No accelerated progression of subclinical atherosclerosis with integrase strand transfer inhibitors compared to non-nucleoside reverse transcriptase inhibitors

Abstract

Background

The role of integrase strand transfer inhibitors (INSTI) in the cardiovascular risk of people with HIV is controversial.

Objectives

To assess the association of INSTI to subclinical atherosclerosis progression measured with the carotid intima-media thickness (cIMT).

Methods

Prospective study in virologically suppressed people with HIV receiving INSTI- or NNRTI-based regimens. cIMT was measured at baseline, 48 and 96 weeks. cIMT progression was analysed both as a continuous and categorical variable, defined as cIMT increase ≥ 10% and/or new carotid plaque. Adjustments through Cox proportional hazard regression and linear mixed models, and propensity score matching were conducted.

Results

190 participants were recruited and 173 completed the 96 week follow-up. 107 (56.3%) were receiving an INSTI-containing, 128 (67.4%) a NNRTI-containing and 45 (23.7%) a NNRTI plus an INSTI-containing regimen. The overall median (IQR) 2-year change of cIMT was 0.029 (−0.041 to 0.124) mm; 87 (45.8%) participants experienced a cIMT increase ≥ 10%, of whom 54 (28.4%) developed a new carotid plaque. Adjusted Cox regression showed no differences between INSTI and NNRTI groups in the categorical 2-year progression of cIMT, both including or excluding participants receiving INSTI + NNRTI. Similar results were observed for the continuous cIMT increase through adjusted linear mixed models. Propensity score matching showed no significant differences in the 2 year cIMT change between treatment groups [0.049 mm (−0.031–0.103) in the INSTI group versus 0.047 mm (−0.023–0.115) in the NNRTI group; P = 0.647]. cIMT progression was associated with traditional cardiovascular risk factors.

Conclusions

INSTI-based regimens are not associated with increased progression of subclinical atherosclerosis when compared to NNRTI.

Introduction

Cardiovascular disease has emerged as one of the primary causes of morbidity and mortality in people with HIV (PWH).¹ The risk of myocardial infarction and cerebrovascular disease in PWH is 2-fold higher compared to the general population.^2,3 Among the contributing pathogenic mechanisms specifically linked to HIV infection, certain antiretroviral drugs have been implicated, predominantly protease inhibitors and abacavir.^4,5

Contemporary ART relies on integrase strand transfer inhibitors (INSTI). In initial studies comparing PI- and INSTI-based regimens, INSTI were shown to have a lesser impact on lipid and cardiovascular biomarker levels, both in ART-naive individuals and in switching studies.^6,7 However, accumulating data suggest that initiating or switching to INSTI could be associated with a higher risk of unfavourable metabolic outcomes, such as weight gain, increases in blood pressure and an enhanced risk of hypertension or diabetes.^8–11 Furthermore, recent evidence suggests a higher incidence of cardiovascular disease in PWH receiving INSTI-based regimens.¹² Nevertheless, the metabolic effects and cardiovascular risks linked with INSTI have not shown consistency across studies, and additional research has yielded conflicting results.^13–15 Assessing the influence of INSTI on the dynamics of atherosclerosis might contribute to expanding our understanding about the cardiovascular impact of this class of ART.

Carotid intima-media thickness (cIMT) assessed by B-mode ultrasonography is a non-invasive imaging modality that allows the detection of asymptomatic atherosclerotic vascular disease.¹⁶ The evaluation of cIMT has proven valuable in assessing the accelerated progression of subclinical atherosclerosis in PWH.¹⁷ Furthermore, cIMT is a predictor of future cardiovascular events,¹⁸ and of mortality within the HIV population.^19,20 The accessibility, low toxicity and low cost, make cIMT measurement a suitable option for repeated measurements to assess atherosclerosis progression.

We conducted a prospective study in a contemporary cohort of virologically suppressed PWH and assessed the contribution of INSTI to subclinical atherosclerosis progression measured with the cIMT.

Methods

Study design, participants and study procedures

A prospective study was carried out at Hospital General Universitario de Elche, Spain. All participants were adults (> 18 years old) with HIV, enrolled between 26 October 2017 and 18 February 2019, and were followed-up until 12 January 2022. Participants included had been receiving ART with regimens based on NNRTI or INSTI, with undetectable viral load (HIV-1 RNA levels < 50 copies/mL) during at least 6 months before inclusion in the study. Participants with known cardiovascular disease, those on treatment with a PI-containing regimen, pregnant women or those becoming pregnant during follow-up, were excluded from the study.

Participants were followed-up for 2 years with face-to-face visits scheduled at baseline, 48 and 96 weeks. On each visit, physical examination and anthropometric measures including blood pressure, weight, height and BMI were collected, and blood samples were obtained for biochemical, metabolic and virological measurements. Blood samples were processed and plasma was obtained and cryopreserved at −80°C. Lymphocyte counts and their subsets were measured in fresh EDTA whole blood (AQUIOS CL Flow Cytometer, Beckman Coulter).

Subclinical atherosclerosis measurement

Carotid artery intima-media thickness measurement was carried out at baseline, 48- and 96-week visits by a single trained ultrasonographer blinded to patients’ ART, following a standardized protocol, as previously described.^21,22 B-mode ultrasound of carotids was performed using a Toshiba system (Toshiba Aplio 400), equipped with a 7–12 MHz linear array transducer. Ultrasonic scans were recorded from right and left common carotid, internal carotid artery and bifurcation. Measurements in millimetres from these six locations, using semiautomated detection software, were averaged and reported as a single cIMT (6-point cIMT).

We analysed the increase of cIMT both as a continuous variable and also as a categorical variable, defining cIMT progression as a 6-point cIMT increase ≥ 10% and/or detection of new carotid plaque (any measurement of cIMT > 1.5 mm).

Inmuno-activation, inflammation and prothrombotic biomarkers

High-sensitivity C-reactive protein (hsCRP) was measured by immunoturbidimetric assay (CRP Gold Latex, DiAgam, Belgium) with an automated instrument (VITROS^® 5600 System, Ortho Clinical Diagnostics). D-dimer [Human D2D (D-Dimer) ELISA Kit], soluble ICAM-1 (sICAM-1) [Human ICAM-1/CD54 (intercellular adhesion molecule 1) ELISA Kit], soluble CD14 (sCD14) [Human sCD14 (Soluble Cluster of Differentiation 14) ELISA Kit] and soluble CD163 (sCD163) [Human sCD163 (Soluble Cluster of Differentiation 163) ELISA Kit] were measured by enzyme-linked immunosorbent assay (ELISA Kits, Elabscience Biotechnology Inc., USA) with an automated instrument (Dynex DS2^® ELISA system).

Statistical analyses

We defined 2-year change in cIMT as the difference between baseline and 96-week measures. Mann–Whitney–Wilcoxon or Student’s t-tests were used for group comparison in continuous variables, according to the result of Shapiro Wilk’s contrast of normality. For categorical variables, comparisons were performed using the χ² for variables with two or more categories, and Fisher’s exact tests for dichotomous variables. For multiple simultaneous comparisons, a Bonferroni adjustment was applied. We used multivariate adjusted Cox proportional hazard regression models to assess the association of cardiovascular risk factors, HIV-related factors, inflammatory biomarkers and ART with cIMT progression defined as a categorical variable incorporating covariates of interest. As a secondary exploratory analysis, linear mixed models with a random term for patient were used to examine associations with cIMT increase as a continuous variable.

To emulate some of the characteristics of a randomized study to estimate the effects of the INSTI-based regimen on subclinical atherosclerosis, treatment groups were balanced through propensity score matching with caliper 0.1. The covariates for propensity matching included durations of exposure to each ART regimen, age, sex, Framingham risk score, cIMT at baseline and baseline CD4 T cells.

For the calculation of the statistical power of our study, we analysed the difference in the 2 year cIMT increase between INSTI- and NNRTI-based groups. Our primary hypothesis was that cIMT progression over 96 weeks would be higher in participants with INSTI-based regimens compared with those on NNRTI-based regimens. With sample sizes of 62 participants on INSTI and 83 on NNRTI, our study provided 80% power using a Mann–Whitney U or Wilcoxon rank-sum test to detect a 2 year difference in cIMT change of at least 0.0192 mm.

Statistical significance was defined at a P value of < 0.05, and for Bonferroni adjustment at P value of < 0.0025 for continuous variables and P value of < 0.003 for discrete variables. Statistical analyses were performed with R version 4.0.3 software (R Foundation for Statistical Computing).

Ethics

The study was approved by the Ethical Committee of Hospital General Universitario de Elche, and all participants signed an informed consent.

Results

Participants’ characteristics

Of 190 participants recruited at baseline, 190 and 173 completed the follow-up at 48- and 96-week visits. Reasons for dropout during follow-up are shown in Figure S1 (available as Supplementary data at JAC Online).

Baseline participants’ characteristics are detailed in Table 1. Median (IQR) age at enrolment was 48 (39–54) years, 154 (81.1%) were male, 111 (59%) had at least one traditional cardiovascular risk factor, 95 (50.5%) were smokers, and 22 (11.7%), 19 (10%) and 5 (2.6%) participants had dyslipaemia, hypertension or diabetes, respectively. The median (IQR) 10-year Framingham and atherosclerotic cardiovascular disease (ASCVD) risk scores were 6.3 (3.7–14.5) % and 2.6 (1.3–7.1) %, respectively, and the median (IQR) BMI was 24.7 (22.5–27.8) kg/m².

Table 1.

Characteristics of patients according to the antiretroviral regimen composition

Antiretroviral regimen
	All patients	INSTI-containing regimen	NNRTI-containing regimen	Both INSTI and NNRTI-containing regimen	P value∗	P value∗∗
Patients, no.	190 (100)	62 (32.6)	83 (43.7)	45 (23.7)	—	—
Male sex	154 (81.1)	46 (74.2)	69 (83.1)	39 (86.7)	0.217	0.579
Age, years	48 (39–54)	49 (42–55)	49 (42–54)	39 (33–51)	0.900	0.164
Baseline CVD risk factors, no. (%)
Any traditional CVD risk factora	111 (59)	37 (60.7)	49 (60)	25 (55.6)	1.000	0.882
≥2 traditional CVD risk factorsa	23 (12.2)	8 (13.1)	13 (16)	2 (4.4)	0.812	0.261
Current smoking	95 (50.5)	30 (49.2)	44 (53.7)	21 (46.7)	0.616	0.466
Hypertension	19 (10)	8 (13)	8 (9.6)	3 (6.7)	0.598	1.000
Dyslipidaemia	22 (11.7)	8 (13)	11 (13.4)	3 (6.7)	1.000	0.648
Diabetes	5 (2.6)	3 (4.8)	2 (2.4)	0 (0)	0.651	1.000
Framingham risk score, %	6.3 (3.7–14.5)	6.6 (3.5–17)	6.5 (4.7–14.6)	4.4 (2.7–9.4)	0.375	0.045
ASCVD risk score, %	2.6 (1.3–7.1)	2.4 (1–9.1)	4 (1.7–8.1)	1.8 (1–4.5)	0.281	0.030
Baseline cardio-metabolic parameters, no. (%)
Blood pressure (BP), mm Hg
Systolic	121 (111–134)	123 (111–131)	123 (111–136.5)	118 (111–129)	0.413	0.188
Diastolic	70 (64–77)	71 (66–77)	70 (63–79)	67 (63–76)	0.522	0.920
BMI, kg/m2	24.7 (22.5–27.8)	24.7 (22–27.8)	25.2 (22.6–27.7)	24.4 (23.1–27.5)	0.513	0.490
Weight, kg	73 (64–83)	71.5 (63–83)	73.5 (66–84)	75.5 (68–83.5)	0.351	0.741
Total cholesterol, mg/dL	171 (149–195)	167.5 (147–185)	180 (150–199)	170 (146–197)	0.057	0.067
LDL-cholesterol, mg/dL	102 (88–125)	97 (85–113)	110 (93–131)	99 (82–124)	0.004	0.006
HDL-cholesterol, mg/dL	46 (38–54)	45 (35–56)	44 (38–53)	48 (44–54)	0.905	0.222
Triglycerides, mg/dL	98 (71–137)	96 (74–153)	101 (73–135)	83 (61–117)	0.770	0.349
Baseline inflammatory biomarkers
hsCRP, ng/mL	0.7 (0.3–2)	0.5 (0.3–1.9)	0.6 (0.3–1.4)	0.8 (0.3–2.3)	0.628	0.939
D-dimer, pg/mL	0.3 (0.2–0.5)	0.4 (0.3–0.7)	0.3 (0.2–0.5)	0.4 (0.2–0.5)	0.020	0.041
sCD14, pg/mL	5034 (4445–5670)	5033 (4346–5668)	5037 (4550–5738)	5056 (4382–5383)	0.665	0.474
sCD163, pg/mL	249 (182–328)	293 (175–396)	256 (196–326)	223 (171–286)	0.414	0.669
sICAM-1, pg/mL	82 (65–107)	89 (71–113)	76 (62–102)	802 (66–97)	0.076	0.144
Baseline immunovirological status and ART
CD4 T cell count nadir, cells/mm3	250 (139–351)	212 (80–339)	250 (155–312)	300 (178–417)	0.383	0.768
CD4 T cell count, cells/mm3	687 (496–866)	690 (433–868)	688 (501–866)	665 (533–852)	0.647	0.693
CD4 T cell percentage, %	36 (30–42)	35.5 (28–41)	35.3 (30–41)	38 (32–42)	0.544	0.894
CD4/CD8 ratio	0.9 (0.7–1.3)	0.8 (0.6–1.3)	0.9 (0.7–1.3)	1 (0.8–1.3)	0.232	0.529
Months of exposure to last ART	33 (19–49)	27.1 (17–38)	49.9 (28–63)	21.4 (11–37)	0.001	0.001
Baseline cIMT, mm	0.808 (0.72–1.03)	0.818 (0.75–1.06)	0.848 (0.72–1.04)	0.773 (0.69–0.92)	0.987	0.333
Two year changes in body weight, BP and cIMT during the study
Body weight increase, kg	1 (−1.5–4.5)	1 (−1.1–4.2)	1.4 (−1.8–4.6)	1.5 (−0.5–4.5)	0.866	0.853
Systolic BP increase, mmHg	0.1 (−12–10.5)	−3.5 (−14.8–9.8)	0.5 (−9–11)	1 (−10–9)	0.270	0.398
cIMT increase, mmb	0.029 (−0.041–0.124)	0.044 (−0.040–0.142)	0.021 (−0.048–0.115)	0.024 (0.027–0.085)	0.415	0.597
cIMT progression, %c	87 (45.8)	30 (48.8)	39 (47)	18 (40)	1.000	0.884

Categorical variables are expressed as number (percentage) and continuous variables as median (IQR).

INSTI, integrase strand transfer inhibitors; CVD, cardiovascular disease; ASCVD, atherosclerotic cardiovascular disease; BP, blood pressure; hsCRP, high-sensitivity C-reactive protein; sCD14, soluble CD14; sCD163, soluble CD163; sICAM-1, soluble intercellular adhesion molecule-1; cIMT, carotid intima-media thickness.

^aTraditional CVD risk factors: current smoking, hypertension, dyslipidaemia and diabetes.

^bcIMT increase represents the difference in mm between median cIMT at baseline and median cIMT at 96-week.

^ccIMT progression was defined as a 6-point cIMT increase ≥ 10% and/or detection of new carotid plaques (any measurement of cIMT > 1.5 mm) during the 96-week follow-up.

^∗ P value for the comparison between participants of the INSTI-containing regimen versus NNRTI-containing regimen

^∗∗ P value for the comparison between participants with INSTI-containing regimen versus non-INSTI-containing regimen.

Among the study participants, 172 (90.5%) had maintained the same ART combination for at least 1 year, and 15 (7.9%) had switched to the current treatment regimen within the previous 6 months. The median (IQR) duration of the latter ART regimen at baseline was 33 (19–49) months. Regarding ART composition at baseline, 128 (67.4%) participants were receiving a NNRTI-containing regimen and 107 (56.3%), an INSTI-containing regimen, of whom 45 (23.7%) were receiving a combination of a NNRTI plus an INSTI. Detailed information about the specific INSTI and NNRTI used is summarized in Table S1. The median (IQR) duration of exposure to INSTI at baseline was 27.1 (17–38) months and to NNRTI 49.9 (28–63) months. During the 96-week follow-up, 10 participants (5.26%) switched to an INSTI-containing regimen, with a median duration of exposure since drug change of 18 (6–22) months. In 27.4% of participants receiving an INSTI, it was their first ART regimen.

Table 1 shows the baseline characteristics of participants according to ART group. Treatment group comparisons used an intention-to-treat approach. In the comparison of participants with INSTI- against participants with NNRTI-containing regimen, the INSTI group showed lower levels of LDL-cholesterol (P = 0.004), total cholesterol (P = 0.057), higher levels of D-dimer (P = 0.020) and fewer months of suppressed viral load at baseline (P = 0.002). However, the prevalence of traditional cardiovascular risk factors was not different among groups. There were no differences in the increase in blood pressure or body weight between treatment groups over the study period. Participants on INSTI-based regimens (including those on INSTI plus NNRTI) had lower Framingham and ASCVD risk scores than those treated with non-INSTI regimens, but no differences were observed in the scores when the group receiving INSTI plus NNRTI was excluded.

Two-year progression in carotid artery intima-media thickness

The overall median (IQR) 2-year change of cIMT was 0.029 (−0.041 to 0.124) mm. A total of 87 (45.8%) participants experienced an increase in the 6-point cIMT ≥ 10%, with 54 (28.4%) of them developing a new carotid plaque. No significant differences in the 2 year change of cIMT were observed in the unadjusted analysis between patients receiving INSTI-containing regimens and those receiving NNRTI-containing regimens (Table 1). Similarly, no differences between ART groups were observed in the progression of cIMT defined as a categorical variable.

Table 2 shows the baseline characteristics of the participants categorized by cIMT progression. In the univariate analysis, age (P < 0.001), systolic (P < 0.001) and diastolic blood pressure (P < 0.001), BMI (P = 0.044), cardiovascular risk indexes [Framingham score (P < 0.001) and ASCVD score (P < 0.001)], HDL-cholesterol (P = 0.013), LDL-cholesterol (P = 0.026), triglycerides (P = 0.005), D-dimer (P = 0.037) and sCD14 (P = 0.040) were associated with cIMT progression. Among HIV-related factors, participants experiencing cIMT progression showed a lower CD4 T cell percentage (P = 0.047). No significant differences between groups were observed in historical exposures to ABC (36.8% of participants experiencing cIMT progression versus 30.1%; P = 0.356, with a median duration of exposure of 2.7 versus 2.1 years; P = 0.357) or to protease inhibitors (39.1% of participants experiencing cIMT progression versus 29.1%; P = 0.167, with a median duration of exposure of 10.9 versus 10.5 years; P = 0.941). A multivariate Cox proportional hazards regression model was carried out adjusting for variables showing statistically significant difference between cIMT progression groups and covariates of interest, specifically age, sex, INSTI-based ART, cIMT at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer. Results showed that age, systolic blood pressure, HDL-cholesterol, LDL-cholesterol, CD4 T cell percentage, D-dimer and sCD14 were independent factors associated with cIMT progression. INSTI-exposure showed a non-significant trend to significance (P = 0.094). When the participants receiving ART with an INSTI plus a NNRTI were excluded from the analysis, no relationship was observed between INSTI and cIMT progression (Table 2). Kaplan–Meier curves show the adjusted probability of cIMT progression during follow-up among participants with INSTI- versus non-INSTI-containing regimens (Figure 1a and b).

Table 2.

Characteristics of patients according to the progression of cIMT during the 2 year follow-up

cIMT progressionb
	Yes		No		P-value	Adjusted HRc	Adjusted HRd
Patients, no.	87	(45.8)	103	(54.2)	—	—	—
Male sex	71	(81.6)	83	(80.6)	1.000	1.24(0.93–1.64)	1.19(0.86–1.65)
Age, years	53	(46–58)	44	(36–51)	0.001	1.03(1.02–1.04)	1.03(1.01–1.04)
Baseline CVD risk factors, no. (%)
Any traditional CVD risk factora	55	(63.2)	56	(55.4)	0.301
≥2 traditional CVD risk factorsa	15	(17.2)	8	(7.9)	0.073
Current smoking	43	(49.4)	52	(51.5)	0.884
Hypertension	13	(14.9)	6	(5.8)	0.051
Dyslipidaemia	17	(19.5)	5	(5)	0.003
Diabetes	3	(3.4)	2	(1.9)	0.662
Framingham risk score, %	11	(5.5–20.3)	4.7	(3.1–8.1)	0.001
ASCVD risk score, %	5.9	(2.4–11)	1.8	(1–3.7)	0.001
Baseline cardio-metabolic parameters, no. (%)
Blood pressure (BP), mm Hg
Systolic	127	(114–138)	118	(110–126)	0.001	1.01(1.00–1.01)	1.01(1.01–1.02)
Diastolic	73	(66–80)	67	(62–74)	0.001
BMI, kg/m2	25.6	(22.7–28.2)	24	(22–27.4)	0.044	0.99(0.96–1.02)	1.04(1.00–1.08)
Weight, kg	73	(66–84)	71	(63–81)	0.119
Total cholesterol, mg/dL	177	(151–199.5)	167	(147–186)	0.115
LDL-cholesterol, mg/dL	110	(90.5–132)	99	(87–116)	0.026	1.01(1.00–1.01)	1.00(1.00–1.01)
HDL-cholesterol, mg/dL	43	(36–51)	48	(40–57)	0.013	0.98(0.97–0.99)	0.99(0.98–1.00)
Triglycerides, mg/dL	107	(73–170.5)	89	(69.2–123)	0.005
Baseline inflammatory biomarkers
hsCRP, ng/mL	0.9	(0.3–2.5)	0.6	(0.2–1.4)	0.072	1.10(0.91–1.34)	1.17(0.93–1.45)
D-dimer, pg/mL	0.4	(0.2–0.7)	0.3	(0.2–0.4)	0.037	1.57(1.19–2.08)	1.12(0.80–1.57)
sCD14, pg/mL	5232	(4640–5870)	4887	(4341–5366)	0.040	1.02(1.01–1.03)	1.02(1.01–1.04)
sCD163, pg/mL	259.2	(192.9–332)	240.9	(169.4–310)	0.299
sICAM-1, pg/mL	84	(68.4–110.3)	80.4	(63.3–96.3)	0.236
Baseline immunovirological status and ART
CD4 T cell count nadir, cells/mm3	246	(142–388)	253	(142–388)	0.388
CD4 T cell count, cells/mm3	654	(450–876)	711	(544–866)	0.286
CD4 T cell percentage, %	34	(29–41)	38	(32–42)	0.047	0.99(0.98–1.00)	0.99(0.98–1.00)
CD4/CD8 ratio	0.9	(0.7–1.3)	0.9	(0.7–1.3)	0.342
Months of exposure to last ART	33	(19–52)	33	(18–44)	0.369
NNRTI treatment at baseline, no (%)	55	(63.2)	69	(67)	0.761
INSTI treatment at baseline, no (%)	48	(55.2)	59	(57)	0.884	1.20(0.97–1.50)	0.98(0.75–1.27)
Baseline cIMT, mm	0.895	(0.754–1.152)	0.763	(0.688–0.893)	0.001	1.02(0.75–1.39)	1.18(0.85–1.62)
Two-year changes in body weight, BP and cIMT during the study
Body weight increase, kg	3	(1–5.2)	2	(0.5–4.8)	0.275
Systolic BP increase, mmHg	14	(4–20)	8	(1.2–14)	0.018

cIMT, carotid intima-media thickness; CVD, cardiovascular disease; ASCVD, atherosclerotic cardiovascular disease; BP, blood pressure; hsCRP, high-sensitivity C-reactive protein; sCD14, soluble CD14; sCD163, soluble CD163; sICAM-1, soluble intercellular adhesion molecule-1; cIMT, carotid intima-media thickness; INSTI, integrase strand transfer inhibitors.

^aTraditional CVD risk factors: current smoking, hypertension, dyslipidaemia and diabetes.

^bcIMT progression was defined as a 6-point cIMT increase ≥ 10% and/or detection of new carotid plaques (any measurement of cIMT > 1.5 mm).

^cMultivariate Cox proportional hazards regression model of factors associated with progression of cIMT > 10% and/or detection of new carotid plaque, adjusted for variables showing statistically significant difference between cIMT progression groups and covariates of interest, specifically age, sex, INSTI-based ART, cIMT at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer.

^dMultivariate Cox proportional hazards regression model of factors associated with progression of cIMT > 10% and/or detection of new carotid plaque and excluding from the analysis participants receiving ART with an INSTI plus a NNRTI, adjusted for variables showing statistically significant difference between cIMT progression groups and covariates of interest, specifically age, sex, INSTI-based ART, cIMT at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer.

Figure 1.

Kaplan–Meier curves to estimate the cumulative proportion of patients developing cIMT progression according to the antiretroviral regimen. (a) Participants with INSTI-containing versus non-INSTI-containing regimen (participants on INSTI plus NNRTI are included in the INSTI group). (b) Participants with INSTI-containing versus NNRTI-containing regimen (participants on INSTI plus NNRTI are excluded from the INSTI group).

Factors associated with cIMT increase analysed as a continuous variable are shown in Table 3. In the univariate analysis, cIMT increase was associated with age, presence of two or more traditional cardiovascular risk factors, dyslipidaemia, higher levels of total, HDL- and LDL-cholesterol, higher levels of D-dimer and higher cIMT at baseline. A multivariate analysis through linear mixed model, including the significant variables associated with cIMT increase and the covariates of interest (age, sex, INSTI-containing ART, cIMT at baseline, presence of two or more cardiovascular risk factors at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer), showed that HDL-cholesterol, LDL-cholesterol and D-dimer were independent factors associated with cIMT increase, and sex and cIMT at baseline were close to significance. No association was found with INSTI exposure. The same results concerning the relationship of INSTI with cIMT increase were observed when patients receiving an INSTI plus a NNRTI were excluded from the analysis (Table 3).

Table 3.

Univariate and multivariate analysis of factors associated with cIMT increase analysed as a continuous variable during the 2 year follow-up

cIMT increase as a continuous variable
Univariate analysis
Variable	Regression coefficient Ω, × 10−3 mm (95% CI)	P value	Adjusted coefficientb, × 10−3	Adjusted coefficientc, × 10−3
Age at baseline, years	3.6 (0.5–6.7)	0.023∗	0.3 (−0.5–1.4)	0.4 (−0.6–1.3)
Male sex	15.2 (−70.0–100.4)	0.726	17.5 (−4.8–37.8)	21.5 (−7–33.5)
≥ 2 traditional CVD risk factorsa at baseline	106.5 (6.6–206.3)	0.037∗	−8.4 (−28.9–14.6)	−10.3 (−36.7–15.1)
Hypertension at baseline, no. (%)	69.7 (−39.3–178.7)	0.209	—	—
Systolic blood pressure (all values), mm Hg	0.3 (−0.1–0.7)	0.126	0.1 (−0.5–0.5)	0.1 (−0.4–0.6)
Systolic blood pressure increase, mm Hg	0.5 (−0.5–1.4)	0.345	—	—
Weight (all values), kg	0.2 (−0.3–0.7)	0.511	—	—
Weight gain. kg	1.8 (−2–5.5)	0.356	—	—
BMI at baseline, kg/m2	−1.2 (−10.9–8.5)	0.805	—	—
BMI (all values), kg/m2	0.6 (−1.3–2.4)	0.536	−0.3 (−2.6–2)	−0.3 (−2.5–2.3)
INSTI-containing regimen (intention to treat)	−36.2 (−105.3–33)	0.303	−4.1 (−16.6–13.4)	−4.2 (−22.5–11.6)
Dyslipidaemia at baseline, no. (%)	155.2 (52.7–257.7)	0.003∗∗	—	—
Total cholesterol (all values), mg/dL	0.3 (0.1–0.4)	0.002∗∗	—	—
HDL-cholesterol (all values), mg/dL	0.3 (0.1–0.7)	0.049	0.4 (0.1–0.8)	0.4 (0.1–0.8)
LDL-cholesterol (all values), mg/dL	0.3 (0.1–0.5)	0.009	0.2 (0.1–0.5)	0.3 (0.1–0.5)
Diabetes at baseline, no. (%)	7.5 (−196.9–212)	0.942	—	—
CD4 T cell percentage (all values), %	−0.3 (−0.2–4.1)	0.347	−0.1 (−1.1–0.9)	−0.4 (−1–0.7)
CD4 T cell count (all values), cells/mm3	0.1 (−0.1–0.2)	0.253	—	—
CD4/CD8 ratio (all values)	−3.4 (−1.6–97.0)	0.628	—	—
hsCRP (log, all values), ng/mL	11.5 (−0.1–23.2)	0.071	0.6 (−11.4–16.1)	−3.0 (−13.5–19.1)
D-dimer (log, all values), pg/mL	34.2 (10.9–57.4)	0.001∗∗	29.1 (6.7–50.3)	38.2 (6.5–50)
sCD14 (log, all values), pg/mL	75.5 (−9.5–158.2)	0.061	26.6 (−66.4–116.2)	−17.2 (−68.2–103.4)
Framingham risk score (all values), %	0.6 (−0.1–1.2)	0.107	—	—
ASCVD risk score (all values), %	1.3 (−1.3–9.5)	0.186	—	—
cIMT at baseline, mm	174.4 (75.6–273.1)	0.001∗∗	24 (−14.3–55.3)	25.3 (−2.9–49.5)

Ω Regression coefficients represent a greater or lesser change in carotid artery intima-media thickness (in millimetre) over 2 years per difference in the corresponding parameter.

cIMT, carotid intima-media thickness; INSTI, integrase strand transfer inhibitors; CV, cardiovascular; ASCVD, atherosclerotic cardiovascular disease; hsCRP, high-sensitivity C-reactive protein; sCD14, soluble CD14.

^aTraditional CVD risk factors: current smoking, hypertension, dyslipidaemia and diabetes.

^bMultivariate analysis through linear mixed model, including the significant variables associated with cIMT increase and the covariates of interest (age, sex, INSTI-containing ART, cIMT at baseline, presence of two or more cardiovascular risk factors at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer). The analysis includes all participants.

^cMultivariate analysis through linear mixed model, including the significant variables associated with cIMT increase and the covariates of interest (age, sex, INSTI-containing ART, cIMT at baseline, presence of two or more cardiovascular risk factors at baseline, and the values at each visit of systolic blood pressure, BMI, LDL-cholesterol, HDL-cholesterol, CD4 T cell percentage, hsCRP, sCD14 and D-dimer). The analysis excludes participants receiving an INSTI plus a NNRTI.

^∗Significance thresholds for main analyses were set at P < 0.05 for continuous and discrete variables.

^∗∗Significance thresholds for Bonferroni adjustment were set at P < 0.0025 for continuous variables and P < 0.003 for discrete variables.

To mitigate selection bias in assessing the impact of INSTI-based versus NNRTI-based regimens on cIMT progression, we conducted an adjusted analysis through propensity score matching. This approach excluded the participants receiving both an INSTI and a NNRTI. Using a caliper width of 0.1, two matched groups of 35 patients were selected, with similar distributions of the matching covariates age, sex, Framingham risk score, baseline CD4 T cell percentage, duration of exposure to each ART regimen, and cIMT measure at baseline (Table 4). In this analysis, no significant differences in categorical cIMT progression were observed (48.6% in the INSTI group versus 42.3% in the NNRTI group, P = 0.810). Similarly, no significant differences were observed in the 2-year change in cIMT between treatment groups [0.049 mm (−0.031–0.103) in the INSTI group versus 0.047 mm (−0.023–0.115) in NNRTI group; P = 0.647].

Table 4.

Patients’ characteristics according to the antiretroviral regimen composition after propensity score matching

	INSTI-containing regimen		NNRTI-containing regimen		P–value	SMD effect size
Patients, no.	35	(50)	35	(50)	—	—
Male sex	29	(82.9)	26	(74.3)	0.561	0.140
Age, years	47	(42–55)	47	(42–54)	0.911	−0.072
Baseline CVD risk factors, no. (%)
Any traditional CVD risk factora	22	(64.7)	20	(57.1)	0.624	0.208
≥2 traditional CVD risk factorsa	6	(17.6)	7	(20)	1.000	0.013
Current smoking	20	(58.8)	18	(51.4)	0.631	0.201
Hypertension	4	(11.4)	4	(11.4)	1.000	0.001
Dyslipidaemia	5	(14.3)	5	(14.3)	1.000	0.095
Diabetes	2	(5.7)	2	(5.7)	1.000	0.001
Framingham risk score, %	9.2	(2.8–18.2)	5.7	(3.8–12.5)	0.972	−0.016
ASCVD risk score, %	3.5	(1.4–9.2)	2.7	(1.4–6.6)	0.989	−0.043
Baseline cardio-metabolic parameters, no. (%)
Blood pressure (BP), mm Hg
Systolic	121	(110–134)	126	(114–139)	0.182	−0.308
Diastolic	72	(67–78)	70	(61–78)	0.553	0.118
BMI, kg/m2	25	(21.9–27.7)	25.2	(22.9–27.7)	0.435	−0.317
Weight, kg	73.5	(63–82.5)	74	(67.2–81.2)	0.707	−0.231
Total cholesterol, mg/dL	168	(148–179)	166	(152–207)	0.131	−0.385
LDL-cholesterol, mg/dL	97	(88–115)	110	(92–131)	0.077	−0.447
HDL-cholesterol, mg/dL	43	(37–52)	44	(37–55)	0.883	0.026
Triglycerides, mg/dL	92	(71–156)	98	(71–124)	0.911	0.200
Baseline inflammatory biomarkers
hsCRP, ng/mL	0.3	(0.2–0.6)	0.6	(0.3–1.1)	0.015	0.047
D-dimer, pg/mL	0.3	(0.2–0.4)	0.4	(0.2–0.6)	0.658	0.174
sCD14, pg/mL	4896	(4103–5658)	5076	(4460–5789)	0.219	−0.357
sCD163, pg/mL	288	(176–335)	260	(196–316)	0.690	0.280
sICAM-1, pg/mL	84.1	(64–99)	75.2	(61–96)	0.384	0.273
Baseline immunovirological status and ART
CD4 T cell count nadir, cells/mm3	220	(94–339)	252	(211–333)	0.369	0.086
CD4 T cell count, cells/mm3	732	(481–867)	600	(448–866)	0.577	0.250
CD4 T cell percentage, %	37.1	(31–42)	34	(32–41)	0.557	0.042
CD4/CD8 ratio	0.9	(0.7–1.3)	0.9	(0.7–1.2)	0.963	0.069
Months of exposure to last ART	32.8	(23.6–39.2)	39.9	(26.8–53.8)	0.102	−0.152
Baseline cIMT, mm	0.763	(0.726–0.978)	0.803	(0.713–0.940)	0.995	−0.126
Two year changes in body weight, BP and cIMT during the study
Body weight increase, kg	1.5	(−0.1–4.5)	1.6	(−2.5–4.9)	0.441	0.100
Systolic BP increase, mmHg	−1	(−14–10.2)	2	(−8.5–18.5)	0.425	0.156
cIMT increase, mmb	0.049	(−0.031–0.103)	0.047	(−0.023–0.115)	0.647	−0.114
cIMT progression, %c	17	(48.6)	15	(42.3)	0.810	0.056

The covariates for propensity matching include age, sex, Framingham risk score, baseline CD4 T cell percentage, duration of exposure to each ART regimen, and cIMT at baseline. For this analysis, participants receiving both an INSTI and a NNRTI were excluded.

INSTI, integrase strand transfer inhibitors; SMD, standardized mean difference; CVD, cardiovascular disease; ASCVD, atherosclerotic cardiovascular disease; BP, blood pressure; hsCRP, high-sensitivity C-reactive protein; sCD14, soluble CD14; sCD163, soluble CD163; sICAM-1, soluble intercellular adhesion molecule-1; cIMT, carotid intima-media thickness.

^aTraditional CVD risk factors: current smoking, hypertension, dyslipidaemia and diabetes

^bcIMT increase represents the difference in mm between median cIMT at baseline and median cIMT at 96-week.

^ccIMT progression was defined as a 6-point cIMT increase ≥ 10% and/or detection of new carotid plaques (any measurement of cIMT > 1.5 mm) during the 96-week follow-up.

Discussion

In this contemporary cohort of virologically suppressed PWH on ART, exposure to INSTI-based regimens was not associated with increased progression of subclinical atherosclerosis when compared to NNRTI. We carried out multiple analyses and thorough adjustments for potential confounding factors, including propensity score matching to ensure balance between treatment groups, and results consistently revealed no significant relationship between INSTI use and enhanced cIMT progression. Our findings align with those reported in analyses conducted through target trial emulation in both the HIV-CAUSAL and the Antiretroviral Therapy Cohort Collaborations¹⁴ and the Swiss cohort,¹³ which did not confirm the increased risk of cardiovascular event development with INSTI-based regimens observed in the respond collaboration.¹² As expected, our data support the pivotal role of traditional cardiovascular risk factors in the progression of atherosclerosis in PWH.

The role of INSTI in the cardiovascular risk of PWH is currently a topic of scientific interest and controversy, arising from emerging concerns regarding the cardio-metabolic complications associated with this class of ART. In this study, we assessed the effect of INSTI compared to NNRTI on the progression of atherosclerosis. NNRTI represent the second most common anchor drug class in contemporary ART regimens.^23,24 This family of antiretroviral drugs has historically been associated with a relatively ‘benign’ impact on the cardiovascular system, which is attributed to its favourable metabolic profile, particularly exhibited by NVP and, to a lesser extent, RPV.^25–27 Available evidence suggests that exposure to NNRTI, either as a class or through specific agents like NVP and EFV, is not associated with an increased risk of myocardial infarction.^28–30 Similarly, exposure to NNRTI was associated in a prospective study with lower progression of cIMT compared to PI.³¹ Our results suggest no differences in subclinical atherosclerosis progression between INSTI and NNRTI drug families.

To assess the progression of subclinical atherosclerosis, our analysis incorporated the development of a new carotid plaque. Carotid plaque is a strong predictor of cardiovascular disease occurrence demonstrated in both the general and the HIV population, where it has been shown to improve the cardiovascular risk stratification.^32–36 Accordingly, carotid plaque assessment using ultrasonography has been endorsed as a risk modifier/enhancer to reclassify cardiovascular risk in the guidelines on cardiovascular disease prevention in the HIV and the general population.^37,38 Existing data about the relationship between regimens containing INSTI and subclinical atherosclerosis progression are currently scarce. In a randomized clinical trial including ART-naive participants, Stein et al. compared the cIMT progression in those initiating ART based on RAL, ATV/RTV or DRV/RTV.³⁹ They did not find differences in cIMT change between antiretroviral regimens, with the exception of a lower cIMT progression at the carotid bifurcation in the ATV/RTV group. An observational prospective study analysed the cIMT increase over a 2-year period in 102 treatment-naive participants starting ART with TAF/FTC in combination with DTG, RAL or EVG.⁴⁰ No control group with a different antiretroviral class was included. To our knowledge, no studies had compared the progression of subclinical atherosclerosis between INSTI- and NNRTI-based antiretroviral regimens.

Weight gain associated with INSTI represents one of the contributing factors through which this antiretroviral class may potentially increase the risk of cardiovascular disease. Several studies have documented increased weight gain associated with INSTI-based regimens compared to NNRTI-based regimens. This phenomenon has been observed in both ART-naive individuals^41–46 and in those switching ART, primarily from regimens based on EFV.^47–50 Follow-up has usually lasted 12–24 months.^42,48–50 Noteworthy, the increase in weight with INSTI was not observed with treatment durations exceeding 2 years in the REPRIEVE trial.⁵¹ Consistent with the findings from the REPRIEVE, we observed no significant differences in weight changes between the INSTI- and NNRTI-based antiretroviral groups in our cohort, where the median duration of the antiretroviral regimens at baseline was over 2 years. Additionally, we did not observe a relationship between changes in weight and the progression of subclinical atherosclerosis. While excess weight is associated with greater risk for cardiovascular morbidity and mortality,⁵² and a higher BMI has been associated with increased cIMT in PWH in two cross-sectional studies in South Africa,^53,54 some studies have neither observed an association between BMI or weight changes and cardiovascular events in the HIV population.^55,56

During the 2-year follow-up period, a high proportion of participants in our cohort showed an increase in the cIMT, and a significant percentage developed new carotid plaques, a frequency similar to that described in other cohorts of PWH.^57,58 As expected, the traditional cardiovascular risk factors emerged as the predominant factors associated with cIMT progression. Participants experiencing subclinical atherosclerosis progression showed a distinct lipid profile, characterized by lower levels of baseline HDL-cholesterol and higher levels of LDL-cholesterol during follow-up. The group receiving NNRTI-based regimens showed higher levels of LDL-cholesterol, which was included among adjusting variables. Our analysis revealed a significant association of hypertension and blood pressure increases with progression of atherosclerosis. Remarkably, we did not observe significant differences between treatment groups in the prevalence of hypertension or blood pressure changes. The existing evidence about the relationship between INSTI and blood pressure is limited and divergent. While several studies have reported a higher incidence of hypertension or blood pressure increases associated with INSTI- compared to NNRTI- and/or PI-containing regimens, both in ART-naive^8,11 and in pre-treated PWH switching to another regimen,^9,10 other studies have not observed such an association.^15,59

In addition to traditional cardiovascular risk factors, chronic inflammation plays a critical role in the genesis and progression of atherosclerosis.⁶⁰ PWH exhibit an enhanced pro-inflammatory state characterized by higher levels of several inflammation and coagulation markers, such as hsCRP and D-dimer, which strongly predict cardiovascular disease event development within this population.^61,62 We observed a significant association of hsCRP and D-dimer levels with subclinical atherosclerosis progression in our cohort. However, we did not find differences in the levels of the biomarkers between participants receiving INSTI- and NNRTI-based regimens. Switch to INSTI from PI, NNRTI or enfuvirtide has been associated with a reduction in the levels of biomarkers of inflammation, insulin resistance and hypercoagulability.^63–66 Likewise, treatment initiation with EVG/CBT versus EFV was associated with greater declines in hsCRP, sCD14 and Lp-PLA2.⁶⁷ Of note, previous studies comparing biomarker changes with NNRTI versus INSTI primarily included EFV, whereas the predominant NNRTI administered in our cohort was RPV. EFV is associated with a more pro-atherogenic profile compared to RPV, primarily due to the lipid changes it induces.⁶⁸ As far as we know, there is a lack of head-to-head studies directly comparing biomarker changes between RPV and INSTI.

The limitations of the study are primarily associated with the non-randomized assignment of the ART regimen. The sample size also represents a limitation, despite the study being powered to detect relatively small differences in cIMT progression between groups. Consequently, if there were an increased risk of subclinical atherosclerosis progression associated with INSTI, it is likely that the effect size would be small. Some of the participants receiving INSTI were concurrently taking NNRTI, which could have influenced the impact of INSTI on cIMT progression. However, the analyses including and excluding the subgroup of participants with an INSTI plus a NNRTI showed similar results. A proportion of participants in the cohort had been exposed to other antiretroviral families prior to inclusion in the study that might have also had an effect on cIMT. Strengths include the prospective study design, with longitudinal analysis of the cIMT progression during a 2-year follow-up, and the carotid measurement conducted by a single sonographer, who was blinded to the ART composition. Participants were pre-treated with antiretrovirals, thus enabling the analysis of the impact of each antiretroviral class on a stable condition. We used stringent criteria to define cIMT progression, aimed at identifying individuals with a higher likelihood of experiencing future cardiovascular events. This variable showed a strong association with traditional cardiovascular risk factors, which reinforces the reliability and consistency of the definition utilized. Finally, we conducted multiple analyses with adjustment for several potential confounders and found consistent results.

In conclusion, among PWH on stable ART, no evidence of accelerated progression of cIMT was observed with INSTI-based compared to NNRTI-based regimens. Instead, the traditional cardiovascular risk factors continue to play a central role in atherogenesis in PWH. Continued surveillance and investigation will contribute to fully understanding the cardiovascular impact of INSTI.

Funding

This work was supported by Spanish National Plan for Scientific and Technical Research and Innovation, European Regional Development Fund (ERDF) and Instituto de Salud Carlos III (PI16/01740, PI18/01861, CM19/00160, CM20/00066, CM21/00186, CM22/00026 and PI22/01949); Centro de Investigación Biotecnológica en Red de Enfermedades Infecciosas (CB21/13/00011); Fundación para el Fomento de la Investigación Sanitaria y Biomédica de la Comunitat Valenciana (A-32 2020); Conselleria de Innovación, Universidades, Ciencia y Sociedad Digital, Generalitat Valenciana (AICO/2021/205).

Transparency declarations

The authors declare no competing interests.

Author contributions

F.G. and M.M. conceived and designed the study. J.G.-A. performed the carotid intima-media measurements. J.A.G. carried out the statistical analysis. M.M., F.G., J.G.-A., S.P., M.F.-G., V.A., P.M. and A.B. participated in patient care, investigation and data collection. M.F.-G. and V.A. performed the biomarkers determinations. M.M. and J.G.-A. wrote the first draft of the manuscript. All authors contributed to the revision of the final version of the manuscript, and approved the final version submitted.

Supplementary data

Figure S1 and Table S1 are available as Supplementary data at JAC Online.