Boston University
Integrase inhibitors and cardiometabolic disease in people living with HIV: evidence from sequential target trial emulations
Abstract
dc:description.abstractOver the last decade, substantial advancements in antiretroviral therapy (ART) have transformed the HIV disease landscape. HIV, once a life-threatening condition, is now considered a manageable chronic illness so long as the individual adheres to long-term ART regimens. However, despite having improved safety profiles and enhanced resistance barriers, the newest class of ART regimens, integrase-strand transfer inhibitors (INSTIs), have been linked to clinically significant weight gain compared to other ART strategies. Whether this weight gain subsequently increases an individual’s risk of cardiometabolic disease, including diabetes, hypertension and cardiovascular disease, remains unclear. Existing literature on the topic are conducted primarily in high-income settings among treatment naïve individuals and show conflicting results. Given the widespread adoption of INSTIs as the recommended first and second line regimens for HIV treatment, understanding their associated cardiometabolic disease risk is essential. In this dissertation, we leveraged data from the global Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE) to address these critical evidence gaps. In the first study, we utilized a sequential target trial emulation approach to estimate the effect of switching to an INSTI-based treatment regimen, versus remaining on a non-INSTI based regimen, on the 5-year risk of incident obesity, diabetes, hypertension, and major adverse cardiovascular events (MACE). Data from emulated trials were pooled and hazard ratios (HRs) were estimated using counterfactual methods including inverse probability of treatment weighted Cox proportional hazard models accounting for competing risks. We found that participants who switched to an INSTI-based regimen, compared to participants who did not switch, had a higher average hazard of obesity (HR: 1.41, 95% confidence interval (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). However, we did not observe a meaningful effect of switching on MACE (HR: 1.17, 95% CI: 0.87, 1.57). In analyses stratified by sex, point estimates were higher among females compared to males. In the second study, we used the sequential target trial emulation approach to emulate a series of trials comparing the effect of switching to an INSTI-based treatment regimen, versus remaining on a non-INSTI based regimen, on changes in blood pressure and incident hypertension. We then applied causal mediation analysis methods to estimate the total intent-to-treat effect, the natural direct effect (NDE), and the natural indirect effect (NIE) of switching to an INSTI on changes in systolic and diastolic blood pressure, mediated by percent weight gain among males and females separately. Additionally, we applied causal mediation methods for survival analyses to estimate the proportion of incident hypertension risk mediated through percent weight gain. Among females, we observed a meaningful increase in systolic blood pressure (SBP) among switchers compared to non-switchers that increased over the duration of follow-up (change in SBP at Month 24 (M24): 0.3mmHg, 95% CI: -0.3, 0.9 mmHg; M36: 0.4, 95% CI: -0.3, 1; M48: 2 mmHg, 95% CI: 1, 3 mmHg; M60: 4 mmHg, 95% CI: 3, 5 mmHg). Notably, weight gain explained only 5% of the 4-mmHg increase at Month 60. We observed similar trends in diastolic blood pressure. In contrast, males showed minimal changes in both systolic and diastolic blood pressure, with weight gain explaining some but not all of this effect. In the overall population, the relative risk of hypertension was 42% higher among switchers compared to non-switchers (HR: 1.42, 95% CI: 1.22, 1.64), with little evidence of mediation through weight gain (6%), though this may reflect the limited number of events. In the third study, we sought to investigate the joint effects of pitavastatin and INSTI-based ART regimens on the risk of obesity, diabetes, hypertension, and MACE. We employed the target trial framework to describe the ideal trial that we would conduct to estimate joint effects of randomization of pitavastatin (versus placebo) and switching to an INSTI (or no switch). However, the data structure did not support a reasonable emulation of the target trial as statin randomization and regimen changes did not occur simultaneously, or in a temporally consistent manner. Consequently, we emulated a series of trials of switching versus not switching to an INSTI-based regimen and considered pitavastatin randomization an effect modifier. We used inverse probability of treatment weighted generalized linear models, stratified by levels of pitavastatin randomization, specifying binominal distribution and appropriate link functions to estimate risk differences (RDs) and risk ratios (RRs). Estimates showed modest increases in the relative risks of obesity, diabetes, and hypertension following a switch to an INSTI in both pitavastatin and placebo arms. For MACE, estimates indicated minimal association with INSTI switch in the pitavastatin arm (RR: 1.06, 95% CI: 0.60, 1.68) but a modest increased risk in the placebo arm (RR: 1.51, 95% CI: 1.05, 2.13), suggesting potential effect modification. In summary, we leveraged a global cohort of people with HIV at low-to-moderate cardiovascular disease risk and employed advanced epidemiologic and analytic methods to estimate causal effects of switching to INSTI-based treatment regimens on cardiometabolic outcomes. While absolute risks were small, our findings suggest a modest but concerning cardiometabolic risk profile emerging after INSTI initiation, warranting close clinical monitoring. We also clarify potential mechanistic pathways, identifying weight gain as one contributor to INSTI-associated blood pressure changes while also suggesting that additional biologic or sex-specific mechanisms may play a role. Finally, we found minimal evidence that statin therapy modified these associations, but observed that INSTI initiation may elevate long-term cardiovascular risk in the absence of statin therapy, highlighting the potential value of pitavastatin as a preventive strategy for people with HIV initiating INSTI-based ART. Together, this dissertation offers a comprehensive assessment of the cardiometabolic impact of INSTI-based ART.
Degree
thesis:*- Name thesis:degree_name
- Doctor of Philosophy
- Level thesis:degree_level
- doctoral
- Discipline thesis:degree_discipline
- Epidemiology
- Grantor
- Boston University
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Kileel, Emma Margaret
- Advisor dc:contributor.advisor
-
- Brennan, Alana T.
- ORCID dc:identifier.orcid
Subjects
dc:subject × 1Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/2144/52944