The Open AIDS

: Background: The association between HIV infection and increased risk of atherosclerotic peripheral artery disease (PAD) has been documented. Still, the relationship between HIV infection and lower extremity medial arterial calcification (MAC) is unknown. Objective: We performed a cross-sectional study to compare the frequency of MAC diagnosed by vascular ultrasound in PAD-asymptomatic people living with HIV (PLWH) with and without chronic kidney disease (CKD) compared to HIV-uninfected participants as a control group. Methods: MAC was defined as smooth, linear, and non-stenotic hyperechogenicity in the arterial wall compared to the surrounding tissues. We studied 191 patients: 50 PLWH (25 with an estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73m 2 and 25 with an eGFR <60 mL/min/1.73m 2 ) and 141 HIV-uninfected patients (68 with eGFR<60 ml/min/1.73m 2 ). Results: MAC was most frequently found in PLWH with CKD (76%). The prevalence of MAC among PLWH was 54.0% (95% confidence interval [CI], 40.4-67.0%), whereas, in HIV-uninfected, it was 34.0% (95% CI, 26.7-42.2%, P=0.013). Age and CKD were consistently associated with MAC in our multivariable models, and there was also a sign that PLWH had higher odds of having MAC. Conclusion: We found a higher prevalence of MAC in PAD-asymptomatic PLWH compared to HIV-uninfected ones and provided evidence that HIV infection could be associated with MAC.


INTRODUCTION
The introduction of antiretroviral therapy has greatly increased the life expectancy of human-immunodeficiencyvirus (HIV) infected persons, also increasing the risk for age-related diseases such as cardiovascular disease (CVD), atherosclerotic peripheral artery disease (PAD) and chronic kidney disease (CKD) [1 -3]. Besides aging, several HIVassociated risk factors for the development of CKD may be present: viral replication, low CD4 T cell count, antiretroviral drugs, and coinfections (hepatitis B and C) [4]. The association between HIV infection and the increased risk of CVD has been widely reported [5]. A large US-based study found that after controlling for traditional factors, HIV infection carried a 19% additional risk for atherosclerotic lower extremity arterial disease [6]. Intimal arterial calcification is associated with atherosclerosis and plaque formation, consequently reducing the arterial lumen and developing atherosclerotic disease [6]. Medial arterial calcification (MAC) differs from atherosclerosis and is frequently present in patients with CKD, diabetes mellitus, and aging [7,8]. Data on the clinical implications of asymptomatic MAC findings are still evolving; however, there is growing evidence that MAC, as a systemic, non-atherosclerotic vascular disorder that increases arterial stiffness, could be an independent risk factor for cardiovascular mortality [9 -11] and atherosclerotic PAD [12] in non-HIVinfected persons. The relationship between MAC and HIV infection has been poorly investigated. The possible role of a class of antiretroviral drugs -protease inhibitors -as a trigger for vascular smooth muscle cell senescence and calcification, a feature of MAC, has been suggested [13]. This study aimed to examine the association between HIV infection, CKD, and MAC defined by vascular ultrasound in people living with HIV (PLWH) compared to HIV-uninfected controls.

Study Design
We performed a cross-sectional study to compare the frequency of MAC diagnosed by vascular ultrasound in PLWH with and without CKD compared to HIV-uninfected participants as a control group. We also examined factors related to MAC and were interested in the relationship between HIV infection and CKD to MAC. The study was supported by the Croatian Science Foundation (project no. IP-2019-04-9702).

Study Population
We investigated 191 patients from two centres: 50 PLWH from the University Hospital for Infectious Diseases (UHID), Zagreb, Croatia, and 141 HIV-uninfected participants presenting to the Nephrology and Dialysis Department, Riuniti Hospital, Anzio (Italy). Consecutive HIV-infected participants were enrolled from the HIV/AIDS Centre at UHID from February 1st, 2018, to September 30th, 2019. The data on HIVuninfected participants were collected retrospectively from medical records among patients who attended Nephrology and Dialysis Department, Riuniti Hospital, Anzio, Vascular laboratory, from September 4 th , 2007, to March 18 th, 2019 Demographic and clinical data were collected from the patient's chart at enrolment.

Inclusion and Exclusion Criteria
Persons aged ≥18 years who provided informed consent were included in the study. HIV-uninfected participants were included if there were younger than 65 years of age and had no known cardiovascular disease, previous kidney transplantation, and previous episodes of acute renal failure. The study population also included persons examined at the Nephrology and Dialysis Department for a regular health check-up and with no known previous kidney disease (N=73).
Twenty-five consecutive PLWH with an estimated glomerular filtration rate (eGFR) ≥ 60 mL/min/1.73m 2 and 25 with an eGFR < 60 mL/min/1.73m 2 were included in the study. All PLWH had confirmed HIV infection by Western-Blot. Exclusion criteria were: pregnancy, breastfeeding, hormonal replacement therapy, corticosteroid and cytotoxic drug use, dementia, or involuntary movement.

Vascular Ultrasound Measures
Vascular ultrasound was performed by one radiology specialist at UHID and one nephrology specialist at Riuniti Hospital Anzio on all participants at a single visit by greyscale B-mode ultrasound for detection of MAC and intimal calcification and duplex color Doppler of the abdominal aorta and inferior extremities' arteries to detect stenosis or occlusion according to the protocol described in the published study by Marinelli and Di Napoli [16]. Vascular ultrasound was performed with the Samsung Medison UGEO H60 machine at UHID and Toshiba Corevision SSA-350A at the Riuniti Hospital Anzio, using a linear probe (7.5 MHz) for analysis of femoral, popliteal, and tibial arteries and a convex probe (3.5 MHz) for the abdominal aorta.
Transcutaneous B-mode ultrasonography was done to detect arterial intimal calcification (AIC), and MAC and duplex colour Doppler sonography was performed to detect stenosis or occlusion. The transducer frequency, focal zone, overall gain, and time-gain compensation were adjusted according to the individual subject's vessel characteristics to optimize the ultrasonography appearance of the arterial wall complex on the anechoic arterial lumen.
The discrete plaque-like calcifications with luminal narrowing were regarded as AIC. MAC was defined as smooth, linear, and non-stenotic hyperechogenicity in the arterial wall compared to the surrounding tissues, disrupting the normal intima-medial wall configuration. MAC was considered as either moderate (spots or isolated segments) or diffuse (more segments) [17,18].
According to guidelines for the assessment of the arterial circulation of the lower limbs, we considered arterial flow as normal in the presence of triphasic (no stenosis) or biphasic waveform (lumen stenosis <50%) and stenotic vessels in the presence of a flow pattern corresponding with lumen narrowing >50%. The vessel was considered occluded if no flow was observed [19].
Atherosclerotic PAD was considered present if any focal or diffuse intimal calcifications were observed in the common femoral, superficial femoral, or tibial arteries and if there were hemodynamically significant stenosis or occlusion.

Statistical Analysis
We described our data on the study population by frequency and percentages or median with the first and third quartiles (Q1-Q3). To compare different characteristics of participants with and without HIV infection and/or peripheral atherosclerosis, we used the Whitney-Mann test for continuous variables and Fisher exact test or χ2 test for categorical variables.
We assessed factors related to MAC by logistic regression with HIV infection and CKD (categorized < 60 and ≥ 60 ml/min/1.73m 2 ) as main predictors. The odds of having MAC was calculated in logistic regression models using the following independent variables: age, HIV status, CKD, BMI, sex, smoking, hypertension, diabetes, and dyslipidemia. We initially included all the variables in the multivariable logistic regression model analysis. However, there were only 16 patients with diabetes; hence the confidence intervals were very wide (1.01-40.96), and there were also concerns about overfitting the model, so diabetes was removed from our main model. By removing diabetes from the model, there were no substantive changes in the coefficients of the main predictors (HIV infection and CKD). The models were also examined for various interactions which were nonsignificant. Five outliers and influential observations were identified by diagnostic plots (e.g., Cook's distance-type index). Since there was a significant change in the coefficient of HIV status after the removal of outliers, we present our multivariable model for the whole population and when outliers are removed. For continuous variables, the Box-Tidwell method checked the linearity assumption, tolerance statistics assessed collinearity, and the Hosmer-Lemeshow test was used to judge the overall fit of the models. Statistical analyses were performed with SAS software, version 9.4. (SAS Institute, Cary, NC, USA). The level of statistical significance was set at P < 0.05.
PLWH had a median of 13 (Q1-Q3: 6-17) years of HIV infection duration with a median current CD4 cell count of 540 per mm 3 (Q1-Q3: 358-720) at the time of vascular ultrasound examination. The most frequent mode of HIV infection transmission was sex between men (58%), followed by heterosexual contact and intravenous drug use (34% and 6%, respectively). At the time of vascular ultrasound examination, 5 patients were not receiving antiretrovirals, and 90% (45/50) of PLWH had plasma HIV-1 RNA < 50 copies/ml.

Lower Extremity Medial Arterial Calcification in HIV-infected Patients
The  Note: a Any stenosis or occlusion plus any intimal vascular calcifications. eGFR, estimated glomerular filtration rate. MAC-medial arterial calcification.

Analysis of Factors Related to MAC
In both univariable and multivariable logistic regression analysis, age and CKD were consistently associated with MAC ( Table 4). PLWH had higher odds of having MAC on univariable and multivariable analyses when outliers were removed ( Table 4). However, when outliers were retained in the multivariable logistic regression model, the association of HIV infection with MAC was not statistically significant (P=0.113) ( Table 4).
Among PLWH, age, hypertension, and eGFR < 60 ml/min/1.73m 2 were strongly associated with MAC. For HIVrelated factors, there was a hint that previous stavudine use and lower CD4 cell counts could be associated with MAC (Table  5).

DISCUSSION
We assessed the presence of lower extremity MAC by comprehensive vascular ultrasound examination in PLWH and HIV-uninfected controls with and without CKD. MAC was most frequently found in PLWH with CKD (76%); it was particularly common in superficial femoral arteries and tibial arteries ( Table 2). MAC was also more frequently found in PLWH (54%) than in HIV-uninfected persons (34%). On multivariable analysis, older age and CKD were consistently related to MAC, whereas the relationship of HIV infection to MAC was less pronounced and was not present in all models ( Table 3). Atherosclerotic PAD was also more frequent in PLWH with MAC compared with HIV-uninfected persons with MAC. a Exact test. b Dolutegravir + lamivudine. Values are frequencies and percentages (in parenthesis) or median and first and third quartiles (in parenthesis). ART, antiretroviral therapy. NRTI, nucleoside analogues reverse transcriptase inhibitors; II, integrase inhibitors; NNRTI, non-nucleoside reverse transcriptase inhibitors; PI, protease inhibitors.
It is problematic to directly compare the prevalence of MAC from published studies because the population under study differs, and the methods used to determine MAC also vary. There are also only a few studies examining MAC in PAD-asymptomatic persons. In a recent study by Konijn et al. [20] using computerized tomography (CT) of the lower extremities, the prevalence of lower extremity calcifications was 55.9% (66/118) in non-atherosclerotic PAD patients. The authors found mainly less severe crural dot-like calcifications in non-PAD patients that seem to fit atherosclerotic intimal calcifications [20]. One small study looking at upper extremity medial calcifications in PAD-asymptomatic CKD patients examined by vascular ultrasound found a prevalence of MAC 26% (11/41) [21]. Older studies in patients with type 2 diabetes found a prevalence of MAC based on radiological findings in 17% [9] and 41.5% [22] patients, respectively. In a more recent study in patients with diabetes, Liu et al. [17] reported that vascular ultrasound is more sensitive in detecting MAC, and MAC was found in 65.8% (173/263) of subjects by ultrasound examination compared to 12.2% when plain radiography was used. CKD is a well-known risk factor for MAC [8]. The prevalence of MAC measured by the ankle-brachial index (ABI ≥1.3) was reported to be 23.7%, 41.7%, 23.1%, and 3.4% in subjects with predialysis, dialysis, and kidney transplant recipients, respectively [23].
The prevalence of MAC has not yet been reported in PLWH. In a large study conducted by Beckman et al. on 28,714 HIV-infected veterans, the authors reported that HIV infection and atherogenic risk factors (older age, hypertension, diabetes, current smoking, and CKD) increased the risk of incident PAD [6]. PAD was defined through administrative claims to individuals who received health coverage from different insurers [24], and the risk of incident PAD was nearly 2-fold higher in persons with sustained CD4 cell counts < 200 cells per mm 3 [6]. In a study of 97 subjects in the Swiss cohort, the prevalence of atherosclerotic PAD based on the ABI index was 20.7% [25]. A recent study from Nigeria reported a higher prevalence rate of low ABI (14.6%) in PLWH compared to the 2% prevalence rate in HIV-uninfected controls [26].
Various diagnostic approaches have been reported in studies on MAC and PAD. The most commonly used noninvasive methods were the ABI and toe-brachial index (TBI) measured by the Doppler technique. An ABI of ≤0.9 and a TBI of ≤0.6 are indicative of atherosclerotic PAD [23,27], and an ABI ≥ of 1.3 indicates MAC [23]. However, a Cochrane review found little evidence about the accuracy of the ABI for diagnosing atherosclerotic PAD in people with exertional leg pain [28]. A recent study suggested that an elevated ABI should not be used to diagnose MAC [29]. Plain foot radiograms have been used [9, 10,22]; however, ultrasound examination is more sensitive in detecting IAC and MAC [17]. Thin-sliced computer tomography (CT) has also been used in differentiating MAC and IAC; however, CT scans are usually done for other purposes [30] and calcifications in arteries are incidental findings. Based on our study's findings, we believe that ultrasound may detect MAC and atherosclerotic PAD in asymptomatic persons and hence could be more frequently used in epidemiological studies.
MAC is considered a vascular disorder distinct from atherosclerosis [8] and is characterized by the accumulation of calcium phosphate with the formation of hydroxyapatite crystals [31]. The pathophysiology is complex and poorly understood, including cellular and molecular mechanisms of inflammation, apoptosis, calcium phosphate metabolism, osteogenic processes, extracellular matrix degradation, matrix vesicle accumulation, and genetic factors [8]. MAC increases vascular stiffening and decreases vessel compliance, affecting localized blood flow [8]. HIV infection is associated with chronic inflammation and monocyte activation [32], which might contribute to MAC. Moreover, perhaps some antiretrovirals might contribute to media calcification [13]. Despite the differences between MAC and atherosclerotic PAD, there are some risk factors common for both (e.g., older age, diabetes, and CKD), which allows clinicians to develop preventive strategies. The correlation of MAC with other viral infections has rarely been studied. Fayed et al. studied the prevalence, severity, and distribution of arterial calcification in hemodialysis patients (stage 5 CKD) with hepatitis B (48 patients) and hepatitis C infection (66 patients). Arterial calcification detected by histopathology of arterial tissue was present in all patients with HCV, of which 31 had medial calcification, and 19 had both intimal and medial calcifications [33].
Our study has limitations primarily because of the relatively small size of the PLWH cohort, which might have reduced the statistical power. In particular, the number of patients in some subgroups was small, resulting in wide confidence intervals (e.g., the presence of diabetes). Because of the small sample size, we did not compare MAC in persons with stage 3 versus stage 4/5 CKD, as reported by Hassan et al. by performing mammography and examining intramammary arteries [34]. Nevertheless, as others did, we found an association between age, CKD, and diabetes with MAC. The association of HIV infection with MAC was present in bivariable analysis and univariable logistic regression, and multivariable logistic regression models when outliers were removed. Because of the small sample size, we could not assess the relationship between the type of antiretroviral drugs used and MAC and whether there might be differences in those on treatment compared to those not treated. PLWH and control group participants were enrolled at different times. However, the vascular ultrasound methodology was the same, so we believe our results are valid. Although widely accepted, we may also argue that the current system for the definition and classification of CKD is not without controversy [35].

CONCLUSION
In conclusion, we found a high prevalence of MAC and atherosclerotic PAD in HIV-infected PAD-asymptomatic persons. PLWH had MAC more frequently than non-HIVinfected patients, and CKD worsened the findings. Age and CKD were consistently associated with lower extremity MAC defined by ultrasound examination. We also provided evidence for an association of HIV infection with lower extremity MAC. Further studies are needed to examine the prevalence of MAC in PLWH and assess asymptomatic MAC's long-term outcome.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
The study was approved by the Ethics Committee of the University Hospital for Infectious Diseases "Dr. Fran Mihaljević," Zagreb, Croatia and Riuniti Hospital, Anzio, Italy (approval number: 01-157-4-2018).

HUMAN AND ANIMAL RIGHTS
No animals were used in this research. All human research procedures followed were in accordance with the ethical standards of the committee responsible for human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2013.

CONSENT FOR PUBLICATION
Written informed consent for participation was obtained from each study participant.

STANDARDS FOR REPORTING
STROBE guidelines and methodology were followed.

AVAILABILITY OF DATA AND MATERIALS
The data supporting this study's findings are available on request from the corresponding author [N.R].

FUNDING
This work was financially supported by the Croatian Science Foundation (Project No. IP-2019-04-9702).