CVD risk prediction models underestimate risk in patients with HIV infection

Cardiovascular disease (CVD) risk prediction models are insufficient for patients with HIV infection and often underestimate the risk for CVD events, according to study findings published in the Journal of the American Heart Association.

Patients with HIV infection have CVD risk that is 50% to 100% higher than that of the general population. The effect of HIV on CVD outcomes appears to depend on sex, age, virologic status, and immunologic status. However, the higher CVD risk in HIV-positive patients cannot be fully explained by traditional CVD risk factors.

Researchers conducted a longitudinal observational study to assess the performance of 3 CVD risk prediction functions across 2 HIV cohorts. Study patients (age range, 30-79 years) were evaluated by the American College of Cardiology/American Heart Association (ACC/AHA) atherosclerotic CVD function; Framingham Heart Study (FHS) hard coronary heart disease (CHD) function; and FHS function for global CVD function risk scores. The predictive power of the FHS CHD function was assessed for outcomes of myocardial infarction and coronary death, and the predictive power of the ACC/AHA and FHS CVD functions was assessed for outcomes of myocardial infarction, stroke, coronary death, and death due to stroke.

Patients with HIV infection were propensity-score matched up to 4:1 against those without HIV infection (controls) from the same health care system on the basis of traditional CVD risk factors and demographic factors. Outcomes were assessed between January 2000 and January 2020.

Among patients (N=9412) included in the final analysis, the median age was 45.1 (IQR, 39.1-51.4) years, 8148 (86.6%) were men, 4933 (52.4%) were White, the median CD4⁺ count was 492 (IQR, 316-690) cells/mm³, 6356 (68%) had HIV RNA levels below 400 copies/mL, and 7833 (83.2%) reported use of antiretroviral therapy (ART).

During a median follow-up period of 5.2 years, 158 (1.7%) patients experienced a CHD event and 309 (3.3%) experienced a CVD event. Compared with those without CVD events, patients with a CVD event were older (median age, 50.9 vs 45.0 years), more likely to be men (88.0% vs 86.5%) and active smokers (36.2% vs 25.8%), and had higher rates of hypertension (43.7% vs 22.9%), hyperlipidemia (50.5% vs 31.1%), and diabetes (13.6% vs 5.8%). In regard to HIV-specific factors, patients who experienced CVD events also had higher rates of viral suppression (71.1% vs 67.9%), longer durations of HIV infection (median, 9.7 vs 6.4 years), more commonly reported ART use (86.1% vs 83.1%), and had lower CD4⁺ counts (median, 451 vs 495 cells/mm³).

For women and men with HIV infection, the median predicted 10-year risk for a CVD event, respectively, was 1.9% and 4.8% via the ACC/AHA function, 0.7% and 3.1% via the FHS CHD function, and 1.3% and 3.7% via the FHS CVD function. Overall, the risk scores generated by the predictive functions had good discrimination for women (c-statistic range, 0.78-0.90) but moderate discrimination for men (c-statistic range, 0.71-0.72), with poor goodness of fit irrespective of sex (χ range, 3.43-24.12).

Across all predictive functions, comparisons between observed and predicted 10-year risk indicated a general pattern in which the risk for a CVD event was underestimated. Among women, the risk was underestimated by all 3 predictive functions for most risk tertiles. For men, the ACC/AHA function underestimated risk for the low and middle predicted risk tertiles, the FHS CHD function overestimated risk, and the FHS CVD function underestimated risk.

The same general pattern in which the predictive functions overestimated risk was also observed among HIV-negative patients in the control cohort, with the exception of the FHS CVD risk score. Moreover, the calibration of all risk functions was poor for both women and men.

Limitations of this study include the low rate of CVD events among women and the possibility of incomplete data due to potential out-of-network events.

Based on these findings, “There is an urgent need for tailored strategies to accurately estimate cardiovascular risk in people living with HIV,” the researchers concluded.

Disclosures: Multiple study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of disclosures.

By Jessica Nye, PhD

References:

Triant VA, Lyass A, Hurley LB, et al. Cardiovascular risk estimation is suboptimal in people with HIV. J Am Heart Assoc. Published online May 18, 2024. doi:10.1161/JAHA.123.029228