Trial design
AVERTAS was a randomized controlled parallel open-label phase 4 trial comparing weight change in PWH switching from dolutegravir, ABC and lamivudine (DTG/ABC/3TC) to dolutegravir and lamivudine (DTG/3TC) vs. continuing DTG/ABC/3TC and followed up after 24 and 48 weeks (Supplementary Fig. 1). The trial protocol has been published elsewhere [10]. The trial was reported according to the CONSORT statement [11].
Owing to slowing accrual, a decline in eligible participants and fewer study sites than planned, the sample size was re-estimated by allowing a statistical power of at least 80%, corresponding to 66 participants. Originally, the power was set to 90%, corresponding to 90 participants. This decision was made by the sponsor. Other minor amendments to the study protocol occurred during the study period. The Ethics Committee of the Capital Region and the Danish Medicines Agency approved all amendments. Please refer to Supplementary Table 1 for protocol amendments.
Participants
Participants were eligible if they were 18 years or older with HIV-1 infection and had been treated with DTG/ABC/3TC for at least 6 months. The female participants had to agree to contraception during the study. The exclusion criteria were preexisting resistance to 3TC or DTG, the presence of hepatitis B antigen (HBsAg) or HBV DNA, cancer within the past five years, unstable cardiovascular disease or diabetes (evaluated by the treating physician), pregnancy, or breastfeeding.
Participants were screened and recruited from outpatient clinics at the Departments of Infectious Diseases at Copenhagen University Hospital – Amager and Hvidovre and Rigshospitalet. Eligible subjects received oral and written information and were subsequently invited to participate in the study. All participants provided oral and written informed consent. All study visits and data collection were conducted at Copenhagen University Hospital – Amager and Hvidovre.
Data collection
Data were collected at baseline, week 24, and week 48, including body weight, body composition (assessed by dual-energy X-ray absorptiometry [DEXA]), and fat distribution (analyzed via visceral adipose tissue [VAT] and subcutaneous adipose tissue [SAT] via low-dose computed tomography [CT]). Liver fibrosis and fat infiltration were evaluated via liver elastography, whereas coronary artery calcium (CAC) was assessed via CT. Blood tests were used to measure metabolic and cardiac parameters, with plasma HIV-RNA and CD4 cell counts taken at each visit and at week four for participants on DTG/3TC. The participants fasted for at least 6 h before the visits. To reduce cumulative X-ray exposure, individuals who had a CT scan within the past decade were not offered a non-enhanced cardiac CT scan. Owing to a radiographer strike, some participants missed baseline or follow-up cardiac CT scans. Additional details on data collection are available in the study protocol [10]. DEXA technicians and CT reviewers were blinded to the treatment allocation.
Intervention
Participants were randomized to either switch treatment to coformulated DTG 50 mg and 3TC 300 mg daily (Dovato, intervention arm) or to the control arm continuing treatment with coformulated DTG 50 mg, ABC 600 mg, and 3TC 300 mg daily (Triumeq). Study treatment was dispensed at the study site at baseline and week 24 and administered by participants.
Adherence
Adherence was assessed by measurement of plasma HIV RNA at baseline, week four (for the DTG/3TC group exclusively), week 24, and week 48. Participants were withdrawn from study participation in cases of virological failure, which was defined as a plasma viral load > 50 copies/ml measured by two consecutive blood samples within 30 days.
Outcomes
The primary outcome was defined as the difference in change in body weight from baseline to week 48 between the two treatment arms. The secondary outcomes included differences in fat distribution, body composition, metabolic parameters, cardiac conditions, liver stiffness and liver steatosis. Details on the secondary outcomes are listed in the study protocol [10].
Sample size
The estimated weight change at 48 weeks was from 0 to −1 kg for the DTG/3TC arm, and + 2 kg for the DTG/ABC/3TC arm.
Based on a 2:1 randomization ratio between the intervention and control arm, a significance level (α) of 5%, a power (β) of 80%, a mean difference in weight of 2 kg (Δ), and a standard deviation of 2.7 kg, the estimated sample size was 64 participants (intervention arm n = 44 and control arm n = 20).
To account for a 5% withdrawal/dropout, the total sample size was set at 70 participants.
Randomization
The participants were randomized after informed consent was obtained by one of two investigators. Randomization was performed electronically in REDCap by study personnel. The participants and the study personnel were unblinded to the randomization. The randomization list was generated by a statistician via block randomization with an allocation of 2:1 (intervention: control) stratified by study center and sex. The blocks had a uniform distribution for all combinations of three to six.
Statistical methods
Baseline characteristics are presented as mean with standard deviation (SD) and were compared by treatment arm via unpaired t tests for normally distributed continuous variables. Skewed variables are shown as median with interquartile range (IQRs) and were compared by the Mann‒Whitney U test. Categorical variables are presented as frequencies and percentages and were analyzed by chi-square tests.
The primary outcome was assessed in the intention-to-treat (ITT) and modified ITT/complete case (CC) populations. The ITT analysis included all randomized participants, regardless of adherence, accounting for those lost to follow-up, those who dropped out, or those who withdrew consent. The modified ITT/CC analysis focused on complete cases adhering to the protocol.
To analyze repeated measurements, a linear mixed model was employed, comparing changes in weight across treatment arms from baseline to week 24 and week 48, assuming a linear relationship over time. Missing data in the ITT analysis were treated as “missing at random” (MAR) and addressed with multiple imputation. The secondary continuous outcomes were analyzed similarly. Within-treatment arm changes were compared from baseline to week 48 via paired t tests or Mann‒Whitney U tests for continuous variables and McNemar tests for categorical paired data. No interim analysis was conducted.
Software
The data were entered and stored in a REDCap database. Data handling and statistical analyses were performed via RStudio 2022.12.0 Build 353.