Libyan Journal of Medicine
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/zljm20
Association between cardio-ankle vascular index
and cardiometabolic risk factors in HIV patients in
Ghana
Kwame Yeboah, Samuel Essel, Jennifer Agyekum & Bartholomew Dzudzor
To cite this article: Kwame Yeboah, Samuel Essel, Jennifer Agyekum & Bartholomew
Dzudzor (2023) Association between cardio-ankle vascular index and cardiometabolic
risk factors in HIV patients in Ghana, Libyan Journal of Medicine, 18:1, 2215636, DOI:
10.1080/19932820.2023.2215636
To link to this article:  https://doi.org/10.1080/19932820.2023.2215636
© 2023 The Author(s). Published by Informa View supplementary material 
UK Limited, trading as Taylor & Francis
Group.
Published online: 19 May 2023. Submit your article to this journal 
Article views: 167 View related articles 
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=zljm20
LIBYAN JOURNAL OF MEDICINE
2023, VOL. 18, 2215636
https://doi.org/10.1080/19932820.2023.2215636
ORIGINAL ARTICLE
Association between cardio-ankle vascular index and cardiometabolic risk 
factors in HIV patients in Ghana
Kwame Yeboah a, Samuel Essela,b, Jennifer Agyekum a,c and Bartholomew Dzudzor d
aDepartment of Physiology, University of Ghana Medical School, Accra, Ghana; bDepartment of Physician Assistant Studies, Central 
University, Accra, Ghana; cMedical Laboratory Unit, Mamprobi Hospital, Ghana Health Service, Accra, Ghana; dDepartment of Medical 
Biochemistry, University of Ghana Medical School, Accra, Ghana
ABSTRACT ARTICLE HISTORY
Human immunodeficiency virus (HIV) infection is associated with increased cardiovascular diseases Received 23 January 2023  
(CVDs) even in patients with viral suppression by combination antiretroviral therapy (cART). Arterial Accepted 15 May 2023  
stiffness is an independent predictor of CVDs in diseased individuals and the general population. 
Cardio-ankle vascular index (CAVI) is an index of arterial stiffness that has been shown to predict KEYWORDS 
target organ damage. CAVI is less studied in HIV patients. We compared the levels of arterial Arterial stiffness; CAVI; HIV; cardiometabolic risk factors; 
stiffness using CAVI and associated factors among cART-treated and cART-naïve HIV patients to cART
those of non-HIV controls. In a case-control design, 158 cART-treated HIV patients, 150 cART-naïve 
HIV patients and 156 non-HIV controls were recruited from a periurban hospital. We collected data 
on CVD risk factors, anthropometric characteristics, CAVI, and fasting blood samples to measure 
plasma glucose, lipid profile, and CD4+ cell counts. Metabolic abnormalities were defined using the 
JIS criteria. CAVI increased in cART-treated HIV patients compared to cART-naïve HIV patients and 
non-HIV controls (7.8 ± 1.4 vs 6.6 ± 1.1 vs 6.7 ± 1.4 respectively, p < 0.001). CAVI was associated with 
metabolic syndrome in non-HIV controls [OR (95% CI) = 2.14 (1.04–4.4), p = 0.039] and cART-naïve 
HIV patients [1.47 (1.21–2.38), p = 0.015], but not in cART-treated HIV patients [0.81 (0.52–1.26), p =  
0.353]. In cART-treated HIV patients, a tenofovir (TDF)-based regimen (β = −0.46, p = 0.023) was 
associated with decreased CAVI and decreased CD4+ cell count (β = −0.23, p = 0.047) was asso-
ciated with increased CAVI. In a periurban hospital in Ghana, compared to non-HIV controls or 
cART-naïve HIV patients, cART-treated HIV patients had increased arterial stiffness measured as 
CAVI. CAVI is associated with metabolic abnormalities in non-HIV controls and cART-naïve HIV 
patients, but not in cART-treated HIV patients. Patients on TDF-based regimens had decreased CAVI.
1. Introduction The cardio-ankle vascular index (CAVI) is a maker 
The global infection rate and burden of the human of arterial stiffness derived from the stiffness para-
immunodeficiency virus (HIV) have reduced; however, meter β, which is reported to have less influence by 
sub-Saharan Africa is responsible for more than two- fluctuation in blood pressure [6]. This makes CAVI 
thirds of the new global HIV infections [1]. Fortunately, unique from other arterial stiffness that is based on 
due to the widespread availability and access to combi- pulse wave velocity (PWV), such as carotid-femoral 
nation antiretroviral therapy (cART), even in poor PWV and brachial-ankle PWV [6,7]. CAVI has been 
resource settings in sub-Saharan Africa, the burden of reported in several studies to be associated with 
AIDS-related morbidity and mortality has been mark- cardiovascular-related organ damage [8,9] and mor-
edly reduced [2,3]. However, there is still high morbidity tality [10]. Several studies conducted in Africa and 
and mortality in HIV patients compared to non-HIV other parts of the world have reported arterial stiff-
individuals, mainly due to chronic diseases such as car- ness in HIV patients using PWV [2,11–14]. We found 
diovascular disease (CVD) [3]. HIV replication is asso- only two studies that have recently reported CAVI 
ciated with immune activation, particularly activation in the Thai HIV population [15,16], and none in sub- 
of CD4 and CD8 T cells, which leads to increased sys- Saharan Africa. There is racial variation of arterial 
temic subclinical inflammation [4]. This may lead to the stiffness across lifespan, with blacks reported to 
formation of atherosclerotic plaques and/or stiffening of have high levels of arterial stiffness compared to 
the large arteries. Arterial stiffness of the medium and Caucasians and Asians at any given age range [17]. 
large arteries has been associated with the future devel- In addition, environmental factors such as diet, 
opment of CVD in diseased patients and the general levels of physical activity and psychosocial para-
population [5]. meters can affect arterial stiffness across different
CONTACT Kwame Yeboah kyeboah@ug.edu.gh Department of Physiology, University of Ghana Medical School, Accra, Ghana
Supplemental data for this article can be accessed online at https://doi.org/10.1080/19932820.2023.2215636.
© 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. 
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), 
which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article 
has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent.
2 K. YEBOAH ET AL.
sociodemographic regions [2,14,17]. There is also with a non-elastic tape measure parallel to the 
conflicting report as to whether HIV infection and/ floor without compressing the skin. The percentage 
or ART medication lead to increase in arterial stiff- of body fat was estimated through a bioelectrical 
ness [2]. Therefore, we compared arterial stiffness impedance analysis with the Body composition 
using CAVI between cART-treated and cART-naïve monitor (BF- 508, Omron Healthcare, Inc., Vernon 
HIV patients with those of non-HIV controls. We Hills, IL, USA). Blood pressure (BP) was measured 
also investigated factors associated with changes using a semi-automated blood pressure monitor 
in CAVI among study participants. We hypothesize (average of two measures for each arm at 1-min 
that, compared to non-HIV controls, individuals with intervals). Hypertension was defined as those hav-
HIV infection and cART medication may increase ing systolic BP ≥140 mmHg and/or diastolic BP ≥90  
arterial stiffness measured as CAVI. mmHg or taking antihypertensive treatment.
We collected 5 ml of fasting venous blood sam-
ples from each participant in appropriate tubes, 
2. Methods centrifuged them at 4000 G, and serum/plasma 
2.1. Study participants, site and design was aliquoted and stored at −80°C until analysis. 
Fasting plasma glucose (FPG), total cholesterol, 
This study was conducted from October 2019 through high-density lipoprotein cholesterol and plasma tri-
February 2020, and the design was a case-control study glyceride levels were analysed using a biochemistry 
with HIV patients as cases and the controls were non- analyser (Contec BC 400, China) and commercial 
HIV individuals who visited the facility for voluntary reagents (Randox Laboratory Reagents, UK). Low- 
testing of their HIV status. HIV patients were classified density lipoprotein (LDL) cholesterol levels were 
as those under cART treatment (cART-treated) and calculated using Friedewald’s formula. CD4 cell 
newly diagnosed patients who had not yet received count was measured using TriTEST reagents follow-
cART treatment (cART-naïve). The study was conducted ing a dual platform protocol and MultiSET and 
at Atua Government Hospital, a 150-bed primary health- Attractors software using a FACScan flow cytometer 
care facility, located at the Agormanya, a periurban (Becton-Dickinson, NJ, USA). Metabolic syndrome 
town in the Eastern region of Ghana. The Agormanya was defined by the joint interim statement criteria 
area has a high prevalence of HIV infection compared to [18] as having three or more of the following: (1) 
the national prevalence. The hospital has approximately abdominal obesity (waist circumference ≥94 cm for 
3000 HIV patients on its registry. Participants with men & ≥ 80 cm for women); (2) high triglycerides 
a diagnosis of diabetes or fasting blood glucose (FPG) ≥1.7 mmol/L; (3) low HDL cholesterol: men <1.0  
> 7 mmol/L, a history of cardiovascular disease or treat- mmol/L or women <1.3 mmol/L; and (4) High BP 
ment and those with an ankle-brachial index <0.9 were (systolic BP ≥130 mmHg and/or diastolic BP ≥85  
excluded from the study. Ethical approval was obtained mmHg); and (5) fasting plasma glucose (FPG) ≥5.6  
from the College of Health Sciences Ethical & Protocol mmol/l.
Review Committee (CHS-Et/M.6–5.17/2018–2019) and CAVI was measured using Vasera 1500N (Fukuda- 
all participants provided voluntary informed consent Denshi, Japan) with the participant resting supine for at 
before joining the study. least 10 minutes before measurement. Electrocardiogram 
electrodes were placed on both wrists, a microphone for 
detecting heart sounds was placed on the sternum, and 
2.2. Data collection
cuffs were wrapped around the upper arms and ankles. 
A structured questionnaire was used to obtain data CAVI values were computed automatically. Briefly, CAVI 
on sociodemographic factors like age, gender, lifestyle corresponds to the stiffness parameter β, calculated from 
factors (smoking, alcohol intake), medical history the values of heart-ankle PWV and BP as follows: 
(hypertension, diabetes, cardiovascular disease), cur-
rent medication (antihypertensive agents, cART), β = (2ρ/ΔP)� [ln (Ps/Pd)]�PWV
2 
occupation, education (school cycles completion), 
marital status. Smoking status was classified as Where ρ indicates blood density; ΔP, pulse pressure; 
never, past (smoking cessation since more than 1  ln, natural log; Ps, systolic BP; and Pd, diastolic BP 
year before the survey) or current smoking. [19,20].
Body weight and height were measured using 
a stadiometer in light clothing with footwear 
removed, with a body-mass index (BMI) calculated 2.3. Sample size calculation
as weight/height2 and categorized as underweight 
(<18.5 kg/m2), normal weight (18.5–24.9 kg/m2), Sample size was computed based on the data from 
overweight (25–29.9 kg/m2), and obese (≥30 kg/ a previous pilot study [21]. A minimum of 133 partici-
m2). Waist and hip circumferences were measured pants was required in each patient group to detect an
LIBYAN JOURNAL OF MEDICINE 3
effect size ≥0.4, with a power of 90% and a significance 3. Results
level of 95%.
3.1. General characteristics of study participants
The mean age of the study participants was 38.4 ±  
2.4. Data analysis 13.7 years with two-thirds being females. There was 
no difference in mean age among various categories 
IBM SPSS version 28 was used to summarise results of participants. There was a high proportion of HIV 
as proportions for categorical variables and means patients who were hypertensive, overweight, and cur-
and standard deviations (SD) for continuous vari- rently or previously smoked. Compared to non-HV 
ables. Shapiro Wilk test was used to check the participants and cART-naïve HIV patients, HIV patients 
normality of the quantitative variables, and non- on cART treatment had higher waist circumference, 
normally distributed variables were transformed waist-hip ratio, percentage of body fat, mean diastolic 
appropriately. Mean differences between groups of and blood pressures, and heart rate. HIV patients on 
patients were analysed by ANOVA and ANCOVA to cART had higher levels of fasting plasma glucose 
adjust for covariates. The distribution of categorical (FPG), triglycerides, and total and LDL cholesterol 
data was analysed by χ2 test, and association compared to non-HIV participants. cART-treated HIV 
between variables using Pearson’s correlation. patients had higher levels of CAVI and MetS com-
Multiple regressions, with all appropriate para- pared to cART-naïve HIV patients or non-HIV partici-
meters forced into the model, were performed to pants, but no difference between cART-naïve and 
determine independent association between non-HIV participants (Table 1). The average duration 
patients’ characteristics and CAVI. Logistic regres- of HIV infection in cART-treated HIV patients was 7.6  
sion analyses were performed to determine the ± 4.6 years and the average duration of cART treat-
association between cardiometabolic abnormalities ment was 7.2 ± 4.5 years. For the cART medication 
and CAVI. p-values <0.05 were considered statisti- regimen, 94 (59.5%) patients were treated with TDF/ 
cally significant. 3TC/NVP or EFV regimens, 52 (32.9%) patients were
Table 1. General characteristics of study participants.
All Participants Non-HIV controls cART-naïve HIV patients cART-treated HIV patients p
N 464 156 150 158
Age, years 38.4 ± 13.7 36.7 ± 14.4 38.2 ± 11.6 39 ± 11.4 0.109
Females, n (%) 312 (67.2) 106 (67.9) 84 (56) 122 (77.2) 0.02
Married, n (%) 198 (42.7) 70 (44.9) 62 (41.3) 66 (41.8) 0.79
Smoking, n (%) 0.029
Current 16 (3.4) 2 (1.3) 4 (2.7) 10 (6.3)
Former 57 (15.9) 9 (5.8) 22 (14.6) 26 (16.5)
Never 187 (80.6) 71 (92.9) 124 (82.7) 132 (83.5)
Alcohol intake, n (%) 102 (22) 38 (24.4) 36 (24) 28 (17.7) 0.55
Waist circumference, cm 87 ± 12 85 ± 11 84 ± 11 90 ± 12#* 0.002
Hip circumference, cm 102 ± 11 103 ± 11 99 ± 11* 103 ± 11 0.074
Body height, cm 164 ± 7 164 ± 8 164 ± 7 162 ± 7 0.194
Waist-hip ratio 0.85 ± 0.08 0.82 ± 0.08 0.84 ± 0.07* 0.88 ± 0.08*# <0.001
Body weight, kg 60 ± 13.2 68 ± 13.3 64.4 ± 7.3 65.8 ± 14 0.204
BMI, kg/m2 24.8 ± 5 25.4 ± 4.7 23.7 ± 4.4* 25.3 ± 5.7 0.061
BMI categories, n (%) 0.024
Underweight 34 (7.4) 6 (3.9) 16 (10.7) 12 (7.6)
Normal 238 (51.5) 70 (45.5) 94 (62.7) 74 (46.8)
Overweight 116 (25.1) 46 (29.9) 18 (12) 52 (32.9)
Obese 74 (16) 32 (20.8) 22 (14.7) 20 (12.7)
Body fat, % 31 ± 12.2 32.4 ± 12.5 27.6 ± 11.8* 32.8 ± 11.6*# 0.014
Systolic BP, mmHg 134 ± 18 132 ± 13 133 ± 19 137 ± 22 0.184
Diastolic BP, mmHg 83 ± 11 80 ± 9 83 ± 12 86 ± 12*# 0.008
Mean BP, mmHg 100 ± 14 98 ± 10 99 ± 14 104 ± 16*# 0.007
Pulse BP, mmHg 51 ± 13 52 ± 10 50 ± 10 51 ± 13 0.672
Heart rate, bpm 74 ± 9 72 ± 8 73 ± 9 80 ± 8*# <0.001
Hypertension, n (%) 162 (34.9) 48 (30.8) 46 (30.7) 68 (43) 0.031
FPG, mmol/l 5.2 ± 0.8 5 ± 0.9 5.1 ± 0.8 5.6 ± 0.8* <0.001
Triglycerides, mmol/l 1.4 ± 0.4 1.4 ± 0.3 1.4 ± 0.4 1.6 ± 0.4* <0.001
Total cholesterol, mmol/l 5.1 ± 1.2 4.8 ± 1.2 5 ± 1.1 5.6 ± 1.1* <0.001
HDL cholesterol, mmol/l 1.5 ± 0.4 1.6 ± 0.4 1.4 ± 0.4 1.5 ± 0.5 0.128
LDL cholesterol mmol/l 3 ± 0.9 2.6 ± 0.9 3 ± 0.8* 3.3 ± 0.7*# <0.001
MetS 178 (38.4) 34 (21.8) 42 (28) 102 (64.6) <0.001
Current CD4 count, cells/mm2 405 (273–562) 430 (327–534) 403 (253–583) 0.804
CAVI 7.1 ± 1.4 6.6 ± 1.1 6.7 ± 1.4 7.8 ± 1.4*# <0.001
Note: BMI, body mass index; BP, blood pressure; FPG, fasting blood glucose; HDL, high density lipoprotein; LDL, low density lipoprotein; CAVI, cardio- 
ankle vascular index; MetS, metabolic syndrome. 
*p < 0.05 compared to non-HIV controls. 
#p < 0.05 compared to cART-naïve HIV patients. 
4 K. YEBOAH ET AL.
on AZT/3TC/NVP or EFV regimens and 12 (7.6%) CAVI than normal BMI participants (Figure 1). Similar 
patients were on LPV/r-based regimens. observations were made when the mean CAVI values 
were adjusted for covariates in ANCOVA analysis 
(Supplementary, S1). When HIV patients were classi-
3.2. Comparison of CAVI among study fied according to their CD4+ cell count, CAVI was 
participants highest in patients with CD4+ cell count <200 cell/ 
mm3, followed by those with CD4+ cell count 200– 
There was no difference in the levels of CAVI between 500 cells/mm3, and those with CD4+ cell count >500 
male and female participants (7.1 ± 1.3 vs 7 ± 1.4, p =  cells/mm3 had the lowest CAVI (Figure 2). Similar 
0.883, respectively). In non-HIV controls (6.8 ± 1.1 vs observations were made when adjusted for covariates 
6.3 ± 1.1, p = 0.023) and cART-naïve HIV patients (7.1 ±  (Supplementary, S2).
1.5 vs 6.5 ± 1.3, p = 0.031), CAVI was higher in those 
with MetS compared to those without MetS, but no 
differences in CAVI was observed between cART- 
3.3. Correlation between CAVI and participants 
treated HIV patients with and without MetS (7.7 ± 1.4 
characteristics
vs 8 ± 1.4, p = 0.141). When CAVI levels were com-
pared among various categories of BMI in all study In all study participants, CAVI was positively correlated 
participants, the obese group had lower CAVI than the with age, waist-hip ratio, BPs, heart rate, FPG, and 
overweight and normal groups, with the underweight total and LDL cholesterol levels; negatively correlated 
group having the highest CAVI. In non-HIV controls, with body weight, hip circumference, BMI, body fat, 
the underweight group had higher CAVI compared to and CD4+ cell count. In non-HIV controls, CAVI was 
the other BMI groups. In cART- naïve and cART- positively correlated with age, BPs, total and LDL 
treated HIV patients, obese participants had a lower cholesterol, and negatively correlated with hip
12
normal overweight obese underweight
10
8
6
4
2
0
All participants Non-HIV controls cART-naïve HIV cART-treated HIV
patients patients
Study participants
Figure 1. Levels of CAVI among various BMI categories in study participants.
12
10
8
6
4
2
0
<200 200 - 500 500+
CD4 cell count levels in HIV patients 
Figure 2. Levels of CAVI levels among CD4+ cell count categories in HIV patients.
CAVI
CAVI
LIBYAN JOURNAL OF MEDICINE 5
Table 2. Correlation between CAVI and characteristics of study participants.
All participants Non-HIV controls cART-naïve HIV patients cART-treated HIV patients
r p r p r p r p
Age 0.72 <0.001 0.72 <0.001 0.67 <0.001 0.67 <0.001
Weight −0.23 <0.001 −0.1 0.242 −0.41 <0.001 −0.25 0.003
Height −0.05 0.344 −0.02 0.852 −0.21 0.02 0.18 0.033
Waist circumference 0.05 0.315 0.14 0.118 −0.14 0.146 −0.11 0.201
Hip circumference −0.23 <0.001 −0.17 0.049 −0.32 <0.001 −0.31 <0.001
WHR 0.33 <0.001 0.35 0.25 0.007 0.2 0.021
BMI −0.21 <0.001 −0.08 0.339 −0.33 <0.001 −0.33 <0.001
Body fat −0.15 <0.001 −0.08 0.342 −0.29 0.002 −0.27 0.002
Systolic BP 0.3 <0.001 0.36 <0.001 0.2 0.027 0.17 0.055
Diastolic BP 0.35 <0.001 0.47 <0.001 0.33 <0.001 0.23 0.006
Mean BP 0.36 <0.001 0.46 <0.001 0.28 0.003 0.23 0.008
Pulse BP 0.05 0.142 0.06 0.474 −0.04 0.704 0.06 0.503
Heart rate 0.13 0.01 0.14 0.114 0.14 0.123 −0.17 0.052
FPG 0.15 0.003 −0.02 0.809 0.21 0.025 −0.03 0.69
Total cholesterol 0.14 0.006 0.19 0.027 −0.01 0.591 −0.05 0.535
Triglycerides 0.09 0.094 0.08 0.373 0.05 0.56 −0.15 0.084
HDL cholesterol −0.01 0.885 0.05 0.55 0.03 0.737 −0.14 0.097
LDL cholesterol 0.19 <0.001 0.23 0.007 −0.09 0.322 0.06 0.498
CD4 cell count −0.45 <0.001 −0.69 <0.001 −0.38 <0.001
Duration of HIV infection 0.32 <0.001
cART treatment duration 0.31 <0.001
Note: WHR, waist-hip ratio; BMI, body mass index; BP, blood pressure; FPG, fasting blood glucose; HDL, high-density lipoprotein; LDL, low-density 
lipoprotein; cART, combination antiretroviral therapy. 
circumference. In cART-naïve HIV patients, CAVI was models, CAVI was significantly associated with 
positively correlated with age, waist-hip ratio, blood increased age, systolic BP, and cART-treated HIV 
pressure indices and fasting blood glucose, and nega- patients compared to non-HIV controls and smokers 
tively correlated with body weight and height, hip compared to non-smokers while CAVI decreased 
circumference, BMI, body fat and CD4+ cell count. In with increasing BMI. In non-HIV controls, CAVI was 
cART-treated HIV patients, CAVI was positively corre- associated with increasing age and systolic BP, as 
lated with age, body height, waist-hip ratio, systolic well as decreasing BMI. In cART-naïve HIV patients, 
and diastolic BP, duration of HIV infection, and cART CAVI was associated with an increase in age and 
treatment; and negatively correlated with body systolic BP while CAVI decreased with an increase 
weight, hip circumference, BMI, body fat, and CD4+ in body height, BMI, and CD4+ cell count. In cART- 
cell count (Table 2). treated HIV patients, CAVI was associated with 
increased age and systolic BP, and decreased body 
height, BMI, triglycerides, and CD4 cell count. CAVI 
3.4. Determinant of CAVI from multiple linear 
decreased in patients on the TDF-based regimen 
regression
compared to the AZT-based regimen (Table 3).
In multivariate regression analyses among all study Association between CAVI and cardiometabolic 
participants with all parameters forced into the abnormalities
Table 3. CAVI Determinants from multivariate linear regression analyses in various Study Groups.
All participants Non-HIV controls cART-naïve HIV patients cART-treated HIV patients
B±SE β p B±SE β p B±SE β p B±SE β p
HIV status (Reference: Non-HIV participants)
cART-naïve −0.09 ± 0.11 −0.06 0.409
cART-treated 0.23 ± 0.12 0.167 0.039
Age 0.06 ± 0.01 0.64 <0.001 0.05 ± 0.01 0.7 <0.001 0.05 ± 0.01 0.41 <0.001 0.07 ± 0.01 0.55 <0.001
Female gender 0.04 ± 0.12 0.03 0.767 0.01 ± 0.22 0.01 0.975 0.09 ± 0.18 0.07 0.617 0.31 ± 0.33 0.22 0.364
Alcohol usage 0.18 ± 0.11 0.13 0.08 0.18 ± 0.14 0.17 0.194 0.17 ± 0.18 0.13 0.345 0.3 ± 0.3 0.21 0.33
Current smoking 0.48 ± 0.23 0.35 0.039 −0.45 ± 0.49 −0.41 0.367 0.77 ± 0.41 0.57 0.061 0.57 ± 0.53 0.41 0.288
Systolic BP 0.02 ± 0.01 0.22 <0.001 0.02 ± 0.01 0.23 0.004 0.01 ± 0.01 0.15 0.01 0.01 ± 0.01 0.23 0.042
Body height −0.01 ± 0.01 −0.07 0.071 −0.01 ± 0.01 −0.06 0.526 −0.05 ± 0.01 −0.25 <0.001 0.01 ± 0.02 0.07 0.502
BMI −0.1 ± 0.01 −0.39 <0.001 −0.09 ± 0.02 −0.41 <0.001 −0.08 ± 0.02 −0.27 <0.001 −0.1 ± 0.02 −0.44 <0.01
Total cholesterol 0.05 ± 0.05 0.04 0.361 0.13 ± 0.07 0.14 0.098 −0.11 ± 0.08 −0.09 0.193 0.26 ± 0.16 0.21 0.113
Triglycerides −0.16 ± 0.16 −0.04 0.324 −0.43 ± 0.27 −0.12 0.111 0.22 ± 0.24 0.06 0.367 −0.96 ± 0.44 −0.26 0.034
FPG −0.03 ± 0.05 −0.01 0.819 0.03 ± 0.07 0.02 0.714 −0.13 ± 0.1 −0.07 0.196 −0.09 ± 0.18 −0.05 0.631
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3 CD4count −0.58 ± 0.09 −0.41 <0.001 −0.23 ± 0.12 −0.23 0.047
cART regimen (Reference: AZT/3TC/NVP or EFV)
TDF/3TC/NVP or EFV −0.64 ± 0.28 −0.46 0.023
LPV/r-based 0.32 ± 0.42 0.23 0.45
Note: BMI, body mass index; BP, blood pressure; FPG, fasting blood glucose; cART, combination antiretroviral therapy; AZT, zidovudine; 3TC, Lamivudine; 
TDF, tenofovir; NVP, nevirapine; EVF, efavirenz; LVP/r, Lopinavir/ritonavir. All parameters were forced into the linear regression model. 
6 K. YEBOAH ET AL.
The association between CAVI and cardiometabolic regimen was associated with decreased CAVI and 
abnormalities were examined using logistic regression decreased CD4+ cell count was associated with 
analyses. In all participants, a unit change in CAVI was increased CAVI.
associated with increased odds of impaired fasting In this study, we found that CAVI increases in HIV 
glucose, high systolic BP, high triglycerides, and patients treated with cART, but is similar among cART- 
MetS in the unadjusted model. After adjustments, naïve HIV patients and non-HIV controls. CAVI is 
high systolic BP and MetS remained associated with a sensitive marker of arterial stiffness and several studies 
CAVI. In non-HIV controls, a unit change in CAVI was have shown that it predicts CVD in patients and the 
associated with increased odds of MetS in unadjusted general population [9,10]. From our literature search, 
and adjusted models. In cART-naïve HIV patients, CAVI we found just two studies that reported CAVI in HIV 
was associated with high systolic BP, high triglycer- patients. These studies were carried out in the Thai 
ides and MetS in unadjusted and adjusted models. In population and reported no differences in CAVI 
cART-treated HIV patients, only high systolic BP was between HIV patients and non-HIV controls [15,16], 
associated with CAVI in unadjusted and adjusted which contrasts with what we observed in our study 
models (Table 4). population. Previous studies conducted in the sub- 
Saharan African population have reported arterial stiff-
4. Discussion ness in HIV patients using PWV. Similar to the findings of 
our study, Msoka et al. reported higher levels of arterial 
The main findings from this study are: CAVI increased stiffness, measured as aortic PWV, in Tanzanian HIV 
in cART-treated HIV patients compared to cART-naïve patients compared to controls [22]. A similar finding of 
HIV patients and non-HIV controls. In study partici- elevated arterial stiffness was reported in Cameroonian 
pants, regardless of HIV or cART status, CAVI was HIV patients using carotid-femoral PWV [14]. In contrast 
associated with increased age and systolic BP, as to our findings, studies conducted in South Africa [4,11], 
well as decreased BMI. CAVI was associated with Cameroon [23] and Ethiopia [12] reported similar levels 
metabolic syndrome in non-HIV controls and cART- of arterial stiffness between HIV patients and non-HIV 
naïve HIV patients, but not in cART-treated HIV controls. Recent metanalysis reported that arterial stiff-
patients. In cART-treated HIV patients, the TDF-based ness measured as carotid-femoral PWV is increased in
Table 4. Association between CAVI and metabolic abnormalities from logistics regression models.
Unadjusted model Multivariable adjusted model
OR (95% CI) p OR (95% CI) p
All study participants
Impaired fasting glucose 1.24 (1.08–1.42) 0.002 1.09 (0.89–1.34) 0.393
High systolic BP 1.46 (1.26–1.69) <0.001 1.69 (1.33–2.15) <0.001
Abdominal obesity 1.1 (0.96–1.25) 0.117 1.06 (0.75–1.51) 0.749
Low HDL 1.01 (0.88–1.17) 0.846 1.14 (0.92–1.41) 0.221
High triglycerides 1.25 (1.08–1.44) 0.002 1.18 (0.92–1.51) 0.19
Metabolic syndrome 1.27 (1.11–1.46) <0.001 1.19 (1.09–1.53) 0.013
Non-HIV controls
Impaired fasting glucose 1.02 (0.74–1.39) 0.923 1.34 (0.75–2.38) 0.326
High systolic BP 1.85 (1.3–2.61) <0.001 1.62 (0.76–3.47) 0.214
Abdominal obesity 1.01 (0.75–1.36) 0.95 0.68 (0.22–2.13) 0.51
Low HDL 0.81 (0.56–1.16) 0.254 1.62 (0.85–3.07) 0.142
High triglycerides 0.9 (0.61–1.33) 0.604 0.7 (0.34–1.45) 0.334
Metabolic syndrome 1.19 (1.07–1.56) 0.022 2.14 (1.04–4.4) 0.039
cART-naïve HIV patients
Impaired fasting glucose 1.44 (1.11–1.85) 0.005 1.33 (0.88–2) 0.176
High systolic BP 1.61 (1.23–2.12) <0.001 1.84 (1.24–2.72) 0.002
Abdominal obesity 1.04 (0.83–1.31) 0.721 1.75 (0.97–3.15) 0.063
Low HDL 0.88 (0.68–1.15) 0.354 0.91 (0.59–1.39) 0.656
High triglycerides 1.22 (1.09–1.62) 0.016 1.45 (1.12–2.54) 0.029
Metabolic syndrome 1.26 (1.07–1.63) 0.031 1.47 (1.21–2.38) 0.015
cART-treated HIV patients
Impaired fasting glucose 0.87 (0.69–1.1) 0.248 1.15 (0.76–1.76) 0.503
High systolic BP 1.36 (1.06–1.73) 0.014 1.51 (1.01–2.24) 0.044
Abdominal obesity 0.91 (0.71–1.16) 0.451 0.52 (0.26–1.07) 0.075
Low HDL 0.99 (0.79–1.25) 0.924 1.05 (0.69–1.59) 0.826
High triglycerides 0.96 (0.77–1.21) 0.751 1.06 (0.7–1.6) 0.791
Metabolic syndrome 0.84 (0.66–1.06) 0.143 0.81 (0.52–1.26) 0.353
Note: IFG, impaired fasting glucose; BP, blood pressure; HDL, high-density lipoprotein cholesterol; cART, combination 
antiretroviral therapy. 
The multivariable models were adjusted for sex, age, BMI, alcohol intake, smoking status, educational level, marital status, 
employment status, CD4 cell count (only in cART-naïve and cART-exposed HIV patients) and cART regimen (only in cART- 
exposed patients). The duration of HIV infection and cART management were excluded from the model due to their low 
tolerance and high valence inflation factor in model diagnostics. 
LIBYAN JOURNAL OF MEDICINE 7
cART-treated HIV patients compared to cART-naïve HIV body fat and protein stores [31]. Previous studies 
patients and non-HIV controls [2], confirming the find- reported that early initiation of cART medication pre-
ings of our study. vented the cachectic and wasted state that charac-
In this study, we found that age and systolic BP terised HIV infection, resulting in suppression of viral 
increase CAVI regardless of HIV or cART status. The replication, a significant reduction in inflammation, 
influence of age on CAVI is similar to what had been and normalisation of CD4+ cell count [32]. This is 
reported in previous studies of patients and healthy likewise supported by what we observed in this 
populations [7,24]. Ageing is associated with several study that HIV patients with low CD4+ cell count 
mechanisms that might have resulted in increased had higher CAVI compared to those with normal 
arterial stiffness, such as fatigue and fracture after levels. Therefore, there is the possibility that weight 
long-standing arterial pulsation in the central arteries gain and high BMI may represent improved cardio-
[25], increased arterial calcification and endothelial vascular health in HIV patients. However, we also 
dysfunction [25,26], and accumulation of advanced observed a lower CAVI in non-HIV obese controls, 
glycated end-products in vascular matrix proteins which implies that further studies are needed to 
[26]. We also found that total plasma cholesterol and investigate the relationship between BMI and arterial 
LDL cholesterol correlated with CAVI in non-HIV con- stiffness in Africans.
trols but not in HIV patients. This is in agreement with Multivariate analysis in cART-treated HIV patients 
the findings of a study by Soska et al. that reported showed that the TDF-based regimen was associated 
a weak relationship between CAVI and total choles- with a decrease in CAVI compared to those on the AZT- 
terol and LDL in non-HIV participants [27]. HIV infec- based regimen. The AZT-based regimen has been 
tion and the use of cART likely affect lipid metabolism, shown to cause oxidative damage, a major mechanistic 
resulting in inflammation, and this may blunt the pathway of arterial stiffness, through the induction of 
relationship between arterial stiffness and plasma mitochondrial dysfunction by inhibition of DNA poly-
cholesterol. However, previous studies in the merase-γ activity [33]. Furthermore, a metabolic study 
Ghanaian diabetic population failed to find any asso- showed that the pathway through which the AZT-based 
ciation between parameters of lipid profile and CAVI regimen leads to oxidative stress is by altering the 
because most diabetic patients were on lipid-lowering metabolism of glutamine and glutamate, glutathione, 
treatment [24,28]. arginine biosynthesis, as well as the metabolism of ala-
We also found that the presence of MetS was nine, aspartate and glutamate [34]. Interestingly, it has 
associated with higher CAVI in non-HIV controls and been demonstrated in a meta-analysis that the TDF- 
cART-naïve HIV patients, but not in cART-treated HIV based regimen has better viral suppression tolerability 
patients. This contrast to the findings of Msoka et al compared to the AZT-based regimen [35,36], making it 
who reported that arterial stiffness was similar in non- suitable for treatment in HIV patients.
HIV controls and cART-naïve HIV patients with and 
without MetS, but higher in cART-treated HIV patients 
4.1. Limitations of the study
with MetS [22]. The contradiction between our find-
ings and that of Msoka et al may be attributed to the The uniqueness of this study is that it is the first to 
different methods used to assess arterial stiffness and report on CAVI in sub-Saharan HIV population. The 
the heterogeneity of the populations studied; we did limitation of this study is that it was carried out in 
not use the strict hepatic, renal and haematological a single healthcare facility and therefore the levels of 
screening protocol employed in the Msoka et al study arterial stiffness measured as CAVI cannot be general-
and we also included HIV patients irrespective of their ised to the entire Ghanaian HIV population. CAVI was 
CD4+ cell count. The findings of our study imply that measured at one time, and therefore we cannot infer 
in non-HIV controls and cART-naïve HIV patients, the causality between HIV infection and CAVI; some HIV 
presence of metabolic abnormalities may partially patients may have increased arterial stiffness from 
explain the increased arterial stiffness as reported in other causes before HIV infection. The sample size 
the Caucasian population [29,30]. was not large enough for us to make a definite con-
In this study, obese individuals had lower CAVI clusion about the effect of various cART regimens in 
compared to those with normal BMI, and BMI was CAVI. It is recommended that future studies use 
negatively associated with CAVI in regression analysis. a multicentre longitudinal design to investigate the 
Similar findings were reported in HIV patients in South development of arterial stiffness and the effect of 
Africa, where low BMI was associated with increased cART medication in a large number of patients to 
arterial stiffness measured as carotid-femoral PWV confirm or refute the findings of the study. 
[13]. This observation may be explained by the phe- Furthermore, other arterial stiffness indices can be 
nomenon of ‘return to health’ which describes desir- used to compare their agreement with CAVI as 
able weight gain after the resolution of a debilitating a marker of arterial stiffness in the Ghanaian HIV 
catabolic illness such as HIV infection that restores population.
8 K. YEBOAH ET AL.
5. Conclusion Availability of data
In a periurban hospital in Ghana, compared to non- A data set supporting the conclusions of this paper is 
HIV controls or cART-nave HIV patients, cART-treated available and can be requested from the corresponding 
HIV patients had increased arterial stiffness measured author.
as CAVI. CAVI is associated with metabolic abnormal-
ities in non-HIV controls and cART-naïve HIV patients, ORCID
but not in cART-treated HIV patients. CAVI increased Kwame Yeboah http://orcid.org/0000-0001-5240-0645
with increasing age and systolic BP, as well as Jennifer Agyekum http://orcid.org/0000-0003-0051-985X
decreasing BMI. Patients on a TDF-based regimen Bartholomew Dzudzor http://orcid.org/0000-0003-2325- 
had decreased CAVI compared to those on an AZT- 7063
based regimen. Multicentre longitudinal studies 
should be conducted to investigate the impact of 
HIV infection and various cART regimen on arterial References
stiffness and cardiovascular end-organ damage in [1] Joint United Nations Programme on HIV/AIDS. Global 
HIV patients in sub-Saharan Africa. HIV & AIDS statistics — Fact sheet. Geneva, 
Switzerland: UNAIDS; 2021 [cited 2023 January 10]. 
Available from: https://www.unaids.org/sites/default/ 
files/media_asset/UNAIDS_FactSheet_en.pdf
Abbreviations [2] Defo AK, Chalati MD, Labos C, et al. Association of HIV 
infection and antiretroviral therapy with arterial stiffness: 
BMI body mass index a systematic review and meta-analysis. Hypertension. 
BP blood pressure 2021;78(2):320–332.
Cart combination antiretroviral therapy [3] Bijker R, Jiamsakul A, Uy E, et al. Cardiovascular 
CAVI cardio-ankle vascular index disease-related mortality and factors associated with 
CVD cardiovascular diseases cardiovascular events in the TREAT Asia HIV 
FPG fasting blood glucose Observational Database (TAHOD). HIV Med. 2019;20 
HDL high-density lipoprotein (3):183–191. DOI:10.1111/hiv.12687
HIV human immunodeficiency virus [4] Fourie CMT, Schutte AE, Smith W, et al. Endothelial 
LDL low-density lipoprotein activation and cardiometabolic profiles of treated and 
MetS metabolic syndrome never-treated HIV infected Africans. Atherosclerosis. 
PWV pulse wave velocity 2015;240(1):154–160.
[5] Kim ED, Ballew SH, Tanaka H, et al. Short-term prog-
nostic impact of arterial stiffness in older adults with-
Acknowledgments out prevalent cardiovascular disease. Hypertension. 
2019;74(6):1373–1382.
We thank all the participants in the HIV clinics at Atua [6] Shirai K, Utino J, Otsuka K, et al. A novel blood 
Government Hospital who voluntarily participated in this pressure-independent arterial wall stiffness parameter; 
study. We thank community health volunteers and healthcare cardio-ankle vascular index (CAVI). J Atheroscler 
providers at HIV clinics for their overwhelming support during Thromb. 2006;13(2):101–107.
the study. Our heartfelt thanks go to Mrs Nneka Essel for her [7] Shirai K, Suzuki K, Tsuda S, et al. Comparison of cardio– 
assistance. ankle vascular index (CAVI) and CAVI0 in large healthy 
and hypertensive populations. J Atheroscler Thromb. 
2019;26(7):603–615.
[8] Aiumtrakul N, Supasyndh O, Krittayaphong R, et al. Cardio- 
Disclosure statement ankle vascular index with renal progression and mortality 
in high atherosclerosis risk: a prospective cohort study in 
No potential conflict of interest was reported by the authors. CORE-Thailand. Clin Exp Nephrol. 2022;26(3):247–256.
[9] Satoh-Asahara N, Kotani K, Yamakage H, et al. Cardio-ankle 
vascular index predicts for the incidence of cardiovascular 
events in obese patients: a multicenter prospective cohort 
Funding study (Japan Obesity and Metabolic Syndrome Study: 
There was no funding for this study. jOMS). Atherosclerosis. 2015;242(2):461–468.
[10] Murakami K, Inayama E, Itoh Y, et al. The role of 
cardio-ankle vascular index as a predictor of mortality in 
patients on maintenance hemodialysis. Vasc Health Risk 
Authors’ contributions Manag. 2021;17:791–798.
[11] Botha S, Fourie CMT, van Rooyen JM, et al. 
KY conceptualized the study, analysed the data and Cardiometabolic changes in treated versus never 
drafted the manuscript. SE collected the data and revised treated HIV-Infected black South Africans: the 
the manuscript. JAA analyzed the data and made scien- PURE study. Heart Lung & Circulation. 2014;23 
tific contributions to the manuscript. BD reviewed the (2):119–126.
manuscript and made scientific contributions to the [12] Gleason RL Jr., Caulk AW, Seifu D, et al. Current 
manuscript. All authors approved the content of the Efavirenz (EFV) or Ritonavir-Boosted Lopinavir (LPV/r) 
manuscript. use correlates with elevate markers of atherosclerosis 
LIBYAN JOURNAL OF MEDICINE 9
in HIV-Infected subjects in Addis Ababa, Ethiopia. PLoS comparison of the cardio-ankle vascular index and 
ONE. 2015;10(4):e0117125. central aortic techniques. BMC Endocr Disord. 2016;16 
[13] Huisman HW, Schutte R, Venter HL, et al. Low BMI is (1):53.
inversely associated with arterial stiffness in Africans. [25] Atkinson J. Age-related medial elastocalcinosis in 
British Journal Of Nutrition. 2015;113(10):1621–1627. arteries: mechanisms, animal models, and physiological 
[14] Lemogoum D, Ngatchou W, Borne PVD, et al. consequences. J Appl Physiol. 2008;105(5):1643–1651.
Cardiometabolic phenotype and arterial stiffness in [26] Lee H-Y, Oh B-H. Aging and arterial stiffness. Circ J. 
HIV-Positive black African Patients. Open J Prev Med. 2010;74(11):2257–2262.
2014;04(04):11. DOI:10.4236/ojpm.2014.44024 [27] Soska V, Dobsak P, Dusek L, et al. Cardio-ankle vascular 
[15] Aurpibul L, Srithanaviboonchai K, Rerkasem K, et al. index in heterozygous familial hypercholesterolemia. 
Prevalence of subclinical atherosclerosis and risk of J Atheroscler Thromb. 2012;19(5):453–461.
atherosclerotic cardiovascular disease in older adults [28] Yeboah K, Antwi DA, Gyan B. Arterial stiffness in non-
living with HIV. AIDS Res Human Retroviruses. 2019;35 hypertensive type 2 diabetes patients in Ghana. 
(11–12):1136–1142. Int J Endocrinol. 2016;2016:6107572.
[16] Jewpakanon T, Satitthummanid S, Songmuang SB. [29] Schillaci G, Socio GVLD, Pucci G, et al.: Aortic stiffness in 
Subclinical left ventricular dysfunction virologically untreated adult patients with human immunodefi-
supressed HIV-infected patients receiving antiretroviral ciency virus infection. Hypertension 2008, 52 
therapy. J Am Coll Cardiol. 2022;79(9_Supplement):1348. (2):308–313.
[17] Schutte AE, Kruger R, Gafane-Matemane LF, et al. [30] Gomez-Sanchez L, Garcia-Ortiz L, Patino-Alonso MC, 
Ethnicity and arterial stiffness. Arteriosclerosis et al. Association of metabolic syndrome and its com-
Thrombosis Vasc Biol. 2020;40(5):1044–1054. ponents with arterial stiffness in Caucasian subjects of 
[18] Alberti KG, Eckel RH, Grundy SM, et al. Harmonizing the the MARK study: a cross-sectional trial. Cardiovasc 
metabolic syndrome: a joint interim statement of the Diabetol. 2016;15(1):148.
international diabetes federation task force on epide- [31] Kumar S, Samaras K. The impact of weight gain during 
miology and prevention; national heart, lung, and HIV treatment on risk of pre-diabetes, diabetes melli-
blood institute; American heart association; world tus, cardiovascular disease, and mortality. Front 
heart federation; international atherosclerosis society; Endocrinol (Lausanne). 2018;9:9.
and international association for the study of obesity. [32] Tate T, Willig AL, Willig JH, et al. HIV infection and 
Circulation. 2009;120(16):1640–1645. obesity: where did all the wasting go? Antivir Ther. 
[19] Sun C-K. Cardio-ankle vascular index (CAVI) as an indi- 2012;17(7):1281–1289.
cator of arterial stiffness. Integr Blood Press Control. [33] Zhao D, Ding Y, Lin H, et al. Mitochondrial Haplogroups 
2013;6:27. N9 and G are associated with metabolic syndrome 
[20] Shirai K, Utino J, Saiki A, et al. Evaluation of blood among human immunodeficiency virus-infected 
pressure control using a new arterial stiffness para- patients in China. AIDS Res Human Retroviruses. 
meter, Cardio-ankle Vascular Index (CAVI). Curr 2019;35(6):536–543.
Hypertens Rev. 2013;9(1):66–75. [34] Sitole L, Fortuin R, Tugizimana F. Metabolic profiling of 
[21] Yeboah K, Musah L, Essel S, et al. P98 impaired pul- HIV infected individuals on an AZT-based antiretroviral 
monary function is associated with increased treatment regimen reveals persistent oxidative stress. 
cardio-ankle vascular index in HIV patients in Ghana. J Pharm Biomed Anal. 2022;220:114986.
Artery Research. 2020;25(Supplement 1):S139. [35] Alzate Angel JC, Duque Molina MM, García García HI. 
[22] Msoka TF, Van Guilder GP, Smulders YM, et al. Zidovudine/Lamivudine vs. Abacavir/Lamivudine vs. 
Antiretroviral treatment and time since HIV-1 diagnosis Tenofovir/Emtricitabine in fixed-dose combinations as 
are associated with large artery stiffness in sub-Saharan initial treatment for HIV patients: a systematic review 
African HIV-1 patients. Artery Research. 2016;16:34–41. and network meta-analysis. Colombia Médica. 2017;48 
[23] Ngatchou W, Lemogoum D, Ndobo P, et al. Effects of (2):70–81.
antiretroviral therapy on arterial stiffness in Cameroonian [36] Dadi TL, Kefale AT, Mega TA, et al. Efficacy and toler-
HIV-infected patients. Blood Press Monit. 2013;18 ability of tenofovir disoproxil fumarate based regimen 
(5):247–251. DOI:10.1097/MBP.0b013e328363ee43 as compared to zidovudine based regimens: 
[24] Yeboah K, Antwi DA, Gyan B, et al. Arterial stiffness in a systematic review and meta-analysis. AIDS Res 
hypertensive and type 2 diabetes patients in Ghana: Treat. 2017;2017:5792925.