Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 20.
Published in final edited form as: Lancet HIV. 2026 Mar 27;13(5):e306–e315. doi: 10.1016/S2352-3018(25)00354-6

Risk of obesity, diabetes, hypertension, and MACE following a switch to an integrase inhibitor: a target trial emulation in REPRIEVE

Emma M Kileel 1,2, Carlos D Malvestutto 3, Markella V Zanni 4, Emma Davies-Smith 5, Carl J Fichtenbaum 6, Jordan E Lake 7, Janet Lo 4, Judith A Aberg 8, Esteban Martinez 9,10, Kunjal Patel 5,11, Gerald S Bloomfield 12, Sarah M Chu 4, Aya Awwad 4, Marissa R Diggs 4, Alex B Lu 4, Sara H Bares 13, Daniel Berrner 14, Marek Smieja 15, Nwora Lance Okeke 16, Princy Kumar 17, Esau João 18, Judith S Currier 19, Michael T Lu 20, Matthew P Fox 1,2, Alana T Brennan 1,2, Pamela S Douglas 21, Steven K Grinspoon 4, Heather J Ribaudo 5
PMCID: PMC13185539  NIHMSID: NIHMS2170313  PMID: 41911941

Summary

Background:

Integrase strand transfer inhibitors (INSTIs) are linked to weight gain, but data on their cardiometabolic effects, particularly major adverse cardiovascular events (MACE), are limited. We aimed to estimate the risk of obesity, diabetes, hypertension, and MACE, after switching to an INSTI among people with HIV (PWH) at low-to-moderate cardiovascular risk.

Methods:

Using data from the global Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE), we emulated a series of trials comparing switching to an INSTI versus remaining on a non-INSTI regimen on risk of incident obesity, diabetes, hypertension, and MACE. Participants without a prior diagnosis of the outcome or prior INSTI use were followed up to five years, until the event, loss to follow-up, death, or study end. Data from emulated trials were pooled and hazard ratios were estimated using weighted Cox-proportional hazard models accounting for competing events.

Findings:

5,114 participants met eligibility criteria for at least one emulated trial. 2,985 (58%) participants were captured as an INSTI switcher in at least one trial. Most switchers (81%) initiated a dolutegravir-based regimen and most (67%) non-switchers were on an efavirenz-based regimen. The average hazard of obesity (HR: 1.41, 95% CI: 1.22–1.59), diabetes (HR: 1.50, 95% CI: 1.24, 1.81), and hypertension (HR: 1.45, 95% CI: 1.26, 1.67) was higher among switchers compared to non-switchers. We did not observe an effect on MACE (HR: 1.17, 95% CI: 0.87–1.57). Point estimates were higher among females compared to males. Results persisted after adjustment for tenofovir alafenamide, tenofovir disoproxil fumarate, prior efavirenz use, and change in BMI.

Interpretation:

The increase in cardiometabolic risk profile suggests long-term observation of PWH switching to INSTIs will be critical to ensure appropriate management of co-morbidities and to assess for future development of MACE.

Funding:

National Institutes of Health, Kowa Pharmaceuticals America, Gilead Sciences, and ViiV Healthcare.

Introduction

Integrase strand transfer inhibitors (INSTIs), a class of antiretroviral therapy (ART) that includes drugs dolutegravir, raltegravir, elvitegravir, bictegravir, and cabotegravir, have been widely recognized for their superior safety profiles and enhanced resistance barriers and are now the recommended ART regimens globally.1 While these drugs have demonstrated improved tolerability and adherence,2 there has been growing concern over their association with weight gain. Recent evidence suggests that INSTIs, compared to protease inhibitors (PIs) or non-nucleoside reverse transcriptase inhibitors (NNRTIs), are associated with excess weight gain, even among those who are virally suppressed at the time of initiation.3 The effect appears to be heightened in certain groups, including women and individuals of Black race.3

This phenomenon of INSTI-associated weight gain is particularly concerning given the well-established connection between weight and cardiometabolic conditions.4 Irrespective of treatment strategy, conditions like diabetes, hypertension, and cardiovascular disease are occurring at higher rates and younger ages among people with HIV (PWH) compared to people without HIV.5 Whether INSTI use and its associated weight gain exacerbates cardiometabolic disease risk remains unclear. While studies evaluating the cardiometabolic consequences of INSTI use have started to emerge, results have been conflicting and follow-up periods, in general, have been limited to 12- and 24-months.6–8 Moreover, most studies have been restricted to specific geographic regions, with the majority conducted in high-income settings like Europe and North America.

The Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE) offers an ideal setting to further our understanding of the metabolic and cardiovascular impact of INSTI use. REPRIEVE was the first large-scale primary prevention trial of cardiovascular disease in PWH, designed to test whether daily pitavastatin calcium reduced the risk of major adverse cardiovascular events (MACE) among PWH at low-to-moderate cardiovascular disease risk.9 Here, we conduct a secondary analysis of REPRIEVE data to evaluate the impact of switching to an INSTI-containing regimen on cardiometabolic outcomes. Leveraging this global cohort of ART experienced PWH, we emulated a series of target trials to estimate the effect of switching to an INSTI-based treatment regimen, versus remaining on an PI- or NNRTI-based regimen, on the risk of incident obesity, diabetes, hypertension, and MACE over 5-years of follow-up.

Methods

Data Source

REPRIEVE (NCT02344290) is a multicenter, phase 3, prospective, double-blind, randomized trial designed to assess the effects of a daily dose of pitavastatin calcium versus placebo on the reduction of MACE among PWH on ART.9 Eligible participants were aged 40–75 years, on a stable ART regimen for at least 6 months with a CD4+ T cell count > 100 cells/mm3, and with low-to-moderate risk for atherosclerotic cardiovascular disease.9 Between March 2015 and July 2019, REPRIEVE enrolled 7,769 participants across 12 countries. The trial was stopped for efficacy in March 2023 after it was found that participants in the pitavastatin arm had a 35% reduction in the hazard of MACE compared to participants in the placebo arm.10,11 Study closeout visits were completed by August 2023, allowing for a minimum of 4 and a maximum of 8 years of follow-up, depending on participants’ enrollment date (median follow-up 5.6 years).

Ethics Statement

REPRIEVE received approval from the Massachusetts General Brigham Human Research Committee (IRB approval number: 2013P001898), and each clinical research site obtained authorization from relevant regulatory bodies. Informed consent was collected from all participants.

Study Design

Target trial specification –

The target trial framework involves first specifying components of the protocol of the hypothetical trial that would have been conducted (target trial) to answer the research question of interest and then emulating this protocol using observational data.12 The protocol of our target trial would be as follows. PWH aged 40–75 years who have been on a stable non-INSTI-containing ART regimen for at least 6 months, with no previous INSTI exposure and no history of the event of interest, would be eligible for the trial. Eligible participants would be randomly assigned to one of two treatment strategies: 1) switch to an INSTI-containing ART regimen or 2) remain on a non-INSTI-containing ART regimen. Participants would be followed from treatment assignment until the event of interest, death, loss to follow-up, or end of trial at maximum of 5 years. Target trials were emulated for four distinct outcome measures: incident obesity, diabetes, hypertension, and MACE. The components of the target trial and the observational emulation, as well as outcome definitions, are described in the Supplemental Appendix p2–3.

Observational emulation of the target trial –

We emulated the target trial described above using data from REPRIEVE. The protocol components of the observational emulations were the same as the target trial components with three exceptions. First, since we did not have participants’ full ART histories in REPRIEVE, we could not conclusively rule out any previous INSTI exposure. However, participants who enrolled in REPRIEVE on an INSTI-based regimen were ineligible for all target trial emulations and those who switched to an INSTI during REPRIEVE follow-up were ineligible for all trials after INSTI switch. Second, participants enrolled in low- and middle-income regions, including Latin America, Southeast Asia, and Sub-Saharan Africa were not considered for trial eligibility until 2018 due to INSTI availability in these regions. Third, as switching to an INSTI was at the discretion of the participants and their clinicians in REPRIEVE, randomization in the observational emulation was assumed conditional on a set of measured baseline covariates.

We identified participants eligible for trial emulation starting on day zero of follow-up in REPRIEVE. Eligible participants were then assigned to the treatment strategy that was compatible with their data over a subsequent 60-day baseline period. Meaning that, if a participant switched to an INSTI-based regimen between days 0–59 of REPRIEVE, they were considered a “switcher”. If a participant did not switch to an INSTI-based regimen during the baseline period, they were considered a non-switcher. Time to the outcome of interest was assessed from the end of the baseline period to a maximum of five years. We repeated this process every 60 days, until REPRIEVE day 2160, emulating a total of 37 sequential trials each with the potential for at least two years of follow-up. For a visual representation of this approach and a description of how a participant could move through trials, see Supplemental Appendix p4–5.

The causal-contrast of interest of the emulated trial was the observational analog of the intent-to-treat effect of a single time-point intervention (switch now vs do not switch now with the potential to switch later), defined as the average effect of assignment to the treatment at that time point, regardless of adherence or subsequent changes in regimen. This approach aligns with the clinical reality that PWH have the potential to change ART regimens at any visit with their clinician.

Statistical Analysis

We used inverse probability of treatment weights (IPTWs) to achieve balance of measured confounders between switchers and non-switchers at the start of each sequential trial. Logistic regression models were used to predict the probability of treatment assignment (i.e., switch=1) based on a set of confounding characteristics identified a priori (Supplemental Appendix p6), including: natal sex, age, race, REPRIEVE enrollment region and year, alcohol, cigarette, and substance use, diet quality, physical activity levels, total ART duration at REPRIEVE entry, nadir CD4+ T cell count, estimated glomerular filtration rate at REPRIEVE entry, statin randomization group, atherosclerotic cardiovascular disease (ASCVD) risk score at REPRIEVE entry, and body mass index (BMI) at the start of each trial. Sequential trial number was included as a covariate in both the numerator and denominator of the IPTW models. To improve precision and reduce variability of the estimates, weights were stabilized by multiplying the IPTWs by the marginal probability of the treatment received. We assessed model specification by calculating standardized differences of covariates after weighting, as well as means and standard deviations of stabilized IPTWs for each trial.

To estimate the effect of switch vs no switch on outcomes of interest, we used IPT weighted cause-specific Cox proportional regression models with data pooled across all sequential trials. Death was considered a competing event. Estimated hazard ratios (HRs) provide an average HR over the follow-up period, reflecting a summary of the relative hazard over 5 years. To account for the repeated use of participants over sequential trials, we utilized non-parametric bootstrapping with 500 replicates to generate percentile-based 95% confidence intervals (CIs). Both the IPTW models and hazard models were estimated within each bootstrap sample, allowing the bootstrap CIs to account for the uncertainty in both the estimation of the weights and the hazard model.13 All analyses were conducted among the overall eligible population and among individuals born male and female separately. Analyses were conducted using SAS version 9.4 on a Linux operating environment.

Sensitivity and post-hoc analyses

We conducted the following sensitivity analyses: 1) adjusting for tenofovir alafenamide (TAF) use at sequential trial baseline due to its association with weight gain;14 2) adjusting for tenofovir disoproxil fumarate (TDF) use at sequential trial baseline due to potential weight suppressive effects;14 3) adding an eligibility criterion requiring participants to be taking TDF at sequential trial baseline to further understand the degree to which weight suppressive effects of TDF contributed to our findings; 4) adding an eligibility criterion requiring participants to be taking efavirenz at trial baseline to evaluate the impact of efavirenz, which has similarly shown to have weight suppressive properties;15 5) adding a criterion requiring switchers to switch to a dolutegravir-based regimen to try and isolate the effects of specific INSTI regimens; 6) adjusting for waist circumference measured at REPRIEVE entry as a proxy for lipodystrophy and body composition changes; 7) for the outcome of MACE, adding an eligibility criterion requiring participants to have an ASCVD risk score > 2.5 to assess whether results differ among those with some degree of baseline ASCVD risk; 8) using a 30-day baseline period to determine the extent to which the size of the baseline period impacted our study; 9) limiting sequential trial emulation to trials with the potential for at least 5-years of follow-up to evaluate whether differing lengths of follow-up in sequential trials impacted our analysis; 10) censoring the sequential trial emulation at 3-years of follow-up to evaluate variability in the average treatment effect over varying follow-up time ;and 11) using a pooled logistic regression with time as a functional form for estimation to ensure approximation to the hazard model.

We conducted the following exploratory, post-hoc analyses: 1) evaluating the impact of weight gain on the outcomes of diabetes, hypertension, and MACE by adjusting for time-updated percent change in BMI; 2) stratifying analyses by region (high income countries and low- to middle-income countries) and 3) evaluating the outcome of hard MACE, a more stringent composite outcome of myocardial infarction, stroke, and cardiovascular death.

Role of the funding source

The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

For each outcome of interest, we emulated 37 sequential trials. Overall, 5,114 participants were eligible for at least one emulated trial (Supplemental Appendix p6). Of these, 2,985 were captured as an INSTI switcher in exactly one trial for at least one outcome and 5,059 were captured as a non-switcher in at least one trial, contributing to a total of 99,357 person-trials. Demographic, behavioral, and cardiometabolic characteristics were largely similar between switchers and non-switchers (Table 1 and Table 2). The majority of non-switchers were on an efavirenz-based regimen with a TDF backbone (Table 3). This was also the most common regimen prior to an INSTI switch. Most switchers initiated a dolutegravir based regimen and the proportion on TAF increased from 7% pre-switch to 25% at the time of switch. For summaries of baseline confounding characteristics among those eligible for each outcome analysis, see Supplemental Appendix p8–11.

Table 1.

Demographic and behavioral characteristics among participants included in the emulation of a target trial of the effect of switching to an INSTI vs not switching on cardiometabolic outcomes nested within the REPRIEVE trial.

INSTI switchers
(N=2981 person-trials; 2981 unique individuals)		 Non switchers
(N=96,376 person-trials; 5059 unique individuals)		 Total
(N=99,357 person-trials; 5114 unique individuals)
Age (years) Median (Q1, Q3) 49 (45, 55) 49 (45, 54) 49 (45, 54)
40–49 1,493 (50.1%) 50,767 (52.7%) 52,260 (52.6%)
50–59 1,238 (41.5%) 38,245 (39.7%) 39,483 (39.7%)
60+ 250 (8.4%) 7,364 (7.6%) 7,614 (7.7%)
Natal sex Male 1,869 (62.7%) 60,957 (63.2%) 62,826 (63.2%)
Female 1,112 (37.3%) 35,419 (36.8%) 36,531 (36.8%)
Gender identity Cisgender 2,927 (98.2%) 94,143 (97.7%) 97,070 (97.7%)
Transgender Spectrum 39 (1.3%) 1,480 (1.5%) 1,519 (1.5%)
Not reported 15 (0.5%) 753 (0.8%) 768 (0.8%)
Race Black or African American 1,328 (44.5%) 36,476 (37.8%) 37,804 (38.0%)
White 687 (23.0%) 25,323 (26.3%) 26,010 (26.2%)
Asian 685 (23.0%) 23,037 (23.9%) 23,722 (23.9%)
Other 281 (9.4%) 11,540 (12.0%) 11,821 (11.9%)
Ethnicity Not Hispanic or Latino 714 (81.0%) 22,483 (78.2%) 23,197 (78.3%)
Hispanic or Latino 158 (17.9%) 5,920 (20.6%) 6,078 (20.5%)
Unknown 9 (1.0%) 346 (1.2%) 355 (1.2%)
Country/Region of Enrollment N.America 955 (32.0%) 30,573 (31.7%) 31,528 (31.7%)
S.America 555 (18.6%) 23,351 (24.2%) 23,906 (24.1%)
Africa 767 (25.7%) 18,113 (18.8%) 18,880 (19.0%)
Thailand 360 (12.1%) 13,687 (14.2%) 14,047 (14.1%)
India 316 (10.6%) 8,917 (9.3%) 9,233 (9.3%)
Spain 28 (0.9%) 1,735 (1.8%) 1,763 (1.8%)
Smoking status Current 618 (20.8%) 20,937 (21.7%) 21,555 (21.7%)
Former 631 (21.2%) 21,102 (21.9%) 21,733 (21.9%)
Never 1,729 (58.1%) 54,260 (56.3%) 55,989 (56.4%)
Alcohol use Missing 2 (0.1%) 123 (0.1%) 125 (0.1%)
Rarely/Never 2,292 (76.9%) 72,275 (75.0%) 74,567 (75.0%)
Usually/Often/Sometimes 687 (23.0%) 23,978 (24.9%) 24,665 (24.8%)
Substance use Current/Former 590 (19.8%) 20,166 (20.9%) 20,756 (20.9%)
Never 2,387 (80.2%) 76,120 (79.1%) 78,507 (79.1%)
Diet Ideal 531 (17.8%) 16,941 (17.6%) 17,472 (17.6%)
Intermediate 1,577 (53.0%) 54,020 (56.2%) 55,597 (56.1%)
Poor 870 (29.2%) 25,229 (26.2%) 26,099 (26.3%)
Physical activity Ideal 346 (11.6%) 11,660 (12.1%) 12,006 (12.1%)
Intermediate 1,392 (46.8%) 46,081 (47.9%) 47,473 (47.9%)
Poor 1,238 (41.6%) 38,386 (39.9%) 39,624 (40.0%)
Open in a new tab

Table 2.

Cardiometabolic and HIV-related health characteristics among participants included in the emulation of a target trial of the effect of switching to an INSTI vs not switching on cardiometabolic outcomes nested within the REPRIEVE trial.

INSTI switchers
(N=2981 person-trials; 2981 unique individuals)		 Non switchers
(N=96,376 person-trials; 5059 unique individuals)		 Total
(N=99,357 person-trials; 5114 unique individuals)
BMI (kg/m2) Median (Q1, Q3) 25.0 (22.1, 28.7) 25.4 (22.4, 29.0) 25.4 (22.3, 28.9)
<25 1,481 (49.7%) 45,114 (46.8%) 46,595 (46.9%)
25–29.9 955 (32.0%) 32,203 (33.4%) 33,158 (33.4%)
30+ 544 (18.3%) 19,053 (19.8%) 19,597 (19.7%)
Fasting Glucose (mg/dL) Median (Q1, Q3) 90 (84, 97) 91 (84, 98) 91 (84, 98)
Systolic Blood Pressure (mmHg) Median (Q1, Q3) 122 (112, 132) 121 (111, 132) 121 (111, 132)
Diastolic Blood Pressure (mmHg) Median (Q1, Q3) 80 (71, 85) 79 (70, 84) 79 (70, 84)
eGFR by CKD-EPI (mL/min per 1.73 mm2) Median (Q1, Q3) 102.3 (88.6, 114.1) 101.4 (88.0, 112.1) 101.4 (88.0, 112.1)
≥ 90 2,026 (68.0%) 64,916 (67.4%) 66,942 (67.4%)
60 - <90 912 (30.6%) 29,917 (31.0%) 30,829 (31.0%)
<60 43 (1.4%) 1,543 (1.6%) 1,586 (1.6%)
ASCVD risk score (%) Median (Q1, Q3) 3.9 (1.8, 6.6) 3.7 (1.6, 6.5) 3.7 (1.6, 6.5)
0-<2.5 947 (31.8%) 33,029 (34.3%) 33,976 (34.2%)
2.5-<7.5 1,504 (50.5%) 46,230 (48.0%) 47,734 (48.0%)
7.5–10 354 (11.9%) 11,493 (11.9%) 11,847 (11.9%)
>10 176 (5.9%) 5,624 (5.8%) 5,800 (5.8%)
Nadir CD4 (cells/mm3) <200 1,522 (51.1%) 46,616 (48.4%) 48,138 (48.4%)
200–349 833 (27.9%) 26,608 (27.6%) 27,441 (27.6%)
350+ 558 (18.7%) 20,745 (21.5%) 21,303 (21.4%)
Unknown 68 (2.3%) 2,407 (2.5%) 2,475 (2.5%)
Total ART use (years) <5 646 (21.7%) 23,408 (24.3%) 24,054 (24.2%)
5–10 933 (31.3%) 30,917 (32.1%) 31,850 (32.1%)
10+ 1,402 (47.0%) 42,027 (43.6%) 43,429 (43.7%)
Unknown 0 (0.0%) 24 (0.0%) 24 (0.0%)
CD4 count (cells per μL) 500+ 2,027 (68.0%) 66,236 (68.7%) 68,263 (68.7%)
<500 954 (32.0%) 30,140 (31.3%) 31,094 (31.3%)
HIV-1 RNA (copies/mL) <LLQ 1,725 (89.5%) 61,170 (90.0%) 62,895 (89.9%)
LLQ −< 400 167 (8.7%) 5,375 (7.9%) 5,542 (7.9%)
400+ 35 (1.8%) 1,456 (2.1%) 1,491 (2.1%)
Open in a new tab

Abbreviations: BMI=body mass index; eGFR= estimated glomerular filtration rate; ASCVD=atherosclerotic cardiovascular disease; ART=antiretroviral therapy; INSTI=integrase strand transfer inhibitors.

Table 3.

Antiretroviral therapy (ART) regimen details among participants included in the emulation of a target trial of the effect of switching to an INSTI vs not switching on cardiometabolic outcomes nested within the REPRIEVE trial.

INSTI switchers
(N=2981 person-trials; 2981 unique individuals)		 Non switchers
(N=96,376 person-trials; 5059 unique individuals)		 Total
(N=99,357 person-trials; 5114 unique individuals)
Pre-switch At switch
ART regimen class NRTI + INSTI - 2,765 (92.8%) - 2,765 (2.8%)
NRTI + NNRTI 2,089 (70.1%) 0 (0.0%) 69,335 (71.9%) 69,335 (69.8%)
NRTI + PI 814 (27.3%) 0 (0.0%) 25,070 (26.0%) 25,070 (25.2%)
NRTI-sparing 23 (0.8%) 107 (3.6%) 555 (0.6%) 662 (0.7%)
Other NRTI-containing 55 (1.8%) 109 (3.7%) 1,416 (1.5%) 1,525 (1.5%)
Choice of INSTI Dolutegravir - 2,412 (81.0%) - 2,412 (81.0%)
Bictegravir - 369 (12.4%) - 369 (12.4%)
Elvitegravir - 172 (5.8%) - 172 (5.8%)
Cabotegravir - 14 (0.5%) - 14 (0.5%)
Raltegravir - 12 (0.4%) - 12 (0.4%)
Choice of NRTI Tenofovir disoproxil fumarate plus emtricitabine or lamivudine 2,096 (70.9%) 1,718 (59.8%) 64,714 (67.5%) 66,432 (67.3%)
Tenofovir alafenamide plus emtricitabine or lamivudine 214 (7.2%) 733 (25.5%) 13,855 (14.5%) 14,588 (14.8%)
Zidovudine plus emtricitabine or lamivudine 401 (13.6%) 41 (1.4%) 9,514 (9.9%) 9,555 (9.7%)
Abacavir plus emtricitabine or lamivudine 177 (6.0%) 202 (7.0%) 6,027 (6.3%) 6,229 (6.3%)
Other NRTI 70 (2.4%) 180 (6.3%) 1,711 (1.8%) 1,891 (1.9%)
Choice of NNRTI Efavirenz 1,532 (72.0%) 1 (19.8%) 47,263 (67.1%) 47,281 (67.0%)
Rilpivirine 253 (11.9%) 61 (67.0%) 15,820 (22.5%) 15,881 (22.5%)
Nevirapine 323 (15.2%) 4 (4.4%) 6,770 (9.6%) 6,774 (9.6%)
Etravirine 18 (0.8%) 5 (5.5%) 465 (0.7%) 470 (0.7%)
Doravirine 1 (0.0%) 2 (2.2%) 139 (0.2%) 141 (0.2%)
Choice of PI Atazanavir 489 (56.7%) 10 (8.3%) 12,245 (46.1%) 12,255 (45.9%)
Darunavir 228 (26.4%) 103 (85.8%) 10,243 (38.6%) 10,346 (38.8%)
Lopinavir 137 (15.9%) 5 (4.2%) 3,825 (14.4%) 3,830 (14.4%)
Other PI 9 (1.0%) 1 (0.8%) 238 (0.9%) 239 (0.9%)
Open in a new tab

Abbreviations: ART=antiretroviral therapy; NRTI=nucleoside reverse transcriptase inhibitor; INSTI= integrase strand transfer inhibitor; NNRTI= non- nucleoside reverse transcriptase inhibitor; PI=protease inhibitor.

*

Total column represents summation between non switchers and switchers at the time of switch.

†

Percents are calculated within each ART regimen class.

During sequential trial follow-up, 4% of switchers discontinued their INSTI and 58% of non-switchers started an INSTI (these individuals were captured as switchers in later sequential trials). Over a median follow-up time of 2.8 years (IQR: 1.6–4.1), 9% of participants who switched to an INSTI developed obesity, 4% developed diabetes, 10% developed hypertension, and 2% experienced MACE. Among non-switchers, 8% developed obesity, 3% developed diabetes, 8% developed hypertension, and 2% experienced a MACE over a median follow-up time of 3.3 years (IQR: 1.9–4.4). Approximately 2% of participants experienced a non-MACE related death.

Weighted standardized differences were ≤ 0.1 for all confounders (Supplemental Appendix p8–11), and mean stabilized IPTWs were near 1 in each trial (Supplemental Appendix p12–13), indicating adequate balance, correct model specification, and no extreme inflation of sample sizes. In pooled Cox proportional hazard models, participants who switched to an INSTI had a higher average hazard of obesity (HR: 1.41, 95% CI: 1.22–1.59), diabetes (HR: 1.50, 95% CI: 1.24–1.81), and hypertension (HR: 1.45, 95% CI: 1.26–1.67), compared to participants who remained on a non-INSTI based regimen. (Figure 1). We did not observe a difference in the average hazard of MACE between switchers and non-switchers (HR: 1.17, 95% CI: 0.87, 1.57).

Figure 1.

Figure 1.

Open in a new tab

Average hazard of obesity, diabetes, hypertension, and MACE over 5-years of follow-up, among participants included in the emulation of a target trial of the effect of switching to an INSTI, vs not switching, nested within the REPRIEVE trial.

*X-axis shown on the logarithmic scale.

†Hazard Ratio estimates use inverse probability of treatment weights to adjust for natal sex, age, race, REPRIEVE enrollment region and year, alcohol, cigarette, and substance use, diet quality, physical activity levels, total antiretroviral therapy duration, nadir CD4 cell count, estimated glomerular filtration rate at REPRIEVE entry, statin randomization status, ASCVD risk score, and body mass index at the start of each sequential trial.

‡Number of events in non-switchers represent the total events contributed by repeated trials. 5909 obesity events among non-switchers corresponds to 541 unique events; 3190 diabetes events among non-switchers corresponds to 264 unique events; 4995 hypertension events among non-switchers corresponds to 444 unique events; 1498 MACE among non-switchers corresponds to 120 unique events.

In analyses stratified by natal sex, point estimates for all outcomes were higher among females, though confidence intervals were overlapping (Figure 1). Notably, the observed risk of MACE was higher among females who switched to an INSTI containing regimen compared to females who did not switch, however this finding was imprecise (HR: 1.42, 95% CI: 0.67, 2.77).

Results of sensitivity and post-hoc analyses are summarized in Table 4. Results were largely consistent with overlapping confidence intervals across all analyses. Findings were consistent after adjustment for TAF or TDF use, and when restricted to those taking TDF. Results were comparable when restricted to those on an efavirenz regimen at sequential trial baseline. In analyses restricted to participants with some degree of baseline cardiovascular risk (ASCVD ≥ 2.5), we observed a greater effect of switching on MACE (HR: 1.26, 95% CI: 0.89–1.55), though this finding was estimated with a wide confidence interval. Results were similar when using a 30-day baseline period (rather than 60-days) and when restricted to sequential trials with at least 5-years of follow-up. The estimated effect of switching to an INSTI on the hazard of obesity and diabetes was slightly larger when emulated trials were censored at 3 years (rather than 5 years), suggesting that the risk may be higher in the first few years after a switch.

Table 4.

Results of sensitivity and post-hoc analyses.

Obesity Diabetes Hypertension MACE
HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI)
Primary analysis estimates 1.41 (1.22, 1.59) 1.50 (1.24, 1.81) 1.45 (1.26, 1.67) 1.17 (0.87, 1.57)
Sensitivity analysis
1) Adjusted for TAF use 1.36 (1.15, 1.56) 1.55 (1.27, 1.86) 1.40 (1.19, 1.65) 1.08 (0.78, 1.49)
2) Adjusted for TDF use 1.44 (1.24, 1.64) 1.44 (1.19, 1.74) 1.38 (1.19, 1.60) 1.05 (0.77, 1.40)
3) Restricted to participants taking TDF at sequential trial baseline† 1.59 (1.07, 2.24) 1.61 (1.02, 2.54) 1.76 (1.57, 2.53) -
4) Restricted to participants taking efavirenz at trial baseline or prior to INSTI switch 1.42 (1.18, 1.69) 1.40 (1.05, 1.85) 1.41 (1.11, 1.72) 1.22 (0.67, 2.09)
5) Restricted switchers to dolutegravir 1.41 (1.16, 1.67) 1.36 (1.05, 1.68) 1.40 (1.12, 1.71) 1.29 (0.87, 1.80)
6) Adjusted for waist circumference 1.28 (1.09, 1.44) 1.45 (1.02, 1.66) 1.42 (1.02, 1.60) 1.19 (0.92, 1.51)
7) Restricted to participants with an ASCVD risk score ≥ 2.5 - - - 1.26 (0.89, 1.72)
8) Using a 30-day baseline period 1.43 (1.27, 1.64) 1.51 (1.25, 1.80) 1.47 (1.29, 1.68) 1.15 (0.85, 1.55)
9) Limited to sequential trials with a full 5-years of follow-up (trials 1–19) 1.27 (1.02, 1.54) 1.60 (1.19, 2.08) 1.39 (1.17, 1.63) 1.27 (0.91, 1.70)
10) Administrative censoring at 3 years 1.61 (1.19, 2.35) 1.82 (1.35, 2.36) 1.42 (1.20, 1.66) 1.17 (0.86, 1.56)
11) Pooled logistic regression* 1.44 1.46 1.43 1.09
Post-hoc analysis
1) Adjusted for percent change in BMI - 1.62 (0.47, 1.93) 1.47 (0.43, 1.71) 0.97 (0.28, 1.32)
2) Restricted to high income countries‡ 1.41 (1.16, 1.73) 1.54 (1.07, 2.10) 1.25 (0.87, 1.63) 0.98 (0.61, 1.45)
3) Restricted to low- to middle-income countries‡ 1.33 (1.11, 1.58) 1.49 (1.17, 1.87) 1.44 (1.17, 1.74) 1.30 (0.85, 1.94)
Open in a new tab

All analyses represent the effect of switching to an INSTI-based treatment regimen versus remaining on a non-INSTI based regimen on outcomes of interest.

Abbreviations: TAF=tenofovir alafenamide; TDF=tenofovir disoproxil fumarate; INSTI=integrase strand transfer inhibitor; ASCVD=atherosclerotic cardiovascular disease; BMI=body mass index.

*

Confidence intervals for the pooled logistic regression sensitivity analysis were not calculated due to the computational intensity of the estimation procedure.

†

Sensitivity analysis restricted to participants taking TDF at sequential trial baseline was not conducted for the outcome of MACE given the limited number of events and resulting lack of precision.

‡

High income countries included United States, Canada, and Spain. Low- to middle-income countries included Brazil, Haiti, Peru, Thailand, India, Botswana, South Africa, Uganda, and Zimbabwe.

In post-hoc analyses adjusting for time-updated percent change in BMI, we found the hazards for diabetes, hypertension, and MACE to be comparable to primary estimates, though less precise. We saw little evidence of regional variation in risk. Additionally, the hazard ratio for hard MACE was comparable to that of primary MACE (HR: 1.20, 95% CI: 0.76, 1.86).

Discussion

In a global cohort of ART-experienced PWH at low-to-moderate cardiovascular disease risk, we found an increased risk of cardiometabolic complications following a switch to an INSTI over five years of follow-up. While we did not observe an increased risk of MACE, the heightened risk of obesity, diabetes, and hypertension – all of which are known to be important contributors to cardiovascular disease risk – raise concern that INSTIs may contribute to an increased risk of MACE over the longer term.

Consistent with prior studies,16,17 we observed a higher risk of obesity among those who switched to an INSTI versus non-switchers. Although mechanisms remain unclear, evidence suggests pre-switch ART regimens may shape post-switch weight trajectories.18 In particular, greater weight gain has been seen in those switching from efavirenz-based regimens compared to PI-based regimens.15 Given that the majority of participants in our study who switched to an INSTI were previously on efavirenz, this may partially account for some of the weight gain observed in this study. However, sensitivity analyses suggest cessation of weight suppressive efavirenz was not driving our overall results. Choice of NRTI has also been shown to influence weight trajectories, with some data suggesting a weight suppressive effect of TDF.15 However, our findings of an elevated risk of obesity after switching to an INSTI were unchanged when restricted to those who continued or initiated TDF.

While the role of the pre-switch regimen and NRTI backbone warrant further investigation, neither pathway explains why females tend to experience greater weight gain post-switch compared to males,3 as was observed in our study. These findings could be due to differences in estrogen levels and hormonal milieu, which may affect adipocytes, leading to a greater propensity for fat accumulation among females compared to males.19,20 Findings from the Women’s Interagency HIV Study (WIHS) cohort demonstrated that menopausal phase at the time of a switch to an INSTI-based regimen differentially impacted post-switch weight, with females in peri-menopause or menopause experiencing faster increases in waist circumference and BMI.21 Women in our study were, on average, in their late 40’s to early 50’s, an age range that, for many, overlaps with the perimenopausal transition. Different results might have been observed in studies conducted in younger or older populations of women with HIV.

Data evaluating the cardiometabolic impact of INSTI use remain sparse and conflicting. While several reports have found no effect of INSTI use on the risk of diabetes,22,23 our findings of an increased risk of diabetes among switchers are consistent with results from the RESPOND cohort which found an increased risk of diabetes among INSTI users compared to non-users (HR: 1.48) over a median follow-up time of 5 years.7 Moreover, in line with previously published observational studies,24,25 we observed a relative increase in hypertension risk following a switch to an INSTI. While it is hypothesized that weight gain may partially mediate this risk, our exploratory post-hoc analysis showed that the association persisted after accounting for changes in BMI. One important aspect to consider is the type of fat gained following a switch to an INSTI. Greater accumulation of visceral fat, as opposed to subcutaneous fat, has been implicated in the promotion of hypertension and other cardiovascular risk factors.26,27 Future studies should look at the composition of fat in individuals initiating INSTIs to determine relative amounts of visceral and subcutaneous fat and the contributions to long-term cardiometabolic disease risk.

Understanding the mechanism through which INSTIs might disrupt cardiometabolic pathways is critical, particularly as it relates to long-term prevention of MACE. Our findings on MACE contrast with that of the RESPOND study, which found an increase in cardiovascular disease risk in the first 24 months after INSTI initiation.8 However, our results align with findings from the Swiss HIV Cohort Study and the HIV-CAUSAL Collaboration/Antiretroviral Therapy Cohort Collaboration, both of which similarly used an explicit target trial emulation approach to minimize bias.28,29 In these studies, as well as our own, an increased risk of MACE or cardiovascular disease was not observed. However, our study sounds the alarm by finding significant increases in the risk of obesity, diabetes, and hypertension, all of which are critical risk factors for MACE. Importantly, our findings apply to the highly relevant population of PWH with low-to-moderate cardiovascular risk, among whom it gives a clear indication of expected risk. Estimates may differ and there may be a larger effect on MACE among those with a greater degree of baseline cardiometabolic risk.

A few limitations must be considered when interpreting the results of this study. First, despite employing a sophisticated modeling approach involving the emulation of sequential target trials and IPTW to estimate the causal effect of switching to an INSTI-regimen, we cannot rule out the possibility of unmeasured confounding. Second, a smaller baseline period would result in better alignment of treatment assignment and time-zero. We favored a 60-day period to ensure a sufficient number of switchers in each trial and avoid violating positivity assumptions, particularly when stratifying by natal sex. We ran models using a 30-day baseline period and found comparable results, suggesting minimal impact of this bias in our study. Third, our estimated ITT effect is assumed to be smaller than the corresponding direct effect of switching to an INSTI in the hypothetical scenario of full compliance to assigned INSTI switch strategy. Our ITT estimate assumes that non-adherence to assigned treatment observed in our study reflects the real-world scenario. Fourth, the number of MACE in our study was relatively small, limiting our statistical power to detect differences between groups. Nevertheless, a major strength is that MACE was the primary adjudicated endpoint REPRIEVE, providing high internal validity and setting our analysis apart from studies that rely on non-adjudicated outcomes identified through coded electronic health record or claims data. Fifth, most of our cohort switched to a dolutegravir-containing regimen, limiting our ability to investigate differential effects by individual INSTI agents. Finally, estimation of hazard ratios as the population summary of interest has limitations.30 HRs cannot be interpreted as unconditional causal effects for the entire randomized population at each time point, and their magnitude may vary over follow-up if the treatment effect is not constant over time. However, HRs remain a valid and informative measure of effect when properly interpreted as a comparison of average hazard functions over a given time period. Recent work has shown that concerns about collider bias vanish under this interpretation and simulation studies have demonstrated that substantial time-varying bias due to depletion of the risk set only arises under extreme heterogeneity in baseline risk31, 32 – a scenario unlikely in our study given the strict eligibility criteria of the parent REPRIEVE trial, which ensured that all participants had relatively comparable levels of baseline risk.

In conclusion, in an ART-experienced, virologically controlled, global cohort of PWH at low-to-moderate CVD risk, switching to an INSTI-containing regimen resulted in an increased risk of new onset obesity, diabetes, and hypertension. While the absolute risk of these outcomes was small, given the elevated risk profile, long-term observation of PWH switching to INSTIs will be critical to ensure appropriate management of developing co-morbidities and to assess for future development of MACE.

Supplementary Material

Supplemental Appendix

Research in Context.

Evidence before this study:

We searched PubMed for literature published between January 1, 2015, and April 1, 2025, using the terms “integrase inhibitors” AND (“weight” OR “weight gain” OR “obesity” OR “diabetes” OR “hypertension” OR “major adverse cardiovascular events” OR “cardiovascular disease”). There is a large body of evidence demonstrating disproportionate weight gain among people with HIV (PWH) initiating integrase inhibitor strand-transfer inhibitor (INSTI) containing antiretroviral therapy (ART) regimens compared to protease inhibitor or non-nucleoside reverse transcriptase inhibitor containing regimens. Studies evaluating the cardiometabolic consequences of INSTI use are more limited and have been conducted primarily in high-income regions with follow-up periods typically ranging from 12- to 24-months. Only four studies have evaluated the cardiovascular impact of INSTI use. A study from the RESPOND cohort found an elevated risk of cardiovascular disease events among those exposed to INSTIs versus those with no INSTI exposure. However, these results were contradicted by findings from the Swiss HIV Cohort and the HIV-CAUSAL Collaboration which found no effect of INSTI use on cardiovascular disease events. These studies were limited to high-income regions, specifically Europe and North America.

Added value of this study:

Our prospective analysis using data from the global REPRIEVE trial contributes critical findings to the existing body of data evaluating the cardiometabolic consequences of INSTI use. Many studies to date have been limited to new ART initiators and do not extend beyond two years of follow-up. Our study evaluates new-onset obesity, diabetes, hypertension, and clinically adjudicated major adverse cardiovascular events (MACE) over five years of follow-up in a global population of ART experienced PWH at low-moderate cardiovascular disease risk.

Implications of all available evidence:

The increasing cardiometabolic risk profile developing among INSTI switchers compared to non-switchers suggests such patients, particularly females, will need attention to long-term management of comorbidities, and close follow-up for the potential development of MACE.

Acknowledgements

The study investigators thank the study participants, site staff, and study-associated personnel for their participation in the trial. In addition, we thank the following: the ACTG for clinical site support; ACTG Clinical Trials Specialists (Laura Moran, MPH, and Jhoanna Roa, MD) for protocol development and implementation support; the data management center, Frontier Science Foundation, for data support; the Center for Biostatistics in AIDS Research for statistical support; and the Community Advisory Board for input for the community.

Funding Statement

This study is supported through NIH grants U01HL123336 and 1UG3HL164285, to the Clinical Coordinating Center, and U01HL123339 and 1U24HL164284, to the Data Coordinating Center, as well as funding from Kowa Pharmaceuticals America, Inc., Gilead Sciences, and ViiV Healthcare. The NIAID supported this study through grants UM1 AI068636, which supports the ACTG Leadership and Operations Center; UM1 AI106701, which supports the ACTG Laboratory Center; and R03AI189203 which supports EMK. This work was also supported by the Nutrition Obesity Research Center at Harvard (P30DK040561 to SKG).

NHLBI Grants Policy Statement

The views expressed in this manuscript are those of the authors and do not necessarily represent the views of the National Heart, Lung, and Blood Institute or the National Institute of Allergy and Infectious Diseases; the National Institutes of Health; or the U.S. Department of Health and Human Services.

Conflicts of Interest

EMK reports no relevant disclosures.

CDM reports institutional research support by Lilly and personal fees from ViiV Healthcare and Gilead Sciences for participation in advisory board meetings outside the submitted work.

MVZ reports grant support through her institution from NIH/NIAID and Gilead Sciences, Inc., relevant to the conduct of the study, support for attending CROI and International Workshop for HIV and Women from conference organizing committee when abstract reviewer and/or speaker; and participation in DSMB for NIH funded studies, as well as grants from NIH/NIAID and NIH/NHLBI.

EDS reports no relevant disclosures.

CJF reports grant support through his institution from Gilead Sciences, ViiV Healthcare, GSK, and Merck, all outside the submitted work.

JEL reports consultancy to ViiV Healthcare and CytoDyn, Inc. outside the submitted work.

JL reports consulting fees from ViiV Healthcare as well as investigator-initiated research grant support from ViiV Healthcare related to this work.

JAA reports grants from Massachusetts General Hospital during the conduct of the study; institutional research support for clinical trials from Gilead Sciences, Glaxo Smith Kline, Janssen, Macrogenics, Merck, Pfizer, Regeneron, and ViiV Healthcare and personal fees for advisory boards from Glaxo Smith Kline/ViiV, Invivyd, Merck and Regeneron; and participation on DSMB for Kintor Pharmaceuticals, all outside the submitted work.

EM has received honoraria for lectures or advisory boards from Gilead, Janssen, MSD, and ViiV, and his institution has received research grants from MSD and ViiV, all outside the submitted work.

KP reports no disclosures.

GSB reports no relevant disclosures.

SMC reports no relevant disclosures.

AA reports no relevant disclosures.

MRD reports no relevant disclosures.

ABL reports no relevant disclosures.

SHB reports serving as a scientific advisor to Gilead Sciences including payment for expert testimony and research grants to their institution from Gilead Sciences, ViiV Healthcare, and Janssen unrelated to the submitted work

DB reports no relevant disclosures.

MS reports research funding from Merck Canada.

NLO reports grants from NIH/NHLBI during the conduct of the study and consulting fees from Gilead Sciences unrelated to the submitted work

PK serves as a consultant and receives research funding from ViiV Healthcare, Merck, Theratechnologies, and Gilead and owns stock in Pfizer, Gilead, Merck, GSK, and Moderna

ESJ reports no relevant disclosures.

JSC reports consulting fees from Merck and Company

MTL reports grant support through his institution from NIH/NHLBI and Kowa Pharmaceuticals America, Inc., for the conduct of the study; grant support to institutions from the American Heart Association, AstraZeneca, Ionis, Johnson & Johnson Innovation, MedImmune, the National Academy of Medicine, the NIH/NHLBI, and the Risk management Foundation of the Harvard Medical Institutions Incorporated, outside of the current work.

MPF reports no relevant disclosures.

ATB reports no relevant disclosures.

PSD reports no relevant disclosures.

SKG reports grants from NIH, KOWA Pharmaceuticals, Gilead Sciences, and ViiV Healthcare during the conduct of the study as well as personal consulting fees from TheraTechnologies and ViiV Healthcare, and service on the Scientific Advisory Board of Marathon Asset Management and Exavir Therapeutics all outside the submitted work.

HJR reports grants from NIH/NHLBI, Kowa Pharmaceuticals, Gilead Sciences, and ViiV Healthcare during the conduct of the study as well as grants from NIH/NIAID, and NIH/NHLBI, NIH/NIDDK, and NIH/NIA outside the submitted work.

Data sharing

Data from the REPRIEVE trial are available on reasonable request to mghreprievetrial@mgb.org. In this case, shared data may include individual participant data that underlie the results reported in this Article, after de-identification (text, tables, figures, and appendices). To gain access, data requestors will need to sign a data access agreement. Study protocols are available on clinicaltrials.gov (https://clinicaltrials.gov/study/NCT02344290). Analysis code is available at https://github.com/ekileel93/seqtrials.

References

  • 1.World Health Organization. Consolidated guidelines on HIV prevention, testing, treatment, service delivery and monitoring : recommendations for a public health approach. World Health Organisation; 2016; : 548. [PubMed] [Google Scholar]
  • 2.Ryom L, De Miguel R, Cotter AG, et al. Major revision version 11.0 of the European AIDS Clinical Society Guidelines 2021. HIV Med 2022; 23: 849–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Lake JE, Wu K, Bares SH, et al. Risk Factors for Weight Gain Following Switch to Integrase Inhibitor–Based Antiretroviral Therapy. Clin Infect Dis 2020; 71: e471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Powell-Wiley TM, Poirier P, Burke LE, et al. Obesity and Cardiovascular Disease: A Scientific Statement From the American Heart Association. Circulation 2021; 143: E984–1010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Gandhi RT, Sax PE, Grinspoon SK. Metabolic and Cardiovascular Complications in HIV-Infected Patients: New Challenges for a New Age. J Infect Dis 2012; 205: S353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.O’Halloran JA, Sahrmann J, Parra-Rodriguez L, et al. Integrase Strand Transfer Inhibitors Are Associated With Incident Diabetes Mellitus in People With Human Immunodeficiency Virus. Clin Infect Dis 2022; 75: 2060–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Rupasinghe D, Bansi-Matharu L, Law M, et al. Integrase Strand Transfer Inhibitor-Related Changes in Body Mass Index and Risk of Diabetes: A Prospective Study From the RESPOND Cohort Consortium. Clin Infect Dis 2025; 80. DOI: 10.1093/CID/CIAE406. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Neesgaard B, Greenberg L, Miró JM, et al. Associations between integrase strand-transfer inhibitors and cardiovascular disease in people living with HIV: a multicentre prospective study from the RESPOND cohort consortium. Lancet HIV 2022; 9: e474–85. [DOI] [PubMed] [Google Scholar]
  • 9.Grinspoon SK, Fitch KV., Overton ET, et al. Rationale and Design of the Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE) RCT# NCT02344290. Am Heart J 2019; 212: 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Grinspoon SK, Fitch KV, Zanni MV, et al. Pitavastatin to Prevent Cardiovascular Disease in HIV Infection. N Engl J Med 2023; published online July 23. DOI: 10.1056/NEJMOA2304146/SUPPL_FILE/NEJMOA2304146_DATA-SHARING.PDF. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Grinspoon SK, Ribaudo HJ, Douglas PS. Trial Update of Pitavastatin to Prevent Cardiovascular Events in HIV Infection. New England Journal of Medicine 2024; 390: 1626–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Hernán MA, Wang W, Leaf DE. Target Trial Emulation: A Framework for Causal Inference From Observational Data. JAMA 2022; 328: 2446–7. [DOI] [PubMed] [Google Scholar]
  • 13.Keogh RH, Gran JM, Seaman SR, Davies G, Vansteelandt S. Causal inference in survival analysis using longitudinal observational data: Sequential trials and marginal structural models. Stat Med 2023; 42: 2191–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Wood BR, Huhn GD. Excess Weight Gain With Integrase Inhibitors and Tenofovir Alafenamide: What Is the Mechanism and Does It Matter? Open Forum Infect Dis 2021; 8: ofab542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Erlandson KM, Carter CC, Melbourne K, et al. Weight Change Following Antiretroviral Therapy Switch in People With Viral Suppression: Pooled Data from Randomized Clinical Trials. Clin Infect Dis 2021; 73: 1440–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Sax PE, Erlandson KM, Lake JE, et al. Weight Gain Following Initiation of Antiretroviral Therapy: Risk Factors in Randomized Comparative Clinical Trials. Clin Infect Dis 2020; 71: 1379–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Lake JE, Wu K, Bares SH, et al. Risk Factors for Weight Gain Following Switch to Integrase Inhibitor-Based Antiretroviral Therapy. Clin Infect Dis 2020; 71: E471–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Wohl DA, Koethe JR, Sax PE, et al. Antiretrovirals and Weight Change: Weighing the Evidence. Clinical Infectious Diseases 2024; 79: 999–1005. [DOI] [PubMed] [Google Scholar]
  • 19.Shi H, Seeley RJ, Clegg DJ. Sexual Differences in the Control of Energy Homeostasis. Front Neuroendocrinol 2009; 30: 396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Brown LM, Clegg DJ. Central Effects of Estradiol in the Regulation of Adiposity. J Steroid Biochem Mol Biol 2009; 122: 65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Abelman RA, Ma Y, Mehta CC, et al. Switch to Integrase Strand Transfer Inhibitors During the Menopausal Transition Is Associated With Accelerated Body Composition Change in Women With Human Immunodeficiency Virus. Clinical Infectious Diseases 2024; 1. DOI: 10.1093/CID/CIAE540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Bhatt I, Khan Z, Lam P, Quale JM. 332. Exploring the Prevalence and Characteristics of Weight Gain and other Metabolic Changes in Patients with HIV Infection Switching to Integrase Inhibitor Containing ART. Open Forum Infect Dis 2019; 6: S176–7. [Google Scholar]
  • 23.Hwang YJ, Lesko CR, Brown TT, et al. Association between switching to integrase strand transfer inhibitors and incident diabetes in people with HIV. AIDS 2024; 38: 1696–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Brennan AT, Nattey C, Kileel EM, et al. Change in body weight and risk of hypertension after switching from efavirenz to dolutegravir in adults living with HIV: evidence from routine care in Johannesburg, South Africa. EClinicalMedicine 2023; 57. DOI: 10.1016/J.ECLINM.2023.101836. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Byonanebye DM, Polizzotto MN, Neesgaard B, et al. Incidence of hypertension in people with HIV who are treated with integrase inhibitors versus other antiretroviral regimens in the RESPOND cohort consortium. HIV Med 2022; 23: 895. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Chandra A, Neeland IJ, Berry JD, et al. The relationship of body mass and fat distribution with incident hypertension: Observations from the dallas heart study. J Am Coll Cardiol 2014; 64: 997–1002. [DOI] [PubMed] [Google Scholar]
  • 27.Hall JE, do Carmo JM, da Silva AA, Wang Z, Hall ME. Obesity, kidney dysfunction and hypertension: mechanistic links. Nature Reviews Nephrology 2019 15:6 2019; 15: 367–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Surial B, Chammartin F, Damas J, et al. Impact of Integrase Inhibitors on Cardiovascular Disease Events in People With Human Immunodeficiency Virus Starting Antiretroviral Therapy. Clin Infect Dis 2023; 77: 729–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Rein SM, Lodi S, Logan RW, et al. Integrase strand-transfer inhibitor use and cardiovascular events in adults with HIV: an emulation of target trials in the HIV-CAUSAL Collaboration and the Antiretroviral Therapy Cohort Collaboration. Lancet HIV 2023; 10: e723–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Hernán MA. The Hazards of Hazard Ratios. Epidemiology 2010; 21: 13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Beyersmann J, Schmoor C, Schumacher M. Hazards Constitute Key Quantities for Analyzing, Interpreting and Understanding Time-to-Event Data. Biometrical Journal 2025; 67: e70057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Abrahamowicz M, Beauchamp ME, Roberts EK, Taylor JMG. Revisiting the hazards of hazard ratios through simulations and case studies. Eur J Epidemiol 2025; 40: 611–29. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Appendix
NIHMS2170313-supplement-Supplemental_Appendix.docx (345.4KB, docx)

Data Availability Statement

Data from the REPRIEVE trial are available on reasonable request to mghreprievetrial@mgb.org. In this case, shared data may include individual participant data that underlie the results reported in this Article, after de-identification (text, tables, figures, and appendices). To gain access, data requestors will need to sign a data access agreement. Study protocols are available on clinicaltrials.gov (https://clinicaltrials.gov/study/NCT02344290). Analysis code is available at https://github.com/ekileel93/seqtrials.

RESOURCES