This modeling study projects the future burden of hypertension-associated cardiovascular disease (CVD) among people living with HIV (PLHIV) compared with HIV-negative adults in Eswatini through 2045. The analysis aims to inform integrated HIV–CVD planning by estimating prevalence, overall CVD mortality, and hypertension-attributable CVD mortality under multiple assumptions about the relative risk (RR) of CVD associated with HIV and antiretroviral therapy (ART).
HIV prevalence trajectories were produced using EMOD-HIV, an agent-based model calibrated to Eswatini’s epidemic. These HIV trajectories were combined with age-standardized CVD estimates from the Global Burden of Disease (GBD) study and published RRs to derive HIV-stratified CVD projections. The study used a logistic generalized additive model to project CVD burden trajectories through 2045 and applied Monte Carlo methods to quantify uncertainty in estimates.
Five scenarios were specified to evaluate how different assumptions about the RR of CVD among PLHIV versus HIV-negative adults affect projections:
These scenarios isolate effects of stable versus increasing cardiometabolic risk profiles—particularly those potentially related to ART regimens such as dolutegravir (DTG)—on projected CVD burden.
Across all scenarios, PLHIV consistently had higher CVD burden than HIV-negative adults.
Scenario 1 (constant RR for prevalence): In 2025, projected CVD prevalence was 11.0% (95% UI: 9.0–13.5%) among PLHIV compared with 6.8% (6.0–7.8%) in HIV-negative adults. These prevalence estimates were stable through 2045 under the constant-RR assumption.
Scenario 2 (constant RR for mortality): CVD mortality rate in 2025 was projected at 0.60% (0.47–0.76%) for PLHIV versus 0.37% (0.31–0.45%) for HIV-negative adults. Mortality rates decline modestly through 2045 but maintain a consistent excess among PLHIV.
Scenario 3 (HTN-attributable mortality under constant RR): Hypertension-attributable CVD mortality was higher in PLHIV and remained stable through 2045. In women, HTN-attributable mortality was 73% (70–76%) for PLHIV versus 64% (61–66%) for HIV-negative women. In men, corresponding estimates were 61% (58–64%) for PLHIV versus 58% (55–60%) for HIV-negative men.
Scenario 4 assessed the effect of a post-DTG increase in RR following Eswatini’s 2021 DTG rollout. Under this scenario, HTN-attributable CVD mortality rose rapidly after DTG introduction and then stabilized:
These projections illustrate how ART-related metabolic shifts temporally associated with DTG rollout could amplify hypertension-related CVD mortality among PLHIV.
Scenario 5 modeled a gradual increase in RR from 2010–2045 to reflect cumulative metabolic and demographic changes. Under this scenario, the gap in HTN-attributable CVD mortality between PLHIV and HIV-negative adults widened substantially by 2045:
These estimates indicate a marked and growing excess burden of hypertension-attributable CVD mortality in PLHIV under scenarios of increasing cardiometabolic risk.
The analysis projects that excess CVD burden among PLHIV in Eswatini will persist through 2045 even under stable relative-risk assumptions. However, ART-related metabolic trajectories—particularly those linked temporally to dolutegravir (DTG)—have the potential to substantially widen the gap in HTN-attributable CVD mortality between PLHIV and HIV-negative adults.
The study highlights that hypertension-attributable CVD mortality is particularly elevated among women with HIV across scenarios. Policy implications noted by the authors include strengthening integrated HIV and noncommunicable disease (NCD) services: routine blood pressure screening, risk-based preventive therapy, and sex-specific counseling regarding DTG-associated metabolic risks to sustain long-term health gains among PLHIV.
Limitations reported in the source include that projections derive from mathematical modeling combining multiple data sources (EMOD-HIV outputs, GBD estimates, published RRs) and thus depend on underlying assumptions about RRs and ART effects. The article is a preprint and has not been peer reviewed; the source states it should not be used to guide clinical practice without further validation.
The analysis used aggregated, publicly available data and previously developed simulation outputs; it did not involve individual-level human subjects data, and institutional ethics approval was not required. The authors declared no competing interests. The article is posted as a medRxiv preprint and carries a CC-BY 4.0 license.