Predictors of subclinical atherosclerosis in HIV

A cross-sectional and multicentre study was carried out by the Lipid and Vascular Risk Unit of the Infanta Elena Hospital together with the Infectious Diseases Units of the Infanta Elena and Juan Ramón Jiménez Hospital in Huelva, Spain. This study was approved by the Research Ethics Committee of the province of Huelva with the code JFS-ATE-2017-01.

The HIV population of the province of Huelva is registered in a database, which currently includes a total of 1208 patients and is managed by the computer program Hospital Control Application (Aplicación de Control Hospitalario—AC&H™) used by the Infectious Diseases Units of the Infanta Elena and Juan Ramón Jiménez Hospitals in Huelva. This program is widely used in several hospitals of the National Health System, is a component of the VIH-Aplicación de Control Hospitalario (VACH) cohort [7] and collects demographic and clinical data, HIV transmission category, serological and immunovirological data, history of antiretroviral treatments performed, comorbidities, opportunistic diseases and specific data on diseases unrelated to HIV.

To enter the study, participants had to meet the following inclusion criteria: HIV infection with more than 10 years of evolution, age between 18 and 65 years and signed informed consent.

The presence of any of the following criteria excluded the participants from the study: class C cirrhosis according to the Child‒Pugh classification; positive serology for hepatitis B virus (HBV); type 1 diabetes mellitus; renal failure (glomerular filtrate < 60 mL/min by Chronic Kidney Disease Epidemiology Collaboration); recorded history of vascular disease including acute coronary syndrome, acute myocardial infarction, aortic aneurysm, ischaemic or haemorrhagic stroke, transient ischaemic attack (TIA), peripheral arterial disease (intermittent claudication, pain or paraesthesia at rest, trophic lesions, established gangrene), coronary revascularization (percutaneous or bypass) or any other arterial revascularization procedure, plaque visualized on coronary angiography or carotid ultrasound or arterial stenosis at any level (> 50%).

After informed consent was signed by each patient, an imaging study of the carotid (common carotid artery, bulbar, right and left internal and external carotid arteries) and femoral (common and superficial vessels, also on both sides) vascular territories was performed with a Toshiba Medical Systems Aplio XG scanner. The presence of atherosclerotic plaque was assessed by high-resolution colour vascular Doppler ultrasound following the Mannheim consensus protocol of 2004–2006 [8], which defines an atheroma plaque as a focal structure that encroaches on the arterial lumen by at least 0.5 mm.

The measurements obtained from the digital images were performed by the same experienced technician who did not know the clinical characteristics of the participants.

The results were included in a database created for this purpose, which complies with confidentiality guarantees and data protection regulations. Access to the database by professionals is restricted, and the use of data for this research was carried out anonymously.

We evaluated the presence of the classic risk factors associated with atherosclerosis in each participant: age, arterial hypertension, diabetes, dyslipidaemia, active smoking and a family history of early cardiovascular events following the definition criteria of the European Atherosclerosis Society [9]. We also considered analytical parameters such as total cholesterol, cholesterol bound to low- (LDL-C) and high-density lipoproteins (HDL-C), triglycerides, CD4/CD8 ratio at the beginning of treatment, CD4 nadir and viral suppression time, which was defined as < 200 copies/mL or < 50 copies/mL of HIV-1 RNA in plasma according to the method used (< 200 copies/mL from 1996 to 2006 and < 50 copies/mL subsequently), as well as the history of coinfection by hepatitis C virus (HCV), immunoglobulin G (IgG) positive for cytomegalovirus (CMV) and ART history. We considered treatments performed when any of the following therapeutic groups were used for at least one year: non-nucleoside reverse transcriptase inhibitors (NNRTIs), nucleoside and nucleotide reverse transcriptase inhibitors (NRTIs), protease inhibitors (PIs) and integrase strand transfer inhibitors (INSTIs).

Descriptive statistics were performed for all variables: the mean and standard deviation for continuous variables and frequencies with percentages for categorical and ordinal variables.

Student t tests were used to compare parametric quantitative variables, and Mann‒Whitney U tests were used for variables that did not follow normality. The chi-square test (or Yates’ correction for continuity) was used for categorical and ordinal variables.

Through logistic regression analysis, we evaluated possible independent predictors of the presence of atheromatous plaque among all the variables studied using the Wald test. The diagnostic performance of the model was obtained by the area under the receiver operating characteristics (ROC) curve (AUC).

The cutoff points were established using the ROC curve. A cutoff point of 0.3 was chosen to achieve a specificity of 100%, and a cutoff point of 1.7 was chosen to achieve a sensitivity of 0%.

The cutoff point 0.7 was selected because it contributed the most information to the model and yielded the highest diagnostic performance. Cutoff points 0.3 and 1.7 did not contribute additional information.

All analyses were performed using the statistical package R Commander UCA, version 4.4.1, released September 26, 2021, from the University of Cádiz (Spain). An alpha error of 0.05 was considered statistically significant.

Our study was developed in Europe, where the prevalence of traditional cardiovascular risk factors (CVRFs), such as smoking, obesity and alcohol abuse, is high [12]. Most of the participants in our study were men (66.6%) and were over 40 years old.

On the other hand, we selected patients with HIV infection with more than 10 years of evolution, which may have influenced the high prevalence of atheromatosis observed in our patients.

High variability is evident among published reports on the SAT prevalence in the HIV population. These diverse findings can be explained by the heterogeneity of the populations studied and the different criteria used to define atheromatous plaque. A total of 83.32% of the plaques found in our patients were in the carotid territory, and if we exclude the plaques found in the femoral territory, the prevalence of plaque was 52%, which was even higher than published rates. However, the number of territories studied must be considered. In most studies, the estimation of atheromatous plaque is performed on the common carotid [11], while in our case, we performed this estimation on the common carotid, bulbar, internal carotid, external carotid, common femoral and superficial femoral territories.

HIV infection is related to selective depletion of CD4 lymphocytes and increased expression of CD8, which lead to an inverted CD4/CD8 ratio (< 1) [13]. This ratio has recently emerged as a marker of immunosenescence and is independently associated with risk factors for both AIDS and non-AIDS events and mortality. In addition, it has been independently associated with intima-media thickness (IMT) progression [14] and SAT [15]. Specifically, a CD4/CD8 ratio < 0.5 or < 0.3 has been determined to be a risk factor for non-AIDS events or mortality, especially when the CD4 count is high (≥ 500 cells/μL) [16].

A decreased ratio can be maintained despite having achieved viral suppression and normalization of CD4 levels with ART [17]. Among the factors associated with a persistent inversion of the CD4/CD8 ratio are the presence of positive CMV serology, ART initiation before 1997, a low CD4 nadir and a short duration of viral suppression [18]. An inverted CD4/CD8 ratio and a low CD4 nadir have been associated with an increased risk of non-AIDS events [19].

The CD4/CD8 ratio tends to present greater stability over time than the absolute counts of CD4 and CD8 cells, and its predictive value for events unrelated to AIDS has been described as exceeding that of the absolute CD8 count in individuals with restored CD4 counts (> 500 cells/µL) [20].

In our cohort, we observed an association between inversion of the CD4/CD8 ratio and the presence of carotid and femoral atheromatosis, which may constitute a parameter to be considered when estimating the risk of CVD in the HIV population without previous events. In addition, the presence of atheromatous plaque was associated with a low CD4 nadir and positive CMV serology, with no differences in viral suppression time found.

Through the use of predictors included in the logistic regression analysis, we obtained a diagnostic yield measured by the ROC curve of 72%.

In our cohort, we observed greater use of PIs in the SAT group. These results must be considered with caution, since patients with HIV follow different therapeutic regimens throughout their disease courses, and within the PI group, the new generations of PIs are very different in terms of side effects and cardiovascular safety.

Although studies related to normalization of the CD4/CD8 ratio with early ART initiation have been published [20‐22], no differences have been found among the different classes of ART used [23].

Although the earliest PIs, such as ritonavir, have been associated with an increased risk of CVD [24], the use of new therapies, such as atazanavir, has been associated with slower progression of atheromatous plaque and a lower incidence of CVD [25].

Vascular ultrasound of the carotid and femoral territories allows assessment of the presence of atheromatous plaque in these locations and provides information on the extent of atherosclerosis at the systemic level.

This approach is fast and innocuous and may be useful to reclassify the vascular risk of people with HIV [26], allowing treatment readjustment with the objective of reducing the vascular risk and avoiding CVD in the future.

Performing ultrasound in primary prevention in intermediate-risk patients is relevant since it can permit vascular risk reclassification.

However, despite the high prevalence of plaque in the HIV population, this technique is normally not included in routine check-ups for several reasons, including a lack of ultrasound time and a necessary learning curve.

Identifying the predictive factors for the presence of SAT is essential to be able to reclassify patients and select those who will obtain a greater benefit from evaluations for SAT.

Therefore, the findings obtained in our cohort may be relevant to vascular risk assessments in these patients.

The CD4/CD8 ratio can play a role in comprehensive assessments of vascular risk in people with HIV, constituting a simple and accessible tool.

Through three clinical-analytical parameters easily extracted from the clinical history of participants (CD4/CD8 ratio less than 0.7, age and active smoking), we were able to predict the presence of SAT in HIV patients without previous CVD.

As the sample size was small due to the design of the study, these results should be considered with caution, and additional studies are necessary to confirm the findings.

A tendency towards a greater incidence of traditional vascular risk factors, such as hypertension and diabetes, was evident in the group with atheromatous plaque, although without significance, probably due to the insufficient sample size. We must also consider the greater use of PIs by patients with atheromatous plaque, which have been associated with an elevated vascular risk in people with HIV in other studies.