Subclinical Atherosclerosis among HIV-Infected Adults Attending HIV/AIDS Care at Two Large Ambulatory HIV Clinics in Uganda

Background The increased immune activation and inflammation of chronic HIV-infection and the characteristic dyslipidemias associated with HIV infection and antiretroviral therapy (ART) contribute to an increased risk of atherosclerotic vascular disease among HIV-infected adults. There is an emerging need to understand determinants of cardiovascular disease (CVD) among individuals aging with HIV in sub-Saharan Africa. We determined the prevalence of subclinical atherosclerosis [carotid intima media thickness (CIMT) ≥0.78 mm] and its correlation with traditional CVD risk factors among HIV-infected adults. Methods In a cross-sectional study, HIV-infected adults (ART-naïve and ART-treated) were consecutively selected from patients' enrollment registers at two large HIV clinics at Mulago Hospital, Kampala, Uganda. We measured traditional CVD risk factors including age, biophysical profile, fasting blood sugar and serum lipid profile as well as biomarkers of inflammation. High resolution ultrasound was used to measure common carotid CIMT. Results Of 245 patients, Median age [Interquartile range (IQR)] 37 years (31–43), 168 (69%) were females; and 100 (41%) were ART-treated for at least 7 years. Overall, 34/186 (18%) had subclinical atherosclerosis; of whom 15/108 (14%) were ART-naïve whereas 19/78 (24%) were ART-treated. Independent predictors of subclinical atherosclerosis included age [odds ratio (OR) 1.83 per 5-year increase in age; 95% confidence interval (CI) 1.24–2.69; p = 0.002], body mass index (BMI); OR 1.15; CI 1.01–1.31; p = 0.041 and high low density lipoprotein (LDL) [OR 2.99; CI 1.02–8.78, p = 0.046]. High sensitivity C-reactive protein (hsCRP) was positively correlated with traditional cardio-metabolic risk factors including waist circumference (r = 0.127, p = 0.05), triglycerides (r = 0.19, p = 0.003) and Total Cholesterol: High Density Lipoprotein ratio (TC:LDL) (r = 0.225, p<0.001). Conclusion The prevalence of subclinical atherosclerosis was 18% among HIV-infected adults in Uganda. Traditional CVD risk factors were associated with subclinical atherosclerosis. We recommend routine assessment of traditional CVD risk factors within HIV care and treatment programs in sub-Saharan Africa.


Introduction
The chronic inflammation and immune activation associated with HIV-infection and the dyslipidemia associated with Antiretroviral Therapy (ART) contribute to increased risk of atherosclerotic vascular disease among HIV-infected adults relative to the general population [1,2]. HIV-infected adults appear to have a significantly greater risk of myocardial infarction and coronary artery disease relative to age-matched HIV-negative individuals [3,4,5,6]. With the increasing life expectancy of HIV-infected adults receiving ART, cardiovascular disease has emerged as an important concern among individuals aging with HIV/AIDS [7,8,9]. With the scale up of life-saving antiretroviral therapy in sub-Saharan Africa (SSA), there is a growing population of individuals aging with HIV [10,11], thus the need to understand predictors of cardiovascular disease (CVD) to inform management of CVD within HIV care and treatment programs.
HIV is associated with dyslipidemia, namely, hypocholesterolemia, hypertriglyceridemia, and low levels of both low density lipoproteins (LDL) and high density lipoproteins (HDL) [12,13]. Whereas elevated serum lipids increase CVD risk, a low HDL contributes to the elevated risk of coronary artery disease (CAD) in HIV patients [14,15]. In addition, ART is associated with dyslipidemia, which tends to be worse with many protease inhibitor (PI) -containing regimens [16]. Similarly, HIV-associated immune activation and inflammation are suggested to contribute to accelerated atherosclerosis and subsequently increased CVD risk among adults living with HIV/AIDS [17,18,19]. There are limited data on CVD and its risk factors within HIV treatment programs in SSA. There is thus a need to understand the relationship of traditional versus non-traditional risk factors with CVD among aging HIV-infected adults in this context.
Given the morbidity and mortality associated with overt CVD [3,4,5,6], early identification and management of subclinical disease is desirable before complications of overt CVD develop. Measurements of carotid intima media thickness (CIMT) have previously been correlated with the extent of coronary atherosclerosis [20,21,22], and increases in CIMT are predictive of future CVD events [22,23]. In this study, we measured CIMT to determine the prevalence of subclinical atherosclerosis and associated factors to inform prevention, early detection and prompt management of CVD within HIV treatment programs in SSA. We also examined the association of inflammatory markers such as high sensitivity CRP (hsCRP) with CIMT and traditional CVD risk factors in order to contribute to the development of clinical and diagnostic algorithms for CVD in settings where CIMT measurements might not be feasible in resource-limited settings.

Study setting
This study was conducted among HIV-infected adults enrolled at two large ambulatory HIV clinics at Mulago National referral hospital. The Makerere Joint AIDS Program immune suppression (MJAP-ISS) clinic enrolls ART-naïve HIV-infected adults that are followed up and initiated on ART according to the 2012 national guidelines of ART initiation at CD4 counts,350 cells/UL. In addition, the Infectious Diseases Institute (IDI), has an HIV treatment cohort that has received ART for seven years, having initiated ART at CD4 counts #200  [24,25]. Protease inhibitor-containing regimens were reserved for second-line therapy if patients failed their first-line regimen.

Study design and participants
In a cross-sectional study, HIV-infected adults $18 years were consecutively selected from the patient registers of ART-naive adults enrolled at MJAP-ISS clinic between February to October 2012, and the ART-treated adults in the HIV treatment cohort with HIV viral loads ,400 copies/ml for at least 6 months prior to the study. Patients with confirmed CVD (stroke, myocardial infarction, and/or peripheral vascular disease), malignancy and active infection (as per the attending physicians at MJAP-ISS/IDI and medical records), were excluded. In addition, we excluded patients receiving glucocorticoids, growth hormone or other anabolic agents within the past 6 months. All eligible patients provided written informed consent to participate in the study. Scientific and ethical approvals were obtained from Makerere University School of Medicine Research and Ethics Committee, and the study was conducted according to the Declaration of Helsinki on research involving human subjects.

Clinical assessment
Upon written informed consent and using a pre-coded and pretested questionnaire as a study instrument, all participants received a detailed clinical assessment to evaluate the relevant history of CVD risk factors including: lifestyle characteristics (exercise, diet, history of smoking and alcohol use), family history of CVD (hypertension, stroke, myocardial infarction and sudden death) as well as any concurrent medication that might affect CVD risk [anti-hypertensive drugs, non-steroidal anti-inflammatory drugs (NSAIDs), and lipid lowering drugs]. In addition, biophysical profile was measured by obtaining the weight in kilograms (Kg) and height in meters (m) to calculate body mass index (BMI), Kg/ m 2 , waist circumference, hip circumference, waist-hip ratio and blood pressure (BP) using an automated blood pressure cuff (OMRON). We also evaluated HIV-specific risk factors such as ART status and stage of HIV disease using WHO clinical stage and CD4 count.

Laboratory tests
A total of 10 mls of blood samples were obtained using aseptic techniques; 2 mls in EDTA tubes for hematology tests [Complete blood count and Erythrocyte Sedimentation Rate at (ESR)] and 3 mls in a general vacutainer for serum lipid profiles at the Uganda Heart Institute laboratory. Fasting blood sugar was measured using a glucometer (CONTOUR R TS). Blood for inflammatory markers was centrifuged for 10 minutes at 1,0006g and the serum was stored at ,270uC at the Makerere University College of Health Sciences immunology laboratory until batch analysis was done. High sensitivity C-reactive protein (hsCRP) was measured by standard ELISA technique using commercial cytokine ELISA kits from ABNOVA cat. KA0238.

CIMT Measurement
This was done by a senior radiologist with experience in vascular imaging at Mulago National Referral Hospital, using carotid Doppler imaging. A linear array transducer with a frequency of 7-14 MHz (PHILIPS HD7) was used to obtain images of the distal 1 cm of the common carotid artery and the carotid bulb. The right and left carotid arteries were studied with the head in the midline position and tilted upward. The probe was adjusted to obtain the near and far walls in a parallel orientation and then positioned to obtain the maximal luminal diameter in the longitudinal plane. Images were recorded digitally for subsequent offline analysis using multiple manual caliper measurements. Three images of the common carotid artery were obtained following the American Society of Echo guidelines at 3 angles (anterior, lateral and posterior) bilaterally along with bilateral images of the carotid bulb at the optimum angle of insonation [26]. A mean value for each segment was obtained and the average of all measurements reported as overall CIMT for the common carotid artery. Subclinical atherosclerosis was defined using a more aggressive cut-off value of CIMT ($0.78 mm), based on the prior observation that on average, a healthy adult reaches a CIMT of 0.78 mm at the age of 76 years [27]. The oldest participant in our study was 74 years old and .95% of participants were below 60 years old, thus justifying this lower cutoff as clearly abnormal.

Data analysis
Data was entered in EPIDATA version 3.1 and exported to STATA 12 for analysis. Subclinical atherosclerosis, defined as CIMT $0.78 mm, was the primary outcome. Predictors of subclinical atherosclerosis were determined using bivariate and multivariate logistic regression. Independent predictors of subclinical cardiovascular disease such as; biophysical measurements, traditional CVD risk factors and the non-traditional CVD risk factors in HIV-infected individuals such as stage of HIV disease and ART status were analyzed. Factors with a p-value of ,0.2 at bivariate analysis were included in a stepwise backward model. Subclinical atherosclerosis was also compared across Framingham risk score categories. We also determined the correlation of hsCRP (as a continuous variable) with CIMT and traditional CVD risk factors. A p-value ,0.05 was considered statistically significant.

Characteristics of study patients
Between February and October 2012, 245 HIV-infected adults were evaluated for CVD risk. One-hundred and forty-five (59%) were ART-naïve patients and 100 (41%) had received suppressive ART for median of seven years ( Figure 1). Of those on ART, 14% were on a second-line PI containing regimen. The median [interquartile range, (IQR)] age was 37 (31-43) years, 168 (69%) were female and 124 (51%) had no higher than primary school education. A majority of the patients 132 (54%) were in WHO stage III and IV. Among the ART-treated adults, 86% were receiving NNRTI-based first-line regimen and only 14 patients were receiving a second-line PI-containing regimen (Table 1).

Subclinical atherosclerosis among HIV-infected adults
Carotid ultrasound imaging was performed in 186 of the study subjects. Fifty-nine subjects were not imaged due to logistical difficulties in arranging the ultrasound examination. Baseline characteristics of these patients were not different from those who had carotid imaging. Subclinical atherosclerosis was observed in (34/186) 18% of patients, 19 (56%) of whom were receiving ART. Only 2 of the 14 receiving a PI-containing regimen as second-line therapy had CIMT of $0.78 mm. Factors independently associated with subclinical atherosclerosis ( ESR was .20 mm/hr among 71 (32%) of HIV infected patients, of whom 13 (18%) were on ART. HsCRP was $3 mg/L in 130 (54%) HIV patients, of whom 56 (43%) were on ART. There was no correlation between CIMT and hsCRP (Spearman's correlation (r) = 20.051 p = 0.49). Hs-CRP was, however, positively correlated with several traditional CVD risk factors (Table 4):  waist circumference (r = 0.127, p = 0.05), triglycerides (r = 0.19; p = 0.003), and TC/HDL ratio (r = 0.225; p,0.001).

Discussion
The prevalence of study-defined subclinical atherosclerosis was 18% among HIV-patients, 56% of whom were receiving ART. This is a lower prevalence compared to what has been found in studies done in developed countries [28,29,30]. In a cross sectional study of 187 HIV-infected patients attending clinics of Hospital Universitari de Santo Joan, Spain, [28]; a high prevalence of subclinical atherosclerosis at 65% was reported [28]. Despite a comparable cut-off of 0.80 mm, patients in the latter study were mainly males and more than 80% were cigarette smokers. Among ART-naïve HIV-infected adults attending hospitals in the Italian coordination group for the study of allergies and HIV infection, De Socio et al [29], reported subclinical atherosclerosis at 41.7% using CIMT cut-off of 0.9 mm. About half (51%) of the subjects in De Socio et al [29] study, had a history of smoking; median age was 40 years and majority of participants (79.2%) were males. It is important to note that our CIMT cut-off of 0.78 mm would have provided even higher rate of subclinical atherosclerosis in these western studies. Our study used this cut-off basing on previous findings from an observational study in the Netherlands, which demonstrated that a healthy adult reaches a CIMT of 0.78 mm at the age of 76 years [27]. In a European cohort Maggi et al [30] found 14.3% of ART naïve and 32.2% of HAART-treated patients to have a CIMT .1.00 mm. Despite a comparable age to our study patients, majority of the patients in the latter study were males (75%) and over two thirds had a history of cigarette smoking, in addition to higher rates of dyslipidemia. Even with our stringent cut-off of 0.78 mm for CIMT, our patients had lower rates of subclinical atherosclerosis despite seven year of ART. We attribute these differences to the generally low prevalence of the traditional CVD risk factors (i.e. smoking) in Uganda relative to the western cohorts [28,29,30,31]. Although quantifiably smaller than other HIV-infected cohorts, our results reveal an emerging risk of future CVD among adults aging with HIV/AIDS in Uganda. This calls for longitudinal studies to understand the modifiable CVD risk factors in this population. In addition, there is need to further understand the mechanisms by which HIV directly or indirectly through inflammation, may accelerate the incidence of non-AIDS defining illnesses to inform the preventive and therapeutic interventions for people living with HIV/AIDS.
Our study shows that traditional CVD risk factors such as advancing age, obesity, and high LDL-Cholesterol were associated with subclinical atherosclerosis among patients with HIV infection in Uganda. Several studies in HIV have previously shown the association of traditional CVD risk factors and atherosclerosis as measured by CIMT [18,29,31,32,33,34,35]. Currier et al [35] found that Carotid IMT was significantly greater with older age, male gender, greater BMI, greater waist-to-hip ratio, increased systolic blood pressure, total cholesterol, glucose, homocysteine, regular smoking, alcohol consumption, lipodystrophy, and antiretroviral therapy. After multivariable adjustment, however, only conventional CVD risk factors were independently associated with increased CIMT in HIV-infected patients. In our study, hsCRP was not associated with subclinical atherosclerosis, despite the fact that levels were elevated ($3 mg/L) in over half of our patients. Our results are comparable with previous reports that hsCRP did not correlate with CIMT, particularly at the Common Carotid Artery (CCA) segment, among ART-treated HIV-infected adults [19,36]. We measured CCA IMT because it is easier to measure and results are more reproducible, relative to CIMT as measured from the internal carotid artery (ICA) [26,37,38]. The American Society of Echocardiography and Society of Vascular Medicine and Biology, in a report on the use of vascular ultrasound in cardiovascular risk stratification, recommended the use of the CCA segment for improved reproducibility since there was no compelling evidence to suggest that combined measurements or measurement of a specific segment is clearly superior [37].
We demonstrated a positive correlation between hsCRP and traditional CVD risk factors including waist circumference, triglycerides, total cholesterol, however there was no correlation between hsCRP and CIMT. Similarly, there are previous studies in some Western cohorts that demonstrated a positive correlation between hsCRP and traditional CVD risk factors such as LDLcholesterol, HDL-cholesterol, BMI, waist-to-hip ratio (WHR) and smoking [39,40,41]. Guimaraes et al found that hsCRP levels correlated positively with waist measurement (p = 0.004), WHR (p,0.001), systolic (p = 0.05) and diastolic (p = 0.03) blood pressure, intra-abdominal fat thickness (p = 0.02), triglycerides (p = 0.001), total cholesterol (p = 0.01), and fasting glucose (p = 0.01) [41]. Similarly, in a cross sectional study of 245 HIV-  Given the cross-sectional nature of our study we could not provide conclusive evidence about the relationship between hsCRP and CIMT since we did not actively exclude the multifactorial causes of raised CRP within our study population such as malaria, helminthes and subclinical hepatocyte dysfunction [42]; among others. In addition, we did not have an HIV-negative control group to provide background data on hsCRP and CIMT in this community. Also, we did not measure other markers of systemic inflammation that have been shown to correlate with cardiovascular disease in HIV, such as IL-6.

Implication of the study
Subclinical atherosclerosis was found in 18% of HIV-infected adults attending HIV care programs at Mulago hospital. Although a lower prevalence than prior studies in developed countries, this subclinical atherosclerosis is likely to lead to clinically important cardiovascular disease for these participants over time. With the global scale up of ART, more HIV-infected adults are living longer, have higher BMI, and experience more dyslipidemias. In this study, we found that these traditional cardiovascular risk factors were associated with subclinical atherosclerosis among HIV-infected adults. This implies that HIV care facilities need to be ready to screen for CVD and identify at risk patients for early intervention. This can be achieved through increasing awareness among health care providers about the need to evaluate CVD among their clients as well as provision of facilities to measure the known traditional risk factors within the HIV care programs. For example; there is a need to train all health workers in CVD assessment, having the necessary equipment to use in CVD risk assessment and detection, for example, sphygmomanometers, scales, and tape measures for measuring blood pressure, BMI, and waist circumference. There is a need to have access to laboratories that can monitor lipid profile for HIV-infected patients. In addition, the HIV care patient charts should have provisions to record these risk factors for the purpose of CVD risk modification. Similarly, ongoing counseling and education on risk reduction is required to prevent CVD among people living HIV.

Study limitations
We recognize that the lack of an HIV-negative control population is a limitation of our study; however, this study was not designed to determine the effect of HIV on CVD risk but rather to explore the magnitude of CVD risk among adults in HIV treatment programs in the SSA context, including the risk associated with ART use. Secondly, we used hs-CRP as a single biomarker of systemic inflammation, although other markers such as IL-6 have been correlated with CVD in HIV subjects. In addition, systemic immune activation, specifically monocyte activation was not measured. Traditional CVD risk factors were strongly correlated, however, with generalized inflammation as measured by hsCRP in our study, suggesting that initial prevention strategies should target traditional CVD risk factors among adults living with HIV in Uganda. Future longitudinal cohort studies among HIV and non-HIV individuals in SSA are needed to further understand relationships between traditional and non-traditional CVD risk factors in our setting.

Conclusion
Overall prevalence of subclinical atherosclerosis among HIVinfected adults was 18% among HIV-infected patients. Traditional CVD risk factors were associated with subclinical atherosclerosis during HIV disease. HsCRP had a positive correlation with traditional CVD risk factors. We recommend routine assessment of CVD using traditional risk factors within HIV care and treatment programs in sub-Saharan Africa.