ABC | Volume 114, Nº1, January 2019

Original Article Leite et al. Carotid thickness in HIV-infected patients Arq Bras Cardiol. 2020; 114(1):90-97 Method Study subjects Cross-sectional analytical study with 115 patients was conducted at Hospital das Clínicas – Federal University of Pernambuco, Northeast, Brazil. Individuals were enrolled by convenience sampling. Ninety-nine patients were infected with HIV (HIV group) and were attended at the Specialized HIV/AIDS Healthcare Service, other 16 individuals were healthy and used as control (non-HIV group); both groups were aged between ≥ 18 and ≤ 60 years. All HIV patients were under ART with two NRTIs analogues and one NNRTI started at any time from their diagnosis, had undetectable HIV‑1 RNA viral load, and were not on therapy for dyslipidaemia. Healthy controls were followers of patients attending in the Urology Service of the same hospital. Low risk for cardiovascular disease was also an inclusion criterion for both groups, calculated by the Framingham Risk Score (FRS). FRS estimates the likelihood of myocardial infarction or death from coronary disease within 10 years in individuals without prior clinical atherosclerosis. Risk calculation uses parameters such as gender, age, total and HDL cholesterol levels, systolic blood pressure, and smoking status. 7 Data collection After patients signed the informed consent form, data were collected with standardized questionnaires, based on medical records and/or interview information as follows: age, gender, race, ART type and time, CD4+ T cells count, HIV-1 RNA viral load, and smoking and diabetes status. CD4+ T-cell counts were estimated with flow cytometry using the FACSCalibur (Becton-Dickinson, USA) and results were expressed in cells/mm 3 . HIV viral load was measured using real-time polymerase chain reaction (RT-PCR) (Roche Diagnostics, Germany) with detection limit of 50 copies/mL. Afterwards, the examinations of lipidogram, the measurements of carotid intima-media thickness (CIMT), and the assessment of inflammatory biomarker levels were carried out. The moment the patient was included in the study, blood was collected for lipidogram and inflammatory biomarker determinations. Blood pressure assessment and carotid Doppler ultrasound were performed as well. Lipidogram Total cholesterol, HDL, and triglycerides were examined using the automated analyser CMD800i (Wiener LAB) with photometric methodology. Blood was collected without anticoagulant and was immediately sent to the laboratory for analysis. LDL and VLDL cholesterol values were obtained through the Friedwald formula. Inflammatory markers Inflammatory markers (IFN- γ , TNF- α , IL-1 β , IL-6, sVCAM-1, and sICAM-1) were assessed using the cytometric bead array (CBA) method. Results were generated in tabular and graphical format using the BD CBA Software FCAP Array, version 3.01. Ultrasensitive C-reactive protein was measured through the latex immunoblottomymetry technique using the CMD800i automated analyzer (Wiener LAB), where it reacts with the specific antibody to form insoluble immunocomplexes. The turbidity produced by immunocomplexes is proportional to the PCR concentration in the sample. Measurements of the carotid intima-media thickness Measurement was performed using an ultrasound device (General Eletric, model LOGIQe BT12), which features DICOM 3.0 software and Auto IMT, with automatic and well-monitored images. Imaging exams were performed by twomedical vascular surgeons. Measurements were performed on the posterior wall of the studied vessel in a plateau-free area and defined as the distance between two echogenic lines represented by the lumen-intima and media adventitia interface of the arterial wall. The mean automatic measurement of the thickened common carotid artery was defined as either right (RCC) or left (LCC). Presence of plaque was considered when intima-media thickening (IMT) > 1.5 mm was observed. 8-10 Statistical analysis Statistical analyses were performed using the STATA software version 11.0. Level of significance was p < 0.05. Variables were also analyzed stratified by age, with cutoff point at 40 years due to the distribution of N in the cases group. Qualitative variables were expressed through absolute and relative frequencies, and the quantitative ones, through descriptive statistics, such as mean, standard deviation, median, 25 th and 75 th percentile. Continuous variables that presented normal distribution were described through mean and standard deviation; in case of non-normal distribution, they were described through median and interquartile range. Normal distribution was observed for quantitative variables age, total cholesterol, HDL and LDL, while triglycerides, CIMT and inflammatory markers did not present normal distribution. Data normality was evaluated by the Kolmogorov-Smirnov test. Nonparametric Kruskal–Wallis and Mann-Whitney tests were used to compare medians. Student’s independent t -test was used to compare means between groups (HIV + and HIV - ). Correlation analysis was performed using the Spearman coefficient. To select the most significant variables, a stepwise logistic regression was used. Variables moderately associated (p < 0.2) with the dependent variable were included in the model, whereas a threshold of p < 0.05 was adopted for the stepwise elimination of variables considered as risk factors. Ethical considerations The Ethics Committee of the Health Sciences Center of the Federal University of Pernambuco (No. 307087) approved this research and the study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsink. All subjects signed an informed consent form. Results Ninety-nine HIV-positive and 16 non-infected individuals participated in the study. Individuals aged 40 years old or older 91