Endothelial Dysfunction in South African Youth Living With Perinatally Acquired Human Immunodeficiency Virus on Antiretroviral Therapy

Abstract

Background

Human immunodeficiency virus (HIV) and antiretroviral therapy (ART) confer cardiovascular disease (CVD) risk in adults with HIV. Few studies have assessed endothelial dysfunction (ED), an early marker of subclinical CVD risk, in youth living with perinatally acquired HIV (YLPHIV).

Methods

Using peripheral arterial tonometry, we compared ED in YLPHIV and age-matched youth without HIV. A reactive hyperemic index ≤1.35 was defined as ED. Eligible participants included those aged 9–14 years and on ART ≥6 months at enrollment.

Results

Overall, 431 YLPHIV and 93 youth without HIV with a median age of 14.1 versus 13.9 years, respectively, were included. YLPHIV had a lower BMI z score (BMIZ; −0.2 vs 0.4; P < .01) but higher rates of hypercholesterolemia (10% vs 1%; P = .01) than youth without HIV. Among YLPHIV, mean log viral load (VL) was 4.83 copies/mL with 21.7% having a CD4 count <500 cell/mm³; median duration on ART was 9.8 years with 38% initiating at <2 years of age. YLPHIV had higher rates of ED than youth without HIV (50% vs 34%; P = .01); this relationship persisted after adjusting for age, sex, BMIZ, elevated BP, and hypercholesterolemia (RR, 1.43; P = .02). Among YLPHIV, CD4 count >500 cell/mm³ (RR, 1.04; P = .76), VL (RR, 1.01; P = .78), and current ART class (protease inhibitor based vs nonnucleoside inhibitor based: relative risk, 0.90; P = .186) were not associated with ED after adjustment.

Conclusions

Even after adjusting for physiologic differences, YLPHIV appear to be at increased risk of ED compared with age-matched youth without HIV. These findings have important implications for the life course of YLPHIV who may be at increased risk of premature CVD and complications.

Youth living with perinatally acquired human immunodeficiency virus (HIV) appear to be at increased risk of endothelial dysfunction compared with age- and sex-matched youth without HIV in South Africa, suggesting that HIV plays a role in the progression of vascular disease and future atherosclerosis in youth living with perinatally acquired human immunodeficiency virus.

Globally, an estimated 2.1 million adolescents are living with human immunodeficiency virus (HIV) [1], most of whom live in sub-Saharan Africa. Before the availability of antiretroviral therapy (ART) up to 25% of children living with HIV manifested significant cardiac dysfunction including left ventricular systolic dysfunction, left ventricular hypertrophy, and left atrial dilation [2]. The advent of effective ART has dramatically reduced overall mortality, with death now predominantly due to comorbid illnesses [3]. Among these comorbidities, CVD and metabolic diseases are among the leading cause of death in adults living with HIV [3, 4].

Youth living with perinatally acquired HIV infection (YLPHIV) comprise a unique and vulnerable population with potential cardiovascular disease (CVD) risk given their lifetime exposure to both HIV and ART. Evidence of deranged lipids due to ART, upregulated inflammatory pathway, metabolic factors, and immunosenescence play a role in the progression of vascular disease and future atherosclerosis [5, 6]. Although clinical symptoms of atherosclerosis may not manifest until adulthood, the process of atherogenesis begins early [7].

Endothelial dysfunction (ED) is one of the earliest stages of vessel-wall alteration leading to atherosclerosis and consequent CVD [8]. Endothelial dysfunction denotes the inability of the artery to sufficiently dilate in response to an appropriate endothelial stimulus [9]. Since the endothelium is not confined to the coronary arteries, less-invasive techniques can be used to assess peripheral vascular endothelial function [10, 11]. One commonly used method for assessing ED is flow-mediated dilation (FMD) [12]; however, this method is operator dependent and has poor reproducibility [13]. To overcome these challenges, the endothelial peripheral arterial tonometry (EndoPAT) device allows noninvasive measurement of vasoreactivity without the disadvantages of conventional ultrasound measurement while being feasible and demonstrating excellent reproducibility in adolescents [14]. EndoPAT detects plethysmographic pressure changes in the fingertips caused by the arterial pulse and translates these into peripheral arterial tone [15]. It is a noninvasive method that measures the reactive hyperemia index (RHI) induced by cuff occlusion of the arm and has been validated to evaluate peripheral vascular endothelial function [15]. EndoPAT has a sensitivity of 80% and specificity of 85% in identifying individuals with ED [16] as well as a strong correlation with FMD [15, 17, 18].

Few studies have assessed ED in children or youth living with HIV [19–21]. Current published studies have been limited by small sample sizes ranging from 49 to 142, and none have been from sub-Saharan Africa where over 90% of the world’s children living with HIV reside.

Our objective was to compare peripheral endothelial function between YLPHIV and youth without HIV using the EndoPAT technique, with the hypothesis that YLPHIV would have poorer ED due to lifelong exposure to HIV and ART.

METHODS

Study Population

The Cape Town Adolescent Antiretroviral Cohort (CTAAC) is an ongoing South African prospective cohort study investigating the long-term health of YLPHIV receiving ART [22, 23]. Youth living with perinatally acquired HIV infection were enrolled between July 2013 and March 2015 from 7 HIV clinics in Cape Town, South Africa, including 3 tertiary health facilities and 4 primary health clinics. Youth without HIV were healthy adolescents, frequency matched by age and sex, who were recruited from the Youth Center, which has a similar socioeconomic background as the sites where YLPHIV were recruited. Eligible YLPHIV enrolled into CTAAC were 9–14 years of age, on ART for at least 6 months, aware of their HIV status, and had confirmed perinatal HIV acquisition from their clinical record.

Study participants were seen every 6 months at the Medical Research Council Unit on Child and Adolescent Health located at the Red Cross War Memorial Children’s Hospital in Cape Town, South Africa. For this study, we only included those who had an EndoPAT measurement available from the 24-month visit. Written informed consent was obtained from a parent or legal guardian, and study participants provided written informed assent. The study protocol was approved by the Human Research Ethics Committee of the Faculty of Health Sciences, University of Cape Town and Stellenbosch University. Approval for the study was also obtained from the Western Cape Provincial Research Committee.

Primary Outcome

Endothelial dysfunction was measured using an EndoPAT device (Itamar Medical Ltd). EndoPAT is a noninvasive method that assesses reactive hyperemia induced by cuff occlusion of the arm and has been validated to evaluate peripheral vascular endothelial function [15]. In addition, it is not operator dependent and has demonstrated good reproducibility [24]. The test was performed with the participant lying in a comfortable position with both hands at the same level using an arm rest in a silent and temperature-controlled (21°C–24°C) environment with dim light.

Three sets of 5-minute recordings were taken: baseline testing period, cuff occlusion period, and postocclusion period. The RHI was automatically calculated by the device, as a ratio of the postocclusion to preocclusion peripheral arterial tone amplitude of the tested arm, divided by the postocclusion to preocclusion ratio obtained in the control arm [15]. An RHI of 1.35 or less was defined as ED [16].

Covariates

Sociodemographic data were collected from the participant and caregiver at the time of enrollment in the study, and the participant’s clinical record was reviewed by a study clinician to record ART history and World Health Organization (WHO) HIV staging [25] from the time of HIV diagnosis. By convention, WHO staging does not change during the course of illness in patients despite a positive response to treatment.

Clinical examination including Tanner pubertal staging, blood pressure (BP), and anthropometry was performed by a trained member of the research team at 24 months at the time when EndoPAT was performed. Elevated BP was defined as greater than the 90th percentile for age, sex, and height [26]. Body mass index (BMI) was calculated, using WHO references, as weight in kilograms divided by height in meters squared (kg/m²) [27].

Laboratory measures included viral load (VL) using Roche Cobas Ampliprep/COBAS TaqMan HIV-1 Test, version 2.0-standard technique, and CD4 cell count measured by Beckman Coulter FC500 MPL analyzers using the PLG Pan Leukocyte Gating method in YLPHIV. Highly sensitive C-reactive protein (hs-CRP) was measured using the Roche Cobas Tina-quant system. Fasting lipid subfractions including total cholesterol (TC), triglycerides (TGs), high-density lipoprotein (HDL), and low-density lipoprotein (LDL) were measured in all participants at the time when EndoPAT was performed. Abnormal lipids were defined as greater than the 95th percentile using the National Health and Nutrition Examination Survey (NHANES) [28]. All of these tests were conducted at 24 months at the time of EndoPAT.

Statistical Analysis

Baseline variables were compared between groups using t tests, Wilcoxon, chi-square, or Fisher’s exact tests, as appropriate. BMI z score for age (BMIZ) and height-for-age and weight-for-age z scores were calculated using WHO references [29]. Basic characteristics of participants were also compared between those with and without available EndoPAT measurements (Supplementary Tables 1 and 2). Since there were less than 5% missing data and data missing at random, we did not perform imputation for missing data. Modified Poisson regression models were used to assess the adjusted association of perinatally acquired HIV infection with ED using a forward selection approach. In addition, hypercholesterolemia [30, 31], sex [32], and BMIZ [31] were considered a priori confounders. BMI has been proven as an independent risk factor for ED due to its association with metabolic syndrome [33], the impaired production of endothelium-derived nitric oxide under hypercholesterolemia has proven it to be a risk factor for ED [34], and the association of hypertension with impaired production of reactive oxygen species will result in ED [35]. For subgroup analyses among YLPHIV we chose CD4 count, log VL, and current ART as specific predictors of interest where we tested the association of each exposure with ED in separate models. All analyses were performed in Stata, version 14.2 (StataCorp).

RESULTS

Among the 625 adolescents (n = 515 YLPHIV and n = 110 youth without HIV) enrolled in the CTAAC study, EndoPAT was performed at the 24-month visit; 570 (91%) completed the 24-month visit. Out of 570, 524 (92%) (431 [91%] YLPHIV and 93 [95%] youth without HIV) had available EndoPAT data (Figure 1). No differences in age, sex, BMIZ, Tanner stage, or HIV status were found between participants who had completed the 24-month visit versus those did not complete the 24-month visit and no differences were found between those who had EndoPAT data versus those who did not at the 24-month visit (Supplementary Tables 1a and 2a).

Figure 1.

Study population flow chart. Abbreviations: CTAAC, Cape Town Adolescent Antiretroviral Cohort; EndoPAT, endothelial peripheral arterial tonometry; HIV, human immunodeficiency virus; YLPHIV, youth living with perinatally acquired HIV.

Median age (14.1; interquartile range [IQR], 12.8–15.5 years; vs 13.9 [12.1, 15.3] years; P = .21) and sex distribution (female: 49.4% vs 57%) were similar between YLPHIV and youth without HIV as per the study design (Table 1). Median BMIZ was lower in YLPHIV (−0.21 [IQR, −0.95, 0.58] vs 0.39 [IQR, −0.56, 1.26] kg/m², P < .01) but waist to hip ratio was higher in YLPHIV (females: 0.84 [IQR, 0.80, 0.89] vs 0.82 [IQR, 0.78, 0.87], P = .01; males: 0.89 [IQR, 0.86, 0.92] vs 0.85 [IQR, 0.82, 0.89]; P < .01). No differences in Tanner stage were found between groups. Systolic and diastolic BP was also lower in YLPHIV (104 [IQR, 97, 111] vs 108 [IQR, 103, 115] mmHg; P < .01; 67 [IQR, 61, 71] vs 70 [64, 74] mmHg; P = .01, respectively), but no difference in rates of elevated BP were noted between groups. Only 11 (2.1% YLPHIV vs 2.2% youth without HIV, P = .97) reported tobacco use within the last 12 months.

Table 1.

Characteristics of Youth Living With Perinatally Acquired Human Immunodeficiency Virus and Youth Without Human Immunodeficiency Virus

	YLPHIV (n = 431)		Youth Without HIV (n = 93)		P
	n	Values	n	Values
Demographics
Age,a years	431	14.1 (12.8, 15.5)	93	13.9 (12.1, 15.3)	.21
Femaleb	431	213 (49.4)	93	53 (57.0)	.19
Tobacco use in last 12 monthsb	427	9 (2.1)	92	2 (2.2)	.97
Growth measures
Height-for-age z scorea	422	−1.25 (−2.03, −0.55)	89	−0.74 (−1.37, 0.06)	<.01
Weight-for-age z scorea	429	−0.83 (−1.63, −0.03)	92	−0.13 (−0.83, 0.75)	<.01
BMI-for-age z scorea	421	−0.21 (−0.95, 0.58)	89	0.39 (−0.56, 1.26)	<.01
Waist circumference,a cm	428	66 (62, 70)	91	69 (62, 76)	.02
Hip circumference,a cm	427	76 (71, 83)	92	84.5 (73.5, 91.5)	<.01
Mid-thigh circumference,a cm	425	39 (36, 42)	92	43 (38, 47)	<.01
Midupper arm circumference,a cm	429	22 (20, 24)	91	23 (21, 26)	<.01
Waist to hip ratiob
Female	211	0.84 (0.80, 0.89)	51	0.82 (0.78, 0.87)	.01
Male	216	0.89 (0.86, 0.92)	40	0.85 (0.82, 0.89)	<.01
Blood pressure,a mmHg
Systolic	431	104 (97, 111)	93	107.5 (103, 115)	<.01
Diastolic	431	67 (61, 71)	93	69.5 (64, 73.5)	.01
Elevated blood pressureb	431		93
Normal		366 (84.9)		75 (80.7)	.306
High		65 (15.1)		18 (19.4)
Tanner stageb	416		91
1		63 (15.1)		12 (13.2)	.75
2		78 (18.8)		16 (17.6)
3		87 (20.9)		25 (27.5)
4		103 (24.8)		21 (23.1)
5		85 (20.4)		17 (18.7)
Laboratory measures
hs-CRP,a mg/L	422	1.09 (0.5, 3.5)	91	0.69 (0.3, 2.1)	.01
Triglycerides,a mmol/L	417	0.9 (0.7, 1.2)	88	0.6 (0.5, 0.8)	<.01
Hypertriglyceridemiab		54 (17.3)		1 (1.9)	<.01
Total cholesterol,a mmol/L	417	4.0 (3.5, 4.7)	89	3.5 (3.2, 4.1)	<.01
Hypercholesterolemiab		42 (10.1)		1 (1.1)	.01
LDL,a mmol/L	417	2.1 (1.8, 2.7)	87	1.9 (1.5, 2.2)	.01
High LDLb		14 (4.5)		1 (1.9)	.39
HDL,a mmol/L	417	1.4 (1.2, 1.7)	89	1.3 (1.1, 1.7)	.16
Low HDLb		35 (8.4)		9 (10.1)	.60

All continuous variables are expressed as medians (interquartile range) or means (SD) and categorical variables as n (%).

Abbreviations: BMI, body mass index; HDL, high-density-lipoprotein cholesterol; HIV, human immunodeficiency virus; hs-CRP, highly sensitive C-reactive protein; LDL, low-density-lipoprotein cholesterol; YLPHIV, youth living with perinatally acquired HIV.

^aContinuous variable.

^bCategorical variable.

Median hs-CRP (1.1 [IQR, 0.5, 3.5] vs 0.7 [IQR, 0.3, 2.1] mg/L; P = .01), TGs (0.9 [IQR, 0.7, 1.2] vs 0.6 [IQR, 0.5, 0.8] mmol/L; P < .01), TC (4.0 [IQR, 3.5, 4.7] vs 3.5 [IQR, 3.2, 4.1] mmol/L; P < .01), LDL (2.1 [IQR, 1.8, 2.7] vs 1.9 [IQR, 1.5, 2.2] mmol/L; P = .01), and rates of hypertriglyceridemia (17.3% vs 1.9%) and hypercholesterolemia (10.1% vs 1.1%; P = .04) were higher in YLHIV (Table 1).

Among YLPHIV, mean log VL was 4.83 copies/mL, with 21.7% having a CD4 count less than 500 cells/μL. Median duration on ART was 9.8 years, with 38% initiating ART at younger than 2 years of age. There were 347 (84.1%) YLPHIV who had WHO stage III and IV at the time of HIV diagnosis (Table 2).

Table 2.

Human Immunodeficiency Virus Disease Severity Measures Among Youth Living With Perinatally Acquired Human Immunodeficiency Virus

	YLPHIV (n = 431)
Viral load,a,b copies/mL
≤50	259 (62.1)
>50	158 (37.9)
Viral load,b,c copies/mL	50 (50–100)
Log viral loadb,c	4.83 (1.8)
CD4 count,a,d cells/μL
<200	19 (4.5)
200–499	73 (17.2)
500–1000	274 (64.6)
>1000	58 (13.7)
WHO HIV staginga,e
I	27 (6.5)
II	39 (9.4)
III	246 (59.6)
IV	101 (24.5)
Age at initiation of ART,c,f years	4.8 (1.9–7.4)
Age at initiation of ARTa
0–2 years	161 (37.8)
3–5 years	125 (29.3)
6–14 years	140 (32.9)
Duration on ART,c,f years	9.8 (6.8–11.6)
Current ART regimena
2 × NRTI + NNRTI	241 (55.9)
2 × NRTI + PI	140 (32.5)
Others	50 (11.6)

All continuous variables expressed as median (interquartile range) or mean (SD) and categorical variables as number (%).

Abbreviations: ART, antiretroviral therapy; HIV, human immunodeficiency virus; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; SD, standard deviation; WHO, World Health Organization; YLPHIV, youth living with perinatally acquired HIV.

^aCategorical variable.

^b14 missing values.

^cContinuous variable.

^d7 missing values.

^e18 missing values.

^f5 missing values.

Over half (55.9%) of YLPHIV were on a nonnucleoside reverse transcriptase inhibitor (NNRTI)–based ART regimen (96.7% of whom were on efavirenz and 3.3% on nevirapine). Approximately one-third of YLPHIV (32.5%) were receiving protease inhibitors (PIs), all of whom were on lopinavir/ritonavir. All YLPHIV received a backbone of nucleoside reverse transcriptase inhibitors (NRTIs). The majority of YLPHIV were on abacavir or lamivudine (68.5% and 83.5% of participants, respectively); smaller proportions of adolescents were receiving tenofovir, zidovudine, emtricitabine, and stavudine (13.7%, 17.9, 13%, and 2.6%, respectively).

Overall, median RHI was lower among YLPHIV (1.36 vs 1.52; P = .01) compared with youth without HIV. Youth living with perinatally acquired HIV also had higher rates of ED compared with youth without HIV (50% vs 34%; P = .01) (Table 3); this relationship persisted even after adjusting for age, sex, BMIZ, elevated BP, and hypercholesterolemia (relative risk [RR], 1.43; P = .02) (Table 4). Among YLPHIV, CD4 count greater than 500 cell/mm³ (RR, 1.04; P = .76), VL (RR, 1.01; P = .78), and current ART class (PI vs NNRTI-based ART; RR, 0.90; P = .19) were not associated with ED after adjusting each model for age, sex, BMIZ, elevated BP, and hypercholesterolemia (Table 5).

Table 3.

EndoPAT Measures

EndoPAT Measures	YLPHIV (n = 431)	Youth Without HIV (n = 93)	P
RHIa	1.36 (1.09, 1.70)	1.52 (1.23, 1.92)	.01
RHIb
Normal	216 (50.1)	61 (65.6)	.01
Low	215 (49.9)	32 (34.4)
AI at 75 bpma,c	−4 (−14, 11)	−5 (−13, 11)	.86

All continuous variables are expressed as medians (interquartile range) or means (SD) and categorical variables as n (%).

Abbreviations: AI, augmentation index; bpm, beats per minute; EndoPAT, endothelial peripheral arterial tonometry; HIV, human immunodeficiency virus; RHI, reactive hyperemia index; SD, standard deviation; YLPHIV, youth living with perinatally acquired HIV.

^aContinuous variable.

^bCategorical variable.

^c16 missing values for YLPHIV and none for youth without HIV.

Table 4.

Regression Model Evaluating the Unadjusted and Adjusted Association Between Perinatally Acquired Human Immunodeficiency Virus Infection and Endothelial Dysfunction

Variable	Unadjusted Relative Risk (95% CI)	P	Adjusted Relative Risk (95% CI)	P
Youth without HIV	Ref		Ref
Perinatally acquired HIV infection	1.44 (1.08, 1.95)	.01	1.43 (1.06, 1.95)	.02
Age,a per 1-year increment	0.83 (.78, .87)	<.01	.81 (.77, .86)	<.01
Gender
Male	Ref		Ref
Female	0.86 (.72, 1.03)	.10	.90 (.75, 1.08)	.26
Elevated BP
Normal	Ref		Ref
High	0.76 (.57, 1.02)	.07	.70 (.51, .96)	.03
BMI z score,a per 1-unit increase	0.90 (.83, .97)	<.01	.91 (.84, .99)	.03
Total cholesterol
Normal	Ref		Ref
High	1.04 (.76, 1.44)	.78	.87 (.64, 1.19)	.38

Abbreviations: BMI, body mass index; BP, blood pressure, CI, confidence interval; HIV, human immunodeficiency virus; Ref, reference.

^aContinuous variable.

Table 5.

Regression Models Evaluating Unadjusted and Adjusted Associations Between Key Human Immunodeficiency Virus–Related Factors and Endothelial Dysfunction Among Youth Living With Perinatally Acquired Human Immunodeficiency Virus

Model	Unadjusted Relative Risk (95% CI)	P	Adjusted Relative Riska (95% CI)	P
CD4 count, cells/μL
<500	Ref		Ref
≥500	1.09 (.86, 1.40)	.46	1.04 (.81, 1.33)	.76
Log HIV viral loadb	1.01 (.96, 1.06)	.79	1.01 (.96, 1.06)	.78
Current ART regimen
2 × NRTI + NNRTI	Ref		Ref
2 × NRTI + PI	1.05 (.85, 1.28)	.66	.91 (.73, 1.11)	.32
Others	.88 (.63, 1.24)	.47	.82 (.60, 1.10)	.19
WHO HIV staging
I	Ref		Ref
II	1.38 (.78, 2.47)	.27	1.14 (.65, 1.97)	.65
III	1.28 (.77, 2.14)	.37	1.05 (.65, 1.69)	.84
IV	1.55 (.92, 2.61)	.10	1.20 (.74, 1.97)	.46

Abbreviations: ART, antiretroviral therapy; BMI, body mass index; CI, confidence interval; HIV, human immunodeficiency virus; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; Ref, reference; WHO, World Health Organization.

^aEach model separately adjusted for age, gender, BMI z score, elevated blood pressure, and hypercholesterolemia.

^bContinuous variable.

DISCUSSION

This is the first study from Africa that demonstrates the pathogenesis of HIV in altering endothelial function among South African YLPHIV. Results are similar to American and European youth [20, 21], even with different genetic and environmental backgrounds. We found that ED is more prevalent in YLPHIV compared with controls without HIV. The presence of ED in this adolescent cohort without many traditional risk factors for CVD suggests that vascular alterations may already be present from a young age in YLPHIV, representing an early sign of microvascular disease or subclinical atherosclerosis [8].

Similar to our findings, studies in North America and Europe [20, 21] also reported worse endothelial function in YLPHIV compared with youth without HIV. A French study compared endothelial function in YLPHIV on ART versus ART-naive individuals but did not observe differences between these groups [20]. In our study, all YLPHIV were on longstanding ART, with a median duration of 9.8 years. A US study also compared ED between youth with perinatally versus those with nonperinatally acquired HIV and found worse ED among those with perinatally acquired HIV [21]. In addition to pediatrics studies, adult studies have also shown a higher prevalence of ED among individuals living with HIV [36, 37].

Several adult studies have also suggested that HIV infection itself along with ART is associated with both progression of atherosclerosis and cardiovascular events [38, 39]. However, the relative impact of HIV, ART, and underlying CVD risk factor profiles remains unclear. These factors are difficult to disentangle in adults because of the interaction between the presence of classic risk factors for atherosclerotic disease and ART with the fact that the timing of HIV infection is often unknown. Among children and youth living with perinatally acquired HIV several factors are likely to contribute to impaired endothelial function such as in utero and early postnatal exposure to both HIV and/or ART, a critical period of development where cells are particularly vulnerable to mitochondrial toxicity and metabolic alterations [40]. Further, lifelong exposure to HIV and ART for YLPHIV may place YLPHIV at risk of inflammation, subclinical immune activation, and immunosenescence [5, 6], which may play role in altering metabolic pathways within the endothelium. We have previously reported that YLPHIV have accelerated aging [41], and the association of accelerated aging with CVD has been shown by a German case-control study [42]. Due to these risk factors, children and adolescents with perinatal HIV may be more likely to experience earlier or more significant long-term HIV-related cardiometabolic complications than adults living with HIV.

Previous studies have demonstrated the association of PIs with atherosclerosis and CVD in adult living with HIV [43]. Pediatric studies [19, 21] have also reported associations of ED and impaired FMD with PI use. One study from the United Kingdom reported impaired FMD in children living with HIV who were PI treated versus those treated with non–PI ART as well as versus those not treated with ART [19]. However, we did not find any differences in ED between PI-treated and non–PI-treated YLPHIV, similar to observations made by Hsue and colleagues [44]. This suggests that perhaps ED is not solely related to PI use, but rather infection by HIV in and of itself plays a role in ED. In fact, a placebo-controlled study of healthy adults without HIV did not find PIs to induce ED [45], and a Swiss study of adults living with HIV did not find an association between PI use and atherosclerosis [46].

We found that RHI increased with age and BMI similar to a prior publication [47–51]. While the relationship between RHI and age is difficult to explain, Kelly et al [48] have suggested that the association of younger age with lower RHI, which they were not able to observe with FMD, could potentially be due to a 1-size finger probe, which might not be suitable for younger children. Other studies have suggested that microvascular development may not be complete until late adolescence, with a potential role of estrogen and dihydroepiandrosterone (DHEA) in improving endothelial function, and this may be the reason for a paradoxical age–RHI association in our youth/adolescent population [49, 51]. The inverse relation of BMIZ and BP found in our study was most likely because of the more pronounced ED finding among younger participants, as younger age is associated with lower BMIZ and lower BP. Other pediatric studies have also reported that low BP is associated with a higher RHI [48, 50]. In addition, an adult study reported this inverse relationship [52]. The mechanism behind this observation remains elusive due to the lack of our understanding of the relationships between BP and actual microvascular function in children.

Currently, there are no recommendations for routine measurement of ED in YPLHIV. Based on the results from CTAAC, it may be important to monitor YLPHIV in sub-Saharan Africa for early risk factors or signs of cardiometabolic disease as well as ED. Noninvasive techniques, such as EndoPAT technology, provide an easy and reliable measure to assess ED. Early detection of this disorder among YLPHIV may have therapeutic and prognostic implications given that ED is potentially reversible.

Limitations

All participants were well established on ART for an extended period, but results may not be generalizable to those who have a shorter duration of treatment or no treatment. Our study is limited by its cross-sectional nature, and therefore causal inference cannot be established. The physical activity of participants was not recorded, although both YLPHIV and youth without HIV were recruited from communities with similar socioeconomic backgrounds, with an assumption that their exercise profile would be similar. We also did not have comprehensive data on the lifetime ART use or all VL measurements since birth for all participants, which hampered our ability to more accurately estimate the association between lifetime cumulative viremic burden and ED. In addition, the vast majority of participants were black African adolescents, thus limiting the worldwide generalizability of our findings, but they may still be generalizable in the sub-Saharan African context.

Conclusions

This is the first study from Africa that demonstrates the pathogenesis of HIV in altering endothelial function among South African YLPHIV with a comparison group of youth without HIV. Our results show that, even after adjusting for physiologic differences, YLPHIV appear to be at increased risk of ED compared with age- and sex-matched youth without HIV in South Africa, highlighting the role of perinatal HIV in altering endothelial function and potential downstream future CVD risk, even in a population with vastly different genetic, socioeconomic, and environmental background from those in North America and Europe. Further longitudinal studies are required to explore risk and mechanisms for the development of CVD in YLPHIV.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank the adolescents and their caregivers who participated in this study, as well as the study staff for their support of this research. The authors also thank Red Cross War Memorial Children’s Hospital for supporting the study.

Financial support. This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (grant number R01HD074051) and the South African Medical Research Council.

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.