Abstract
Dear Editor:
Patients with HIV infection have an increased mortality, largely attributable to cardiovascular disease (CVD). 1 In addition to traditional and not traditional cardiovascular risk factors, HIV infection and antiretroviral therapy (ART) have also been implicated in the premature development of atherosclerosis and coronary heart disease. 2,3 There is the need to stratify the risk of coronary events in HIV patients. Framingham risk score is an extensively studied index to predict cardiovascular risk in the general population. 4 It includes current smoking status, gender, age, cholesterol concentrations and blood pressure, and estimates the risk of coronary events by stratifying individuals into three risk categories: low (<10% risk of an event in 10 years), intermediate (10–20%), and high (>20%). Framingham risk score could underestimate prediction in HIV-infected patients 5,6 because there are many aspects that contribute to it. A inflammatory reaction associated with HIV infection may play a role in the atherosclerotic process. 7 In addition, ART is associated with metabolic abnormalities that increase cardiovascular risk.
Endothelial dysfunction and increased intima medial thickness (IMT) are considered an early event in the development of atherosclerosis. 8 In general population have shown to correlate with coronary atherosclerosis, 9 and have been directly associated with an increased risk of myocardial infarction and stroke in older adults without a history of CVD. 10
We aimed to investigate whether Framingham risk score predict subclinical atherosclerotic disease (SAD) in HIV infected patients with low cardiovascular risk. We also studied the prevalence of SAD, the association of SAD with cardiovascular risk factors and HIV-related factors, and its relationship with plasma levels of several proatherosclerotic biomarkers.
The study was conducted at the HIV outpatient clinic of the University Hospital Reina Sofia of Murcia, Spain. Consecutive healthy HIV-infected adults visiting from January 2009 to June 2009 were invited to participate. The investigation was approved by the Ethics Committee for Clinical Research and the patients' informed consent was obtained.
Details were taken for age, HIV transmission category, stage of HIV infection according to Center for Disease Control and Prevention (CDC) criteria, duration of ART, current antiretroviral medication, coinfection with hepatitis B or C virus, cardiovascular risk factors and pharmacologic treatment of dyslipidemia, diabetes, or hypertension.
Blood samples were collected after an 8-h overnight fast for measurement of glycemia, total cholesterol, high-density lipoprotein (HDL) cholesterol, direct low-density lipoprotein (LDL) cholesterol, triglycerides, creatinine, insulinemia, CD4+ T-cell count, and HIV plasma viral load. An additional sample was processed by centrifugation. Plasma aliquots obtained were stored at −80°C. All frozen samples were subsequently defrosted and plasma levels of several proatherosclerotic biomarkers were simultaneously measured. More precisely, we determined levels of vascular cell adhesion molecule (VCAM), secretory phospholipase A2 (SPLA-2), thiobarbituric acid reactive species (TBARS), superoxide dismutases (SOD), resistin, adiponectin, high-sensitivity C-reactive protein (hsCRP), and antioxidant capacity. Plasma concentrations of sVCAM-1, SPLA-2, adiponectin, and hsCRP were measured using commercially available enzyme-linked immunosorbent assay (ELISA) kits. TBARS, SOD, and antioxidant capacity were measured using colorimetric assay.
The 10-year risk of developing myocardial infarction or coronary death was calculated for each patient with the Framingham equation.
To determine SAD we measured endothelial function and common and bulb carotid intima media thickness (c-IMT). SAD was considered if common c-IMT >0.8 mm or bulb c-IMT >1 mm or flow-mediated dilatation (FMD) <5%.
Endothelial function was evaluated measuring FMD. Patients were required to fast and not use any tobacco-containing products for 8 h before the study. Patients were placed in a supine position and a blood pressure cuff was placed on the widest part of the proximal right forearm, approximately 1 cm distal to the antecubital fossa. Using a high-resolution (≥7 MHz) linear array vascular ultrasound transducer (Philips iE33, Philips, Andover, MA), the brachial artery was located above the elbow and scanned in longitudinal sections. Baseline vessel diameter was determined with the mean of three measures. The forearm cuff was inflated to 240 mm Hg for 5 min to induce reactive hyperemia. FMD of the brachial artery was measured 1 min after cuff deflation as described above. After a 10-min rest period, nitroglycerine-mediated vasodilatation (NTGMD), a marker of endothelium-independent vasodilatation, was measured 4 min after the administration of sublingual nitroglycerine (400 μg). FMD was calculated as the ratio of the brachial artery diameter after reactive hyperemia to baseline diameter and was expressed as a percentage of change. NTGMD was calculated in an analogous fashion. Measurements were performed by a single sonographer blinded to patient information and treatment.
For determining the carotid IMT, B-mode high-resolution ultrasound was used following a standard procedure as described previously. 11 To quantify the degree of carotid artery wall thickening, the mean of six measures performed in the posterior wall of the left common and bulb carotid artery was taken. IMT was considered elevated whether was higher than 0.8 mm in common or higher than 1 mm in bulb carotid artery.
Descriptive analysis of the baseline characteristics of patients was performed using frequency distributions for categorical variables and median and standard deviation (SD) for continuous variables, respectively. Differences in demographic and clinical characteristics between patients with and without SAD were assessed using the χ2 or Fisher's exact test for categorical variables and Mann–Whitney test for continuous variables.
Binary logistic regression analysis was carried out using forward logistic regression models to obtain an adjusted measure of the effect of cardiovascular related variables and ART on SAD. A p value of <0.05 was considered to be statistically significant.
To evaluate concordance index between Framingham risk score and SAD we used κ index. The discriminative power of other Framingham risk cutoff points was also explored using receiver operating characteristic (ROC) curves and the SAD values as gold standard. Statistical analyses were performed using SPSS, version 17.0 (SPSS Inc., Chicago, IL).
One hundred fifty-three patients were evaluated, 109 (71.2%) of whom had Framingham risk score lower than 10%.
The main baseline characteristics (mean [SD] unless otherwise indicated) of the 153 patients included in the study were: age, 44.2 (9.8) years; male gender (n [%]), 131 (85.6); current antiretroviral therapy (n [%]), 134 (87.5); time of exposure to antiretroviral therapy, 6.5(6) years; HIV-RNA viral load less than 50 copies per milliliter (n [%]), 109 (71.2); CD4 cell count, 565.8 (311.16) cells/mL; lipodystrophy (n [%]), 31 (20.2); body mass index, 24.9 (4.5) kg/m2; LDL cholesterol, 106.1 (38.67) mg/dL; HDL cholesterol, 46.06 (14.98) mg/dL; current smoking (n [%]), 89 (58.16); hypertension (n [%]), 49 (32); family history of early coronary disease (n [%]), 38 (24.8).
Sixty-seven (56% [95 % CI: 48.2–63.8]) patients had SAD, 22 (14.4%) patients had common c-IMT >0.8 mm, 38 (24.8%) had bulb c-IMT >1 mm, and 34 (22.2%) patients had FMD <5%. Patients with SAD were older (48.2 [9.8] versus 41.1 [8.6] years; p<0.001), had higher levels of tryglicerides (195.5 [136.7] mg/dL versus 155.4 [90.1] mg/dL; p=0.009), were more likely to receive ART (91% versus 74.7%; p=0.036) and PI (84.05% versus 33.7%; p=0.023), had previous PI exposure (62.3% versus 42%; p=0,009), had lipodystrophy (34.9% versus 11.39%; p=0.001), had lipoatrophy (29% versus 8.8%; p=0.002) and had higher Framingham risk score (9.1 [7.8] versus 5.97 [6.41]; p=0.007) and were more likely to have Framingham risk score higher than 10% (38% versus 20.6%; p=0.013) than those without SAD.
Forty-one patients (37.6% [95% CI: 32.4–41.6]) with Framingham risk score lower 10% had SAD. Comparing with patients with higher Framingham risk score, χ2=5.87; p=0.015.
Nine (8.2%) patients had common c-IMT >0.8 mm, 19 (17.4%) had bulb c-IMT >1 mm, and 21 (19.26%) had FMD <5%. Patients with Framinghan risk score lower 10% and SAD were older (44 [6] versus 39 [7] years; p<0.001), had higher abdominal circumference (90.6 [11.9] cm versus 83.35 [8.56] cm; p=0.009), were more likely to receive PI (51.21% versus 34.8%; p=0.05), had previous PI exposure (58.53% versus 36,36%; p=0,02), had lipodystrophy (32.43% versus 11.86%; p=0.015), had lipoatrophy (27% versus 10.1%; p=0.032) and had metabolic syndrome (35% versus 18.4%; p=0.048) than those without SAD.
Univariate and multivariate analysis of factors associated with SAD are shown in Tables 1 and 2.
All numeric variables are presented as means (standard deviation [SD]) and no. (%) of patients.
SAD, subclinical atherosclerotic disease; NNRTI, non nucleoside reverse-transcriptase inhibitor; PI, protease inhibitor; NRTI, nucleoside reverse-transcriptase inhibitor.
SAD, subclinical atherosclerotic disease OR, odds ratio; CI, confidence interval; ART, antiretroviral therapy; PI, protease inhibitors.
Patients with previous PI exposure had higher levels of VCAM (1604.74 [1130.69] ng/mL versus 926.33 [428.08] ng/mL; p=0.01] and resistin (5.63 [2.09] ng/mL versus 4.47 [1.35] ng/mL; p=0.024) than those without previous PI exposure. Patients receiving ART had lower levels of VCAM (1075.7 [755.16] ng/mL versus 1776.06 [828.41] ng/mL; p=0.008) than those without ART. Patients with lipodystrophy had lower levels of SOD (0.25 [0.11] U versus 0.33 [0.14] U; p=0.05) and adiponectin (7.06 [3.4] μg/mL versus 9.06 [3.32] μg/mL; p=0.05) than those without lipodystrophy. Age was directed correlated with VCAM (correlation coefficient Spearman's ρ 0.26; p=0.018). Not associations were found with others proatherosclerotic biomarkers.
There was a low level of concordance between Framingham risk score and SAD (κ index 0.186; [p=0.015]). According to ROC curve (area under the curve 0.636 (95% confidence interval [CI], 0.548–0.724), SD 0.045], Framingham risk >10% had 38% sensitivity and 79.1% specificity to predict SAD; Framingham >5% had 56.7% sensitivity and 64% specificity to predict SAD; and Framingham >2% had 71.6% sensitivity and 47.7% specificity to predict SAD.
The most important finding in this study is that Framingham risk score is not useful to predict SAD in HIV-infected patients, mainly in patients with low cardiovascular risk, and could underestimate cardiovascular risk in those patients. Patients with low Framingham risk score had high prevalence of SAD and increased coronary artery disease risk. Age, receiving ART, and having lipodistrophy were independently associated with SAD, and previous PI exposure was associated in patients with low cardiovascular risk. Several proatherosclerotic biomarkers, mainly VCAM, adiponectin, resistin, and SOD were associated with age, ART, lipodistrophy, and previous PI exposure indicating that those factors could develop atherosclerosis through a proinflamatory process and/or modifying oxidative stress.
In conclusion, patients with low cardiovascular risk according to Framingham risk score could have increased cardiovascular disease risk. Additional risk factors remain for patients with HIV infection. Our analyses suggest the need to develop specific cardiovascular risk equations for people with HIV infection both receiving and not receiving ART. Clinicians should carefully monitor the risk of CVD in their patients receiving ART and intervene appropriately to reduce these risk factors.
Footnotes
Acknowledgment
Supported in part by Instituto de Salud Carlos III (Fondo de Investigaciones Sanitarias-FIS-EXP PI08/90914) and proyectos FEDER.