Association of HIV Infection With Cardiovascular Pathology Based on Advanced Cardiovascular Imaging: A Systematic Review

Abstract

Importance: HIV-associated cardiovascular disease is increasing in prevalence, but its mechanisms remain poorly understood.

Objective: To systematically review data from advanced cardiovascular imaging studies evaluating computed tomographic coronary angiography, positron emission tomography (PET), and cardiac magnetic resonance (MR), in people living with HIV compared with uninfected individuals.

Data sources: Three databases and Google Scholar were searched for studies assessing cardiovascular pathology using computed tomographic coronary angiography, cardiac MR, PET, and HIV from inception to February 11, 2022.

Study selection: Two reviewers selected original studies without any restrictions on design, date, or language, investigating HIV and cardiovascular pathology.

Data extraction and synthesis: One investigator extracted data checked by a second investigator. Prevalence ratios (PRs) and differences in inflammation among people living with HIV and uninfected individuals were qualitatively synthesized in terms of cardiovascular pathology. Study quality was assessed using the National Heart, Lung, and Blood Institute quality assessment tool for observational studies.

Main outcomes and measures: Primary outcomes were computed tomographic coronary angiography-defined moderate to severe (≥50%) coronary stenosis, cardiac MR-defined myocardial fibrosis identified by late gadolinium enhancement, and PET-defined vascular and myocardial target to background ratio. Prevalence of moderate to severe coronary disease, as well as myocardial fibrosis, and PRs compared with uninfected individuals were reported alongside difference in vascular target to background ratio.

Results: Forty-five studies including 5218 people living with HIV (mean age, 48.5 years) and 2414 uninfected individuals (mean age, 49.1 years) were identified. Sixteen studies (n = 5107 participants) evaluated computed tomographic coronary angiography; 16 (n = 1698), cardiac MRs; 10 (n = 681), vascular PET scans; and 3 (n = 146), both computed tomographic coronary angiography and vascular PET scans. No studies originated from low-income countries. Regarding risk of bias, 22% were classified as low; 47% moderate; and 31% high. Prevalence of moderate to severe coronary disease among those with vs without HIV ranged from 0% to 52% and 0% to 27%, respectively, with PRs ranging from 0.33 (95% CI, 0.01-15.90) to 5.19 (95% CI, 1.26-21.42). Prevalence of myocardial fibrosis among those with vs without HIV ranged from 5% to 84% and 0% to 68%, respectively, with PRs ranging from 1.01 (95% CI, 0.85-1.21) to 17.35 (95% CI, 1.10-274.28). Differences in vascular target to background ratio among those with vs without HIV ranged from 0.06 (95% CI, 0.01-0.11) to 0.37 (95% CI, 0.02-0.72).

Conclusions and relevance: In this systematic review of studies of advanced cardiovascular imaging, the estimates of the associations between HIV and cardiovascular pathologies demonstrated large amounts of heterogeneity. The findings provide a summary of the available data but may not be representative of all individuals living with HIV, including those from low-income countries with higher HIV endemicity.

Figure 1.. The Geographical Distribution of Advanced Cardiovascular Imaging Studies

Cartogram for the geographical distribution of cardiovascular imaging studies included in this systematic review. Numbers in parentheses represent total number of participants with HIV included in the studies evaluated in this review. Blue indicates high-income countries and beige indicates upper middle-income countries classified by the World Bank Income Group. There were no studies performed in low-income countries defined by a Global National Income per capita of $1085 or less as of 2021.

Figure 2.. Risk of Moderate to Severe Coronary Artery Disease Detected by Computed Tomographic Coronary Angiography in People Living With HIV and Uninfected Comparators

^aOf the 16 studies included, 10 reported the prevalence of more than 50% coronary stenosis in people living with HIV and uninfected individuals. Three studies reported both more than 50% and more than 70% stenosis,^,, and no studies reported only more than 70% stenosis. Of these 16 studies evaluating moderate to severe coronary disease detected on computed tomographic coronary angiography, 11 provided the prevalence in HIV infected and uninfected comparators with the prevalence ratio shown in this plot. ^bSee eFigure 3 in the Supplement for the pooled prevalence of moderate to severe coronary stenosis on computed tomographic coronary angiography in people living with HIV. ^cAbsolute difference represents the difference in risk of coronary stenosis detected on computed tomographic coronary angiography in people living with HIV and uninfected controls. ^dThe prevalence ratio for coronary stenosis is an unadjusted ratio (unless statistical adjustments are indicated) calculated from the absolute risk of coronary stenosis in people living with HIV over the absolute risk of stenosis in HIV negative individuals. See eFigure 7 in the Supplement for a pooled estimate. ^eA continuity correction has been applied for 0 cell count. The data points represent the central estimate of the prevalence ratio; error bars, 95% CI. The size of the boxes represent the number of participants in the study.

Figure 3.. Risk of Myocardial Fibrosis Determined by Late Gadolinium Enhancement on Cardiac Magnetic Resonance Imaging in People Living With HIV and Uninfected Comparators

^aOf the 15 studies evaluating cardiac magnetic resonance (CMR), 9 provided the prevalence of late gadolinium enhancement across HIV infected and uninfected comparators with the prevalence ratio shown in this plot. ^bSee eFigure 13 in the Supplement for the summary of the prevalence of late gadolinium enhancement in people living with HIV only. ^cAbsolute difference represents the difference in risk of late gadolinium on cardiac MR in people living with HIV and uninfected controls. ^dThe prevalence ratio is calculated as an unadjusted prevalence ratio unless stated. See eFigure 18 in the Supplement for a pooled estimate. ^eA continuity correction has been applied to the HIV-negative comparator group. The data points represent the central estimate of the prevalence ratio; error bars, 95% CI. The size of the boxes represent the number of participants in the study. Absolute differences or prevalence ratios could not be calculated for studies without HIV-negative comparison groups.