Age-related differences in the vascular function and structure of South Africans living with HIV

Background As the life expectancy of people living with the HIV increases because of antiretroviral treatment (ART), their risk for vascular co-morbidities and early vascular ageing (EVA) also increases. Objective We aimed to investigate whether HIV infection relates to vascular structure and function in black South African adults and whether this relationship is age dependent. Method This cross-sectional study carried out in urban and rural areas of North West province, South Africa, included 572 age- and sex-matched people living with HIV (PLWH) and without HIV. Participants from the EndoAfrica study and PURE study were stratified according to tertiles of age. Measures of vascular structure (carotid intima-media thickness) and function (carotid-femoral pulse wave velocity, central systolic blood pressure, central pulse pressure and pulse pressure amplification) were determined. Results Blood pressure measures were lower in PLWH compared with their controls (all P ≤ 0.001), especially in the younger and middle-aged groups (all P ≤ 0.031), whilst vascular measures did not differ (all P ≥ 0.611). In multivariate linear regression analyses, vascular measures were not associated with a HIV- positive status in either the total or any of the age groups. Conclusion Black South Africans living with HIV have a less adverse blood pressure profile than their counterparts without HIV. The HIV-positive status was not associated with measures of vascular structure or function in any age group. The results suggest that HIV does not contribute to EVA in this population; however, further longitudinal investigation is warranted.


Introduction
indicated that the effect of HIV on the vasculature may differ through the lifespan of PLWH. 27 HIV and/or ARTrelated vascular alterations may contribute to EVA. 8,9,10,28 However, there is a lack of literature regarding the development of EVA in PLWH from South Africa, where the vast majority are infected with HIV type-1 subtype C phenotype. 29 Therefore, we investigated whether HIV infection relates to vascular structure and function in South African adults on first-line ART at different ages.

Research methods and design Setting and study population
This study includes data of participants residing in North West province of South Africa who participated in both the EndoAfrica study (vascular endothelial dysfunction: the putative interface of emerging cardiovascular risk factors affecting populations living with and without HIV in SSA) and the Prospective Urban and Rural Epidemiology (PURE) study. Data from the EndoAfrica study were collected from 2017 to 2018, and data collection for the PURE study took place in 2015. Both the studies were designed to determine the risk of developing cardiovascular disease amongst South Africans living with and without HIV. Both studies were previously described in more detail elsewhere. 30,31 The EndoAfrica study recruited participants from seven local clinics in and around Potchefstroom; the participants for the PURE study were recruited from urban and rural communities, which include Potchefstroom (urban), Ganyesa and Tlakgameng (rural). Participants in the EndoAfrica study were 18-60 years and in the PURE study were 42-88 years old. For this study, we included 286 PLWH and 286 HIV-free participants (control group). The participants in the group with HIV were matched for age and sex with those who were HIV free ( Figure 1). The PLWH using treatment received first-line fixed dose combination ART. 26 The exclusion criteria for both studies include PLWH using second-and third-line ART, individuals who refused HIV testing or to provide a blood sample, women who were pregnant or less than three months post-partum. Additional exclusions were made for missing data of the main variables and/or an unknown HIV status.

Questionnaire and anthropometric measurements
A standardised questionnaire provided information regarding the age, sex, lifestyle patterns (which included tobacco use and alcohol consumption) and medication use in both studies.
Anthropometric measurements were performed according to the standardised methods of the International Society for the Advancement of Kinanthropometry. 32

Cardiovascular measurements
The brachial blood pressure, including systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate of each seated participant were taken twice after 10 min of rest (OMRON M6 automatic digital blood pressure monitor, Omron Healthcare, Kyoto, Japan). The final measurement was used in subsequent analyses. We identified participants as hypertensive if SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg 33 and/or if they used anti-hypertensive medication. Mean arterial pressure (MAP) was calculated, MAP = (SBP(2xdbp))/3. Central SBP (cSBP), central pulse pressure (cPP) and carotidfemoral pulse wave velocity (cfPWV) were measured by arterial waveform analyses (SphygmoCor ® XCEL System, AtCor Medical Pty. Ltd, Sydney, Australia). The transitdistance method was used to measure cfPWV along the descending carotid-femoral artery. The 80% rule was applied for the calculation of the distance used. 34 The cfPWV measurement was performed twice and repeated a third time if the two measurements differed by more than 3 m/s. The average was used for further analyses. Pulse pressure amplification (PPA) was defined as the ratio of the amplitude of the pulse pressure between a distal and proximal location.

Blood sampling and biochemical analyses
Blood samples were collected, centrifuged and aliquoted according to standardised methods, and were stored in −80 C bio-freezers for future analyses. Samples collected from the rural site during the PURE study were prepared at the on-site laboratory and immediately stored on dry ice (−18 °C) for a maximum of 5 days. These samples were then transported to the laboratory and stored at −80 °C.
In both studies, the Cobas Integra ® 400 Roche ® Clinical System (Roche Diagnostics, Indianapolis, IN, US) was used to determine glucose, total cholesterol (TC), high-density lipoprotein (HDL), low-density lipoprotein (LDL) and triglyceride levels using an enzymatic colourimetric method.
High-sensitivity C-reactive protein (CRP) was analysed with the particle-enhanced turbidimetric assay method and gamma-glutamyl transferase (GGT) with the enzymatic colourimetric assay method.
In the EndoAfrica study, plasma samples were sent to the National Health Laboratory Services (NHLS) to determine the CD4+ cell count (Beckman COULTER ® EPICS ® XLTM Flow Cytometer, GMI Inc., Fullerton, CA, US). The fingerprick blood and a point-of-care device (PIMATM CD4, Alere, Jena, Germany) with the fixed volume cytometry analysis were used in the PURE study.

Ethical considerations
This study complies with the requirements of the Declaration of Helsinki, and the Health Research Ethics Committee of the North-West University approved the EndoAfrica study (NWU-00045-15-A1), the PURE study (NWU-00016-10-A1), as well as this study (NWU-00367-20-A1). All procedures were explained to the participants before any measurements were made, and all participants gave written informed consent.

Results
The characteristics of the study population are presented in Table 1. The PLWH and without HIV were matched for age (P = 0.098) and sex (P = 1.000), and their locality was comparable (P = 0.073). People living with HIV had a lower waist-to-height ratio (WtHR) and BMI (both P < 0.001) compared with the HIV-free group. With regard to cardiovascular measurements, all brachial blood pressures were lower in PLWH (all P < 0.049); however, no differences were found in either cfPWV, cIMT or PPA (all P ≥ 0.204). When comparing biochemical markers, GGT was higher in those with HIV (P < 0.001), whilst levels of glucose, glycated haemoglobin, TC, triglycerides and LDL-cholesterol (P ≤ 0.049) were lower. The CD4+ count was found to be 525.5 cells/µL for PLWH. Both tobacco use and alcohol consumption (P ≤ 0.014) were more frequent in PLWH.
In Figure 2 and Appendix Table 1   cSBP, cPP and PPA) between the groups with and without HIV in different age tertiles (on the left), as well as across different age groups within the HIV and HIV-free groups, respectively (on the right). PLWH in the younger and middle age groups had lower cSBP (both P ≤ 0.031) and in all the age groups had lower cPP (all P ≤ 0.027) than those without HIV. Carotid-femoral pulse wave velocity, cIMT, cSBP and cPP increased, whilst PPA decreased across increased tertiles of age in both the HIV and HIV-free groups (all P ≤ 0.028).
We further determined whether vascular measures associated with the HIV status in the total group and in the respective age groups (Figure 3 and Appendix

Discussion
This study found that in a population of PLWH in South Africa, brachial blood pressure measurements (cSBP and cPP) were lower in younger and middle age groups and vascular markers (cfPWV, cIMT and PPA) were comparable with the HIV-uninfected population. In addition, we showed a lack of association between HIV infection and any of the vascular markers in any age group.
To the best of our knowledge, this was the first study to investigate age-related differences in the vascular structure and function between South Africans living with HIV and their age and sex-matched controls without HIV. We also determined whether an association between the HIV-positive status and these vascular measures exists within different age groups.
In addressing our first objective, we found that the group with HIV presented with less adverse central and brachial Heart rate (beats/min)  blood pressure profiles compared with their controls. In addition, the age-related deterioration of the vascular profile in the group with HIV was not as pronounced as expected, with cSBP being lower in the younger and middle age groups when compared with the group without HIV. Moreover, cPP was lower in PLWH in all the age groups when compared with their HIV-free counterparts. Data on the vascular profile of PLWH are controversial, and studies on this topic in a SSA context are scant. Previous studies both contradict 35,36,37,38 and correspond 30,39,40,41 to the study findings. Contrary to the study results, an Italian study in age-matched PLWH and individuals without HIV older than 18 years showed higher cSBP in those with HIV, which may have been attributed to renal damage. 35 Msoka and colleagues concluded in a metaanalysis that cIMT, an early marker for atherosclerosis, was higher in PLWH compared with their controls in crosssectional studies. 37 The increased cIMT in the PLWH was associated with elevated levels of CRP and an adverse lipid profile. 37 High CRP levels and an adverse lipid profile commonly occur in immunosuppressive states, such as HIV, where chronic low-grade systemic inflammation is known to prevail. 38 Similar to our results, a narrative review of studies conducted in SSA concluded that treated PLWH had lower SBP and DBP, and hypertension compared with ART-naïve or individuals without HIV. 39 We recently investigated the cardiovascular profile of PLWH and individuals without HIV in the EndoAfrica study, with a mean age of 42 years. 30 The cardiovascular profile of these individuals included vascular markers similar to those in this current study, such as cfPWV, cIMT, cSBP and cPP. 30 Neither the cardiovascular profile nor the CRP levels of those with HIV and participants without HIV in the EndoAfrica study differed, 30 supporting the current findings that EVA was not evident in this study population with HIV. In another cross-sectional study, no increase in arterial stiffness, atherosclerosis or inflammation was found in PLWH, despite the presence of endothelial activation, when compared with ART-naïve or control groups. 41 Considering the effect of ART on metabolic markers, it was reported that the lipid profile of Chinese PLWH who used Efavirenz-based ART was less adverse than those receiving Lopinavir or Ritonavir-based ART. 42 The more favourable lipid profile of the group with HIV was also found in this South African study population using Efavirenz-based ART.

Lifestyle variables
Although there was no obvious effect on LDL-cholesterol in the Chinese study, 42 this study found lower levels of LDLcholesterol, TC and triglycerides in the HIV group (Table 1). This study's less adverse glycaemic profile supports the rising notion that the cardiovascular profile of PLWH is less detrimental compared with their uninfected counterparts. 43 These results support the findings of Shet et al., which showed better treatment outcomes for the HIV subtype C phenotype in South Africa, despite a higher baseline viral load, compared with other subtypes of HIV. 44 Upon investigating whether an association exists between HIV infection and measures of vascular structure and function in different age groups, no significant results were found. In contrast with the study findings, a systematic review of studies conducted in SSA populations showed a positive association between HIV infection and atherosclerosis. 45 It is known that atherosclerosis develops as a result of inflammation, 46 and it is further suggested that the progression thereof is a plausible link to vasculopathy 47 and EVA. 28,48 However, Dillon et al. associated HIV infection with a lower SBP, DBP and BMI in a systematic review and meta-analysis on studies conducted in SSA. 40 In our study population living with HIV, we found a lower cSBP and BMI compared with their counterparts without HIV in all the three age groups. However, we found no associations between cSBP and HIV infection in any of the age groups. The findings from previous studies, which found no association between HIV and arterial stiffness, 49 hypertension 36 or atherosclerosis, 50 are in line with the study findings. Monteiro et al. 49 reported a higher frequency of hypertension in the uninfected controls, which was also evident in this study.
As the carotid-femoral segment of the vasculature is highly sensitive to increases in blood pressure, 51  PWV, pulse wave velocity; IMT, intima-media thickness; SBP, systolic blood pressure; PP, pulse pressure; CI, confidence interval.

FIGURE 3:
Associations between vascular measures and HIV status in the total group and in different age tertile groups. Models that were not significant (carotid IMT -middle and older group; central SBP -older group) were not included. Confounders included in all the models: HIV status, sex, body mass index, total cholesterol, gamma-glutamyl transferase, glycated haemoglobin, C-reactive protein, tobacco use, antiretroviral treatment use and antihypertensive medication use. Mean arterial pressure was additionally included in the models, with cfPWV, cIMT and PPA as the dependent variables. Heart rate was also included in the model, with PPA as the dependent variable.
http://www.sajhivmed.org.za Open Access and hyperglycaemia, 54 none of which was encountered in the group with HIV of whom 76% received ART in this study.
The results on the association of HIV with vascular markers should be interpreted in light of the fact that we did not have ART adherence or viral load data. This cross-sectional study design limited us to infer cause and effect, and the study results need to be confirmed in a larger cohort. To the best of our knowledge, this was the first study to investigate the effect of age on the association between vascular structure and function and HIV infection in a South African population.

Conclusion
The results of this study revealed that HIV-positive status did not associate with measures of vascular structure and function in any of the age groups. People living with HIV did not have a worse cardiovascular profile, and EVA was not evident in these people compared with those without HIV.
The study findings warrant further longitudinal investigation that may observe individual vascular changes over time to determine whether South Africans living with HIV have an increased risk of developing CVD.
A.L. contributed to data acquisition, drafted the manuscript, performed statistical analyses and interpreted the results. C.M.T.F., S.B.-L.R. and Y.B. contributed to conceptualisation and the acquisition, analyses, and interpretation of the data. All authors critically revised the manuscript, gave input and final approval, and agreed to be accountable for all aspects of the work in order to ensure integrity and accuracy.

Funding information
Financial

Data availability
The data sets used and/or analysed during the current study are available from the corresponding author, Y.B., on reasonable request.

Older group
Age range (years)