Atherosclerosis 246 (2016) 301e308

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Randomized trial of a multidisciplinary lifestyle intervention in HIVinfected patients with moderate-high cardiovascular risk Maria Saumoy a, *, Carlos Alonso-Villaverde b, Antonio Navarro a, Montserrat Olmo a, Ramon Vila c, Josep Maria Ramon d, Silvana Di Yacovo a, Elena Ferrer a, Jordi Curto a, Antonio Vernet e, Antonia Vila a, Daniel Podzamczer a a

HIV Unit, Infectious Disease Service, Bellvitge University Hospital, Bellvitge Biomedical Research Institute, Hospitalet de Llobregat, Spain Unitat Clínica CS Llevant, Hospital Santa Tecla, Tarragona, Spain Vascular Surgery Service, Bellvitge University Hospital, Hospitalet de Llobregat, Spain d Preventive Medicine Service, Bellvitge University Hospital, Hospitalet de Llobregat, Spain e Department of Mechanical Engineering, Universitat Rovira i Virgili, Tarragona, Spain b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 July 2015 Received in revised form 29 December 2015 Accepted 9 January 2016 Available online 11 January 2016

Objective: To assess the impact of a multidisciplinary lifestyle intervention on cardiovascular risk and carotid intima-media thickness (c-IMT) in HIV-infected patients with Framingham scores (FS) > 10%. Design: Randomized pilot study; follow-up 36 months. Methods: Virologically suppressed adult HIV-1-infected patients with FS >10% were randomized 1:1 to the intervention group (multidisciplinary lifestyle intervention) or control group (routine care). At baseline and months 12, 24 and 36, lipid parameters were analyzed and carotid ultrasound was performed to determine c-IMT and presence of plaques. Biomarkers were measured at baseline and month 36. The primary endpoints were lipid and FS changes at 36 months. Results: Fifty-four patients were included, 27 in each arm. Median age was 50.5 years, all patients but one were men, and FS was 16.5%. Relative to controls, total and LDL cholesterol had significantly decreased in the intervention group at 24 months (p ¼ 0.039, p ¼ 0.011, respectively). However, no differences between groups were found at month 36 in lipid variables, neither in FS. Tobacco use decreased in the intervention group (p ¼ 0.031). At baseline, 74.5% of patients had subclinical atherosclerosis, and at month 36, we observed a progression in c-IMT that was greater in the intervention group (p ¼ 0.030). D-dimer increased (p ¼ 0.027) and soluble intercellular adhesion molecule-1 decreased (p ¼ 0.018) at 36 months. Conclusions: In this cohort of HIV-infected patients with FS>10% and a high percentage of subclinical atherosclerosis, a multidisciplinary lifestyle intervention resulted in a slight improvement in some cardiovascular risk factors and the FS during the first 2 years, but did not prevent c-IMT progression. © 2016 Elsevier Ireland Ltd. All rights reserved.

Keywords: Lifestyle intervention Cardiovascular risk Carotid intima-media thickness Carotid plaques Biomarkers

1. Introduction HIV-infected patients are at a higher risk of developing cardiovascular disease than the general population [1,2]. HIV infection itself, combined antiretroviral therapy (c-ART), and traditional cardiovascular risk factors interact in the pathogenesis of atherosclerosis [3]. HIV is associated with persistent inflammation and

* Corresponding author. HIV Unit, Infectious Disease Service, Hospital Universitari de Bellvitge, C/Feixa Llarga s/n, 08907 Hospitalet de Llobregat, Spain. E-mail address: [email protected] (M. Saumoy). http://dx.doi.org/10.1016/j.atherosclerosis.2016.01.014 0021-9150/© 2016 Elsevier Ireland Ltd. All rights reserved.

immune activation. As compared to healthy controls, several biomarkers related to the pathogenesis of atherosclerosis are elevated in treatment-naïve HIV-infected patients. Initiation of c-ART leads to a decrease in some of these markers, whereas others remain elevated or higher than in controls [4e6]. In addition, c-ART, mainly protease inhibitors and some nucleoside reverse transcriptase inhibitors, are associated with disturbances in lipid metabolism (elevated total cholesterol, low-density lipoprotein cholesterol, and triglycerides) and glucose metabolism (insulin resistance, glucose intolerance, and diabetes) [7]. Finally, traditional cardiovascular risk factors such as dyslipidemia, hypertension, and diabetes are more common and present at a younger age in HIV-infected

302

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

patients than in the general population [1]. Lifestyle interventions are the first step in cardiovascular risk management [8], but information is limited regarding the impact of these interventions on metabolic disorders [9e11] and subclinical atherosclerosis [12] in HIV-infected patients. We conducted a pilot study in which patients with estimated cardiovascular risk greater than 10% determined by the Framingham Score (FS) were randomized to a multidisciplinary lifestyle intervention (intervention group) versus routine care (control group) with the aim of assessing the impact of the intervention on the patients’ cardiovascular risk status, lipid profile, carotid intimamedia thickness (c-IMT), and cardiovascular biomarkers over a 36month follow-up. 2. Methods 2.1. Study design, patient population, and endpoints This is a randomized, controlled pilot trial carried out in the HIV unit of a university teaching hospital. The inclusion criteria were documented HIV infection, age older than 18 years, stable antiretroviral regimen, undetectable viral load for the previous 3 months, and estimated cardiovascular risk greater than 10% based on the FS. The exclusion criteria were previous cardiovascular disease, diabetes mellitus, microalbuminuria, dyslipidemia due to hypothyroidism, nephrotic syndrome or renal insufficiency (estimated glomerular filtration measured by the Cockcroft-Gault equation 6.5%), the patient was referred to a general practitioner or endocrinologist. Tobacco use was recorded at each visit.

2.3. Laboratory assessment Venous blood samples were obtained after an overnight fast every 4 months. Total cholesterol (TC), high-density lipoprotein cholesterol (HDL-c) and triglycerides (TG) were measured using an enzymatic colorimetric method. LDL-c was calculated using the Friedewald equation (LDLc ¼ TC e HDLc e (TG/2.17)). At baseline, and months 12, 24, and 36, insulin was measured by an immunochemiluminescent assay (Immulite 2000, Siemens Health Care Diagnostics). At baseline and month 36, a blood sample was centrifuged and plasma was refrigerated at 70  C for subsequent analysis. The following biomarkers were measured on a Luminex 200 xMAP system (Millipore Corporation, Billerica, MA, USA): inflammatory biomarkers (interleukin-6 [IL-6], tumor necrosis factor-alfa [TNF-a], high-sensitivity C-reactive protein [hsCRP], fibrinogen), endothelial biomarkers (monocyte chemoattractant protein-1 [MCP-1], soluble intercellular adhesion molecule-1 [sICAM-1], L-selectin, and fractalkine), coagulation biomarkers (sCD40, D-dimer, and plasminogen activator inhibitor-1 [PAI-1]) and adiponectin. Asymmetric dimethylarginine (ADMA) was measured using a double-antibody sandwich enzyme-linked immunosorbent assay (ELISA) (Central Plains Road, Yangpu District, Shanghai, China). Soluble CD163 was measured by ELISA (Boston Biochem, 840 Memorial Drive, Cambridge, MA, USA). The estimated overall cardiovascular risk over the next 10 years was calculated using the FS [13].

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

2.4. Ultrasound measurement of carotid intima-media thickness High-resolution B-mode ultrasound imaging of the carotid arteries was performed at baseline and once a year, using a Siemens Antares Ultrasound system with a 5e10 MHz linear array probe. Patients were placed in supine position with the neck extended and the head slightly rotated toward the side opposite to the side examined. Intima-media thickness at the far wall of the distal common carotid artery was assessed using ArtWAS software, analyzing the segment 1 cm proximal to the bifurcation. This MATLAB-based software allows semiautomatic measurement of cIMT [14]. Focal atherosclerotic plaque was excluded, and the average of the right and left measurements was used in the analysis. Comprehensive screening of the carotid bifurcation for plaque was performed starting at 3 cm proximal to the flow divider and covering up to 1 cm distal to the flow divider (corresponding to the internal carotid). Special care was taken at each follow-up examination to reproduce the same image of the bifurcation as that obtained in the baseline study so that c-IMT and atherosclerotic plaque would be determined in the same location. Subclinical atherosclerosis was defined as the presence of a plaque (focal structure within the arterial lumen of at least 0.5 mm or 50% of the surrounding IMT value or c-IMT 1.5 mm) in the common carotid, carotid bulb, or internal carotid, or common c-IMT greater than the 75th percentile of an age- and sex-adjusted reference population [15,16]. 2.5. Ankle-brachial blood pressure index The ankle-brachial blood pressure index (ABI) was measured using a bidirectional Doppler ultrasound blood flow detector (Dopplex SD2 Hutleigh Healthcare) and sphygmomanometer. Systolic arterial pressure was obtained from the left and right brachial, posterior tibial, and dorsalis pedis arteries. ABI was calculated as the ratio between the highest systolic pressure reading in the legs (dorsalis pedis or posterior tibial) and the highest pressure in the arms. An ABI measurement 3.37 mmol/L (66.7% vs 33.3%; p ¼ 0.008), whereas in the control group there were no significant changes (46.2% vs 25%; p ¼ 0.508). Differences between arms were not significant. No changes in the body mass index, glucose, or insulin levels were observed in either arm or between arms at month 36. 3.3. Cardiovascular risk assessment A significant decrease relative to baseline in the FS-estimated cardiovascular risk was observed in the intervention arm at months 12 (p ¼ 0.046) and 24 (p ¼ 0.009), but there were no significant changes within or between arms at week 36. After correcting for the RTM effect, similar result were obtained (p ¼ 0.051 and p ¼ 0.001, at month 12 and 24 respectively) (Fig. 2). With regard to tobacco use, 81.5% (22/27) of intervention participants were smokers at baseline versus 54.2% (13/24) at month 36 (p ¼ 0.031). All patients had stopped smoking during the first year, but 2 restarted tobacco consumption, which yielded an abstinence rate at month 36 of 31.6%. In the control group, 3 participants had stopped smoking by month 36 (66.7% at baseline and 54.2% at month 36; p ¼ 0.625); the abstinence rate was 20%. There were no significant differences between arms (p ¼ 0.697). Systolic and diastolic blood pressure showed a tendency to increase in the intervention group at month 36 (13 mmHg [50 to 47]; p ¼ 0.073 and 2 mmHg [15 to 27]; p ¼ 0.177, respectively), with no changes in the control group (1.5 mmHg [34 to 40];

304

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

Table 1 Baseline characteristics of the participants. Variable Age, years Sex, men, n (%) HIV acquisition, n (%) Heterosexual Homosexual Injecting drug use Other/unknown Current ART, n (%) PI NNRTI Duration of c-ART exposure, months CD4 cell count, cells/mL Cardiovascular risk factors Current smoker Hypertension Family history of CHD Cardiovascular risk estimation, FS >20%, n (%) Ankle-brachial index 5.18, n (%) HDL-c, mmol/L HDL-c 3.37, n (%) TC/HDL-c LDL-c/HDL-c Triglycerides, mmol/L Triglycerides >1.7, n (%) Glucose, mmol/L Glucose > 6.9, n (%) Insulin, pmol/L Insulin >120, n (%) Common c-IMT, mm c-IMT > P75, n (%) Presence of carotid plaques, n (%) Common carotid or bulb Common carotid Bulb Subclinical atherosclerosis, an (%) Plasma biomarkers Interleukin-6, pg/mL TNF-a, pg/mL hs-CRP, mg/L Fibrinogen, ng/mL MCP-1, pg/mL sICAM-1, ng/mL L-Selectin, ng/mL Fractalkine, ng/mL sCD40, ng/mL D-dimer, ng/mL Adiponectin, ug/mL CD163, ng/mL ADMA, Umol/mL

Control group n ¼ 27

Intervention group n ¼ 27

p-value

51 (39e72) 27 (100)

50 (41e76) 26 (96.3)

0.856 —

6 12 8 1

7 (25.9) 15 (55.6) 5 (18.5)

0.647

(22.2) (44.4) (29.6) (3.7)

0

9 18 118 483

(33.3) (66.7) (32e142) (36e1343)

11 17 108 651

(40.7) (63) (25e156) (149e1558)

0.573 0.776 0.533 0.058

18 6 4 16 6 1.05 4 136 80 26.7

(66.7) (22.2) (14.8) (10e27) (22.2) (0.5e1.33) (14.8) (100e180) (60e104) (21e36.2)

22 11 4 17 7 1 4 135 80 25.9

(81.5) (40.7) (14.8) (10e30) (25.9) (0.5e1.38) (14.8) (106e177) (60e100) (18.9e38.4)

0.214 0.143 1 0.108 0.750 0.899 1 0.853 0.301 0.566

(51.9) (44.4) (11.1) (37)

0.613

5.93 24 1.08 10 3.59 18 5.88 3.10 2.4 22 5.5

(5.69e6.58) (88.9) (1e1.26) (37) (3.18e4.09) (66.7) (5.27e6.57) (1.37e4.62) (1.8e4.2) (81.5) (5e6)

0.099 0.467 0.820 0.776 0.138 0.132 0.326 0.128 0.957 0.340 0.945

108 12 0.73 13

(64e154) (44.4) (0.47e1.17) (48.1)

0.931 0.402 0.699 0.413

(70.4) (14.8) (70.4) (85.2)

0.260 1 0.260 1

(1.1e11.4) (3.2e31.9) (0.7e41.2) (0.6e11.5) (349e1862) (49.1e286) (0.4e5.8) (8.9e270) (1.3e37.8) (79e332) (3.5e65) (192e2267) (0.42e13.31)

0.842 0.243 0.551 0.219 0.748 0.264 0.271 0.924 0.287 0.210 0.264 0.574

11 10 1 5

(40.7) (37) (3.7) (18.5)

5.71 21 1.11 9 3.16 12 5.26 2.80 2.7 19 5.6

(5.24e6.12) (77.8) (0.95e1.43) (33.3) (2.65e3.84) (46.2) (4.21e6.69) (1.08e4.77) (1.6e4.2) (70.5) (5e6.2)

67 9 0.74 16

(45e144) (33.3) (0.4e1.07) (59.3)

0

14 12 3 10

0.129

0

15 4 15 22 2.5 12.2 4 1 940 112.6 1.3 40.4 23.9 149.5 12 644 0.92

(55.6) (14.8) (55.6) (81.5) (1.3e10) (6e25.9) (0.6e130.9) (0.1e6.9) (371e1562) (51.6e306) (0.2e4.4) (5e549) (2.6e37.9) (78e356) (0.9e83.8) (209e1243) (0.41e1.91)

19 4 19 23 2.4 10.8 7.8 1.5 753 105.8 1.46 31.5 24.1 185 7.7 554 0.91

Values for the quantitative variables are expressed as the median (minimumemaximum). ART, antiretroviral therapy; CHD, coronary heart disease; c-IMT, carotid intima-media thickness; FS, Framingham score; HDL-c, high-density lipoprotein cholesterol; hs-CRP, high-sensitivity C-reactive protein; LDL-c, low -density lipoprotein cholesterol; MCP-1, monocyte chemoattractant protein-1; NNRTI non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; sICAM-1, soluble intercellular adhesion molecule-1; TC, total cholesterol; TG, triglycerides. TNF-a, tumor necrosis factor-alfa. a Presence of a plaque in the common carotid or carotid bulb, or common c-IMT >75th percentile of a reference population, adjusted by age and sex.

p ¼ 0.915 and 1 mmHg [-24 to 26]; p ¼ 0.258, respectively). There were no differences between arms. Five participants started

antihypertensive therapy in the control group and none in the intervention group.

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

∆ at month 12 1.0

305

∆ at month 24

∆ at month 36

** Differences between groups: - At month 12 in TG (p=0.024 ) - At month 24 in TC (p=0.039) and LDL-c (p=0.011 )

Median ∆ (mmol/L)

0.5 .27 .15 .06

.08

.06

0.0

-.01 -.05

-.02 -.01

-.06

-.07

-.03 -.14

-.26

-0.5

-.38

-.34

-.32 0.008 *

-.37

-.67 0.026 *

-.78 0.012 *

-.30

-.60 0.004 *

-.70 0.021 *

-.76 0.001 *

-1.0 TC

LDL-c

HDL-c

TG

TC

Control group

LDL-c

HDL-c

TG

Intervention group

* Statistically significant intra-group differences from baseline: Wilcoxon test or t-test for repeated measures. ** Statistically significant differences were found between the intervention group and controls: Mann Whitney U test or t-test for independent

HDL-c, high-density lipoprotein cholesterol; LDL-c, low -density lipoprotein cholesterol; TC, total cholesterol; TG, triglycerides. Fig. 1. Median changes in plasma lipid values from baseline at months 12, 24, and 36 in the intervention and control group.

Fig. 2. Cardiovascular risk estimation measured by the Framingham Score at baseline and months 12, 24, and 36 in the intervention and control group.

3.4. Cardiovascular biomarkers Plasma cardiovascular biomarkers were measured at baseline and month 36. There were no differences between groups at

baseline. At month 36 in the total sample, a decrease in sICAM-1 (18.1 ng/mL [147 to 161]; p ¼ 0.018) and increases in D-dimer (20.8 ng/mL [143.5 to 299.1]; p ¼ 0.027) and ADMA (0.09 umol/ mL [12.7 to 0.45]; p ¼ 0.057) were found relative to baseline, with

306

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

no differences between arms. Fractalkine increased only in the intervention group (5.4 [103.1 to 127.1]; p ¼ 0.061) (between arms, p ¼ 0.049). There were no significant changes within or between arms for any of the other biomarkers analyzed (Table 2). 3.5. Carotid ultrasound findings and ankle-brachial index At baseline, subclinical atherosclerosis was detected in 45 (83.3%) participants. The annual delta c-IMT was calculated only in patients who had undergone carotid ultrasound examination at the 4 study time points (n ¼ 32; 18 in the intervention group and 14 in the control group). There were no significant differences in baseline characteristics between this subgroup and the overall group. The mean (SD) annual delta c-IMT was 0.032 (0.053) mm/year, and a significant difference was found between arms: 0.050 mm/year in the intervention group and 0.009 mm/year in the control group (p ¼ 0.030). During the study, 20 participants (62.5%) showed c-IMT progression (86.4% in the intervention group and 61.9% in the control group; p ¼ 0.066). Furthermore, an increase in new carotid plaque formation was seen at month 36 (63% baseline vs. 89% month 36; p ¼ 0.004), with no differences between arms. The percentage of patients with subclinical atherosclerosis did not change within or between arms at month 36. The variables related to c-IMT changes in the univariate and multivariate analyses are shown in Table 3. Baseline D-dimer, hsCRP, fibrinogen, and presence of carotid plaque at baseline were associated with c-IMT progression on univariate analysis. Nevertheless, only older age, a higher baseline TC, and changes in the TC/ HDL-c ratio and CD4 cell count remained associated with c-IMT progression on multivariate analysis. The remained unchanged in the two groups over the study period (ABI ¼ 0.628 and p ¼ 0.529). 4. Discussion In this 3-year study performed in HIV-infected patients with moderate to high cardiovascular risk, a multidisciplinary lifestyle intervention was associated with a favorable change in several lipid variables and contributed to improving efforts to stop smoking. This strategy also achieved a transient decrease in cardiovascular risk assessed by the FS. However, the high burden of subclinical atherosclerosis at baseline worsened over the study period in a significant number of participants, and the intervention was not

associated with improvements in cardiovascular disease-related biomarkers. Lifestyle interventions including dietary counseling and exercise have been associated with improvements in blood lipids, blood pressure, and glycemia parameters in the general population [10,19], and are considered the first step in the management of cardiovascular risk factors. At the end of our 3-year study, the intervention imparted was associated with a small improvement in TC and LDL-c, but there were no changes in triglycerides or glycemia. There is little published data on the efficacy of lifestyle interventions on plasma lipids in HIV-infected patients [9,10]. One study showed that a lifestyle intervention did not improve lipid variables in HIV patients with metabolic syndrome, although a decrease in waist circumference and blood systolic pressure was achieved [9]. Studies reporting the benefits of a dietary intervention on lipid measurements such as this one and others performed in the general population, had a duration of 3e12 months [9,20,21]. In the present study, we observed a significant improvement in lipids and the FS in the group undergoing the intervention during the first and second year of the study. The worsening observed during the third year may be attributable to exhaustion in following the lifestyle recommendations. This reinforces the idea that patients should be closely followed up and ways should be sought to stimulate them to maintain the recommendations to achieve longterm success. Inflammation has been related to achieving less marked reductions in TC and LDL-c after a low-fat/low-cholesterol diet [22]. At baseline in our study, CRP was higher than 1 mg/L in 90.6% of patients and D-dimer was greater than 126 ng/mL in 62.5%; both cut-offs being associated with cardiovascular risk. Furthermore, these values did not improve and some even worsened over the study period, reflecting a proinflammatory environment that may be related to the fact that only small lipid improvements were observed in our study [23,24]. Of note, smoking cessation was achieved in 31.6% of participants in the intervention group, a slightly higher percentage than has been reported previously in HIV-infected patients [25]. Another finding worth mentioning is the high percentage of participants (83.3%) with subclinical atherosclerosis at baseline, which concurs with the results of a previous study in which 76% of HIV-infected patients with FS >10% had subclinical atherosclerosis [26]. In our study, atherosclerosis progressed at a mean c-IMT increase of 0.032 mm/year, similar to the value reported by Hsue et al.

Table 2 Changes in plasma biomarkers at month 36. Variable

Interleukin-6, pg/mL TNF-a, pg/mL hs-CRP, mg/L Fibrinogen, ng/mL MCP-1, pg/mL sICAM-1, ng/mL L-Selectin, ng/mL Fractalkine, ng/mL sCD40, ng/mL D-dimer, ng/mL Adiponectin, ug/mL CD163, ng/mL ADMA, Umol/mL

Control group, n ¼ 24

Intervention group, n ¼ 23

D at month 36

p-value

0.0 0.09 1.41 0.13 98.5 22.7 0.23 2.3 1.22 20.3 0.38 22.6 0.05

0.689 0.380 0.086 0.170 0.108 0.317 0.797 0.349 0.819 0.060 0.391 0.290 0.536

(5.8 to 8.3) (14.4 to 4.9) (108 to 84) (2.1 to 1.9) (851 to 389) (147 to 161) (2.6 to 2.5) (201 to 83) (30 to 31) (120 to 299) (40 to 11) (307 to 406) (0.6 to 0.4)

a

D at month 36

p-value

0.03 0.86 1.43 0.16 53 16.6 0.08 5.42 7.3 26.1 0.2 11.1 0.12

0.355 0.688 0.465 0.394 0.627 0.012 1 0.061 0.275 0.188 0.808 0.738 0.055

(2.7 to 10.3) (13.1 to 12.4) (32 to 37) (10.3 to 3.8) (1023 to 649) (127 to29.8) (4.7 to 3.6) (103 to 127) (31 to 28) (143 to 261) (19 to 19) (446 to 380) (12.7 to 0.4)

p-valueb

Overall (n ¼ 47) a

D at month 36

p-value

0.03 0.14 0.37 0.04 3.1 18.1 0.05 3.1 4.1 20.8 0.3 19.6 0.09

0.491 0.819 0.505 0.735 0.462 0.018 0.751 0.688 0.179 0.027 0.440 0.295 0.057

(5.8 to 10) (14.4 to 12.5) (108 to 84) (10.3 to 3.8) (1023 to 649) (147 to 161) (4.7 to 3.6) (201 to 127) (31 to 31) (143 to 299) (40 to 19) (446 to 406) (12.7 to 0.4)

a

0.890 0.255 0.085 0.180 0.150 0.398 0.915 0.049 0.173 0.573 0.733 0.567 0.655

Values are expressed as the median (minimumemaximum). ADMA, asymmetric dimethylarginine; hs-CRP, high-sensitivity C-reactive protein; MCP-1, monocyte chemoattractant protein-1; PAI-1, plasminogen activator inhibitor-1; sICAM-1, soluble intercellular adhesion molecule-1; TNF-a, tumor necrosis factor-alfa. a p-value of the change at month 36 within each group. b p-value of the change at month 36 between groups.

M. Saumoy et al. / Atherosclerosis 246 (2016) 301e308

307

Table 3 Univariate and multivariate linear regression analyses of variables related to c-IMT change at month 36. Variable

Univariate analysis, b (p-value)

Multivariate analyses b (p-value)

Age Treatment group Baseline diastolic blood pressure Baseline D-dimer Baseline hs-CRP Baseline fibrinogen D TC/c-HDL D Glucose D Systolic blood pressure D Diastolic blood pressure D CD4 cell count Presence of carotid plaques at baseline Baseline total cholesterol higher than 5.18 mmol/L

0.004 0.110 0.005 0.001 0.006 0.028 0.002 0.072 0.003 0.006 0.0004 0.044 0.133

0.006 (0.009)

(0.135) (0.048) (0.020) (0.024) (0.010) (0.025) (0.098) (0.092) (0.027) (0.007) (0.020) (0.043) (0.037)

0.002 (0.041)

0.001 (