Incidence of Dyslipidemia in People With HIV Who Are Treated With Integrase Inhibitors Versus Other Antiretroviral Agents: The RESPOND Study Group

Dathan M. Byonanebye; Mark N. Polizzotto; Josip Begovac; Katharina Grabmeier-Pfistershammer; Irene Abela; Antonella Castagna; Stéphane DeWit; Cristina Mussini; Jörg J. Vehreschild; Antonella d'A Monforte; Ferdinand W.N.M. Wit; Christian Pradier; Nikoloz Chkhartishvili; Anders Sönnerborg; Jennifer Hoy; Jens Lundgren; Bastian Neesgaard; Loveleen Bansi-matharu; Lauren Greenberg; Josep M. Llibre; Vani Vannappagari; Joel Gallant; Coca Necsoi; Piotr Cichon; Peter Reiss; Inka Aho; Tengiz Tsertsvadze; Marianna Mennozzi; Andri Rauch; Camilla Muccini; Matthew Law; Amanda Mocroft; Lene Ryom; Kathy Petoumenos


AIDS. 2021;35(6):869-882. 

In This Article


Study Design and Participants

The analysis was conducted within RESPOND, a collaboration of 17 observational cohorts with 29 432 PLWH across Europe and Australia.[29] All cohorts collect data on demographics, HIV treatment, comorbidities, clinical events, and laboratory biomarkers of metabolic diseases. Data, including random laboratory-determined LDL, HDL, TC, and TRIGs levels, were collected using standard case report forms via the HIV Cohorts Data Exchange Protocol ( and transmitted to a coordinating center annually. All cohorts ensure data completeness and accuracy by adhering to quality control checks.[29]

Participants were eligible if they were adults (age ≥18 years), and initiated or switched to ART comprising two NRTI and either an INSTI [dolutegravir (DTG), raltegravir (RAL) or cobicistat-boosted elvitegravir (EVG/c)], NNRTI [rilpivirine (RPV) or efavirenz (EFV)], or PI/b [cobicistat or ritonavir-boosted atazanavir (ATV/b) or cobicistat or ritonavir-boosted darunavir (DRV/b)], between 1 January 2012 and 31 December 2018. ART-experienced participants were naive to the ART class they were switched to. The participants had no dyslipidemia (TC > 240 mg/dl and/or HDL < 35 mg/dl, TRIGs > 200 mg/dl, and/or receiving lipid-lowering therapy) at baseline and had lipid test result(s) after the baseline date.

Study Outcomes

The primary outcome was the incidence of dyslipidemia in PLWH treated with INSTI versus NNRTI or PI/b. Dyslipidemia was defined as TC more than 240 mg/dl and/or HDL less than 35 mg/dl, TRIGs more than 200 mg/dl, and/or initiation of lipid-lowering therapy, consistent with prior D:A:D analyses.[18] We conducted an analysis involving all PLWH and separate analyses for those who were ART-naive and ART-experienced. The secondary outcomes were the incidence of dyslipidemia using the thresholds defined in the National Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (ATP III) guidelines[30] (i.e. TC ≥ 240 mg/dl and/or HDL < 40 mg/dl and/or LDL ≥ 160 mg/dl, and/or TRIGs ≥ 200 mg/dl) and the incidence of serious lipid abnormalities (TC ≥ 300 mg/dl, LDL ≥ 190 mg/dl, or TRIGs > 500 mg/dl) as well as any lipid abnormalities (TC ≥ 200 mg/dl, LDL ≥ 130 mg/dl, or TRIGs ≥ 150 mg/dl), both defined by the division of AIDS (DAIDS).[31] The DAIDS-defined events were used to determine the incidence of dyslipidemia grades milder and more severe than the ATP III criteria. We compared the incidence of dyslipidemia in participants taking specific antiretroviral drugs versus DTG as well as incidence in those taking tenofovir alafenamide fumarate (TAF) or abacavir versus tenofovir disoproxil fumarate (TDF). Only drugs with at least 100 events were considered for drug comparisons.

Baseline Covariates

Baseline covariates were within 12 months before and up to 7 days following ART initiation or switch (i.e. baseline date). Where multiple baseline measures existed, the measure closest to the baseline date was considered. We defined hypertension as SBP greater than 140 mmHg and/or DBP more than 90 mmHg and/or use of antihypertensive drugs. Diabetes mellitus was defined as initiation of treatment and/or blood glucose level at least 11.1 mmol/l or hemoglobin A1c at least 6.5% and/or confirmed diagnosis. Participants were hepatitis C virus (HCV) positive if they had a positive antibody test or RNA or had genotype tests or received HCV therapy. Hepatitis B virus (HBV) positivity was defined as a positive HBV surface antigen test and/or a prior HBV DNA test. Chronic kidney disease (CKD) was defined as two successive estimated glomerular filtration rates (Chronic Kidney Disease Epidemiology Collaboration) 60 ml/min per 1.73 m2 or less at least 90 days apart and/or CKD diagnosis. CVD was defined as stroke, and/or acute MI, and/or evidence of the invasive coronary procedure. Lipid-lowering therapy was defined as the use of statins and/or fibrates, consistent with most guidelines that recommend statins and/or fibrates as first-line lipid-lowering therapy.[32]

Statistical Analysis

We conducted univariate followed by multivariate poisson regression analysis to determine the adjusted incidence rate ratio (aIRR) of dyslipidemia in PLWH treated with NNRTI and PI/b versus INSTI. Follow-up was censored after 3 years, or upon ART regimen discontinuation or last lipid test date or 31 December 2019, whichever occurred first.

The covariates considered in the univariate analysis were baseline lipid values, NRTI backbone, age, ethnicity, sex, mode of HIV acquisition, baseline calendar year, smoking status, BMI, hypertension, diabetes mellitus, prior AIDS, CVD, CKD, HBV, HCV, HIV RNA, nadir and baseline CD4+ cell counts, duration since HIV diagnosis, and prior ART exposure. Covariates were considered in the multivariate model if they had a global P less than 0.2 in the univariate analysis. The backward elimination method based on the Akaike information criterion was used to fit the final model. All statistical tests were two-sided with P less than 0.05. Data were prepared using SAS Enterprise Guide 8.2 (SAS Institute Inc., Cary, North Carolina, USA) and analyzed using Stata version 16.0 (Stata Corp., College Station, Texas, USA).

Sensitivity Analysis

We performed several sensitivity analyses, including the incidence of abnormal individual lipids [hypertriglyceridemia (>200 mg/dl), low HDL (<35 mg/dl), or elevated TC (>240 mg/dl)], the incidence of initiation of lipid-lowering therapy, exclusion of cohorts in which at least 20% of participants had missing data on lipid, and exclusion of events within 3 months of the baseline.