonckheere-Terpstra test for ordered alternatives. Normal and slow metabolizers may also be referred to as extensive and poor metabolizers, respectively. Abbreviations: ABC, abacavir; TDF, tenofovir disoproxil fumarate.

Figure 5.

Forest plots of multivariable associations between CYP2B6 metabolizer genotype and change in weight at week 48 in a clinical trials cohort of human immunodeficiency virus–positive individuals who were randomized to initiate efavirenz-containing regimens in prospective AIDS Clinical Trials Group (ACTG) studies. For these analyses, CYP2B6 genotype was categorized into 10 ordinal levels that predict progressively greater plasma efavirenz exposure. Linear regression model results are shown. For “all participants” and nucleos(t)ide reverse transcriptase inhibitor (NRTI) analyses, covariates included weight, age, race/ethnicity, and sex at baseline, as well as the first 3 principal components generated from genome-wide genotype data. For race/ethnicity analyses, the same covariates as for all participants were included except principal components. For sex analyses, the same covariates as for all participants were included except sex. β coefficients and 95% confidence intervals (CIs) are shown. Negative β coefficients indicate that CYP2B6 genotypes that predict increased plasma efavirenz concentrations were associated with less weight gain (or greater weight loss). The tenofovir disoproxil fumarate (TDF) arms are shown in blue; abacavir (ABC) arms are shown in red. A, Associations among all evaluable participants from ACTG protocols A5095, A5142, and A5202, stratified by concomitant NRTI, self-reported race/ethnicity, and sex. B, Associations among all evaluable participants from A5202, stratified by concomitant NRTI, self-reported race/ethnicity, and sex. C, Associations among all evaluable participants from A5095 and A5142, stratified by concomitant NRTI, self-reported race/ethnicity, and sex.