Standardized Comparison of Cardiovascular Risk Factors Prevalence in Spanish Women and Men Living with HIV and in the General Population

People living with HIV (PLWH) have an increased risk of cardiovascular (CV) disease, likely due to a higher prevalence of CV risk factors. We compared the age-standardized prevalence and management of CV risk factors in PLWH to that of the general population in Spain. Blood pressure, lipid, glucose, and anthropometric profiles were cross-sectionally compared along with the treatment of hypertension, dyslipidemia, and diabetes in a general population cohort and a PLWH cohort. Prevalence rates were standardized by the direct method by 10-year age groups in European standard populations and stratified by gender. We included 47,593 individuals aged 35 to 74 years, 28,360 from the general population cohort and 19,233 from the PLWH cohort. Compared to the general population, PLWH had a higher concentration of triglycerides (>35 mg/dL in women and >26 mg/dL in men) and a higher prevalence of smoking (>23% and >17%) and diabetes (>9.9% and >8.5%). The prevalence of treated diabetes, hypertension, and dyslipidemia were up to three-fold lower in both women and men living with HIV. There was a significant difference in PLWH compared to the general population in the lipid, glucose, and anthropometric profile. In addition, PLWH were less often treated for diagnosed diabetes, hypertension, and dyslipidemia.


Introduction
The introduction of combined antiretroviral therapy (cART) in the 1990s significantly reduced the morbidity and increased the survival of people living with HIV (PLWH). cART has become the standard of care for PLWH, leading to lifelong suppression of viral replication. cART has steadily improved undetectable viral load at 48 weeks from 48% in 1995 to 78% in 2010 and >90% in the most recent analysis [1]. This success has gained several decades of life expectancy for these patients. As PLWH live longer with their disease, they face age-related chronic comorbidities [2].
PLWH tend to smoke and develop cardiovascular (CV) diseases, hypertension, diabetes, and chronic kidney disease more often than the general population [3][4][5]. PLWH have been reported to have almost 60% greater prevalence of hypertension, coronary artery disease, peripheral arterial disease, or chronic kidney disease as compared to the general population [6]. This increase is probably due to a combined effect: (1) higher prevalence of CV risk factors; (2) chronic inflammation and immune activation from the host response; (3) adverse events of cART; and (4) metabolic effects induced by HIV infection [7][8][9]. These factors may also contribute to the management of PLWH and their comorbidities. Half of PLWH have one or more comorbidities and receive at least one medication concomitant to cART [10].
While comparison of CV risk factors' prevalence in PLWH and the general population has been reported in some studies, age-standardized risk factor prevalence has never been used to control for differences in age distribution. For instance, PLWH tend to be younger than the general population in these studies. Furthermore, few studies have focused on the sex-related distribution of CV risk factors in PLWH, and available results are biased toward men living with HIV (MLWH). Women living with HIV (WLWH) continue to be under-represented in clinical trials and epidemiological studies, despite representing over half of PLWH worldwide, and their socio-economic and clinical conditions can be different than what is observed in men [11,12]. In Spain, the prevalence of HIV infection is lower among women (0.1%) than in men (0.65%), a distribution that is characteristic of high-income countries and prompts differential analysis by sex [13].
The objective of the present study was to compare the age-standardized prevalence and management of CV risk factors in Spanish WLWH and MLWH with that of the Spanish general population.

Study Design and Population
We used a cross-sectional design and two large Spanish cohorts to compare the prevalence of CV risk factors in the general population with that in PLWH aged 35 to 74 years. We excluded individuals outside the defined age range and those records with missing data. For each individual included, we analyzed the most recent measurements available.
Data on the general population were obtained from the epidemiology wing of a study of dyslipidemia, atherosclerotic risk, increased high-sensitivity C-reactive protein (hsCRP), and inflammatory and oxidative status (DARIOS study). DARIOS is a pooled cohort of individual data from 11 population cohorts recruited in 10 Spanish autonomous communities in 2000-2010, which included 28,887 participants aged 35 to 74 years at baseline [14,15].
Data on PLWH were obtained from the VACH cohort, a Spanish registry of PLWH consecutively recruited since 1997 at 23 hospital-based HIV clinics nationwide. The VACH cohort includes 29,217 HIV patients who were 16 years or older at hospital admission or at recruitment to the cohort [16,17].

Data Collection and Measurements
Measurements and questionnaires from the DARIOS and VACH cohorts were obtained during an examination and interview at study inclusion and at their first clinic visit, respectively. Standardized World Health Organization questionnaires [18] were used to collect information on sociodemographic variables, smoking status, and previous history of high blood pressure, dyslipidemia, and diabetes. Physical examinations and blood tests were performed to obtain data on anthropometrics, blood pressure, lipid profile, and glycaemia.
Weight, height, and waist circumference were measured by professional healthcare workers. Body mass index (BMI) was calculated as weight divided by squared height (kg/m 2 ). Systolic and diastolic blood pressures were measured in seated individuals with an automatic sphygmomanometer after a 5 min rest. Two measures were taken, and the lowest was recorded. Hypertension was defined as previous diagnosis or treatment and/or systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg.
In DARIOS, blood samples were taken after 10-14 h fasting and stored at −80 • C. Total cholesterol, high-density lipoprotein cholesterol (HDL-c), triglycerides, and glucose were determined using enzymatic methods. Low-density lipoprotein cholesterol (LDL-c) was calculated with the Friedewald equation when triglycerides were <300 mg/dL. Diabetes was defined as (i) previous diagnosis or treatment or (ii) glucose >125 mg/dL. Quotients of total cholesterol/HDL-c and triglycerides/HDL-c were also calculated. VACH measurements were taken from hospital laboratories' records.
Metabolic syndrome was defined as the presence of three or more of the following five conditions: BMI > 30, waist circumference ≥ 102 cm for men or ≥88 cm for women, triglycerides ≥ 150 mg/dL, HDL-c < 40 mg/dL for men or <50 mg/dL for women, blood pressure ≥ 130/≥85 mmHg, or fasting glucose ≥ 100 mg/dL.

Statistical Analysis
The effect of age differences between the VACH and DARIOS cohorts was controlled by direct standardization of risk factor prevalence using the 1976 European Standard Population (ESP) [19]. The 1976 ESP was chosen over the 2013 ESP to allow direct comparison with previous publications (we used the weights 14, 14, 11, and 7 for the age groups of 35-44, 45-54, 55-64, and 65-74 years, respectively) [20].
Categorical variables including age-standardized prevalences were described as frequencies and compared using the chi-squared test. Continuous variables were described with the mean and 95% confidence interval (CI) and compared using the Student t-test. Triglycerides were log-transformed to normalize their distribution.
All statistical analysis was performed with the R software version 3.6.2 [21].

Ethics
All participants were duly informed and signed their consent to participate. The present study was approved by the Parc de Salut Mar Drug Research Ethics Committee (Authorization CEIm PSMAR 2018/8347-I).

Results
In this study, we included 47,593 individuals aged 35-74 years, 28,360 from the general population (DARIOS cohort) and 19,233 with HIV (VACH cohort). We excluded 527 individuals from the DARIOS and 9984 from the VACH cohort due to missing data or age out of range ( Figure S1).
Differences in CV risk factor prevalence between women from the general population and WLWH are presented in Table 1. WLWH had significantly higher systolic blood pressure, triglyceride, and glucose concentration levels. WLWH had also higher agestandardized triglycerides/HDL-c quotient and higher prevalence of diabetes, metabolic syndrome, and smoking. WLWH had lower age-standardized concentration of HDL-c, LDL-c, and total cholesterol and lower prevalence of obesity than women from the general population. On the one hand, the analysis of crude prevalences showed similar results except for the following parameters. Systolic blood pressure, glucose concentration levels, and prevalence of hypertension and metabolic syndrome were lower in WLWH. On the other hand, concentration of triglycerides, HDL-c, LDL-c, and total cholesterol, triglycerides/HDL-c rate, and prevalences of diabetes, obesity, and smoking were similar between WLWH and the general population. Table 2 presents CV risk factor prevalence in men. Compared to men from the general population, MLWH had a higher concentration of triglycerides, quotient of triglycerides/ HDL-c, and prevalence of diabetes, metabolic syndrome, and smoking, but lower agestandardized blood pressure, HDL-c, LDL-c, and total cholesterol concentrations, quotient of total cholesterol/HDL-c, and prevalence of obesity.
When comparing crude and standardized prevalences in MLWH, we found the following differences. Crude estimates for MLWH yielded lower glucose concentration, hypertension, and metabolic syndrome prevalence, and similar concentration of blood pressure, triglycerides, HDL-c, LDL-c, and total cholesterol, quotients of total cholesterol/HDL-c and triglycerides/HDL-c, and prevalence of diabetes, obesity, and smoking than in men from the general population. The percentage of individuals with treated hypertension, diabetes, and dyslipidemia was lower in PLWH than in the general population in both women and men (Tables 1 and 2). In WLWH, we estimated a 2.9-fold decrease in the prevalence of treated hypertension, a 2.8-fold decrease in treated diabetes, and a 1.6-fold decrease in treated dyslipidemia compared to their general population counterparts. Similar differences were observed in MLWH (2.2-, 3.0-, and 1.5-fold decreases, respectively) compared to men in the general population when prevalences were age-standardized.

Discussion
This study compared CV risk factor prevalence between Spanish PLWH (VACH cohort) and the Spanish general population (DARIOS cohort), stratified by sex and, importantly, standardized for age. The most prevalent CV risk factors observed in PLWH were a higher triglycerides concentration (>35 mg/dL higher in women and >26 mg/dL higher in men) and a higher prevalence of smoking (>23% and >17%, respectively) and diabetes (>9.9% and >8.5%) as compared to the Spanish general population.
The higher concentration of triglycerides observed in the VACH cohort agrees with results of the United States study by Önen et al. [22], the Copenhagen study by Gelpi et al. [23], and the French study by Savès et al. [24]. The French cohort was receiving cART with protease inhibitors (PI), which could have increased triglycerides concentration. Most PLWH from the VACH cohort also received cART containing a PI or a nucleoside reverse transcriptase inhibitor (NRTI), which is associated with the development of dyslipidemia and dysregulation in glucose homeostasis [25][26][27][28][29].
The difference in metabolic syndrome prevalence between PLWH and the general population is likely due to the higher concentration of triglycerides and prevalence of diabetes and the lower HDL-c concentration observed in PLWH. The increased diabetes prevalence in PLWH observed in our study could possibly be associated with causes similar to the increased triglycerides concentration. Noubissi et al. [5] described an association between cART and diabetes, although Önen et al. [22] and Savès et al. [24] found no difference in diabetes prevalence between PLWH and the general population.
The PLWH cohort in our study was less often treated for hypertension, diabetes, and dyslipidemia than the Spanish general population. Prior studies in other countries have yielded contradictory results on this topic. Gelpi et al. [23] described a similar proportion of treated diabetes and hypertension between PLWH and the general population; on the other hand, the PLWH group had more dyslipidemia-treated patients than in the general population of Copenhagen. In Australia, Dharan et al. [30] also found that PLWH purchased more dyslipidemia medication but observed that fewer PLWH received treatment for diabetes compared with the non-infected population; the authors attributed these findings to differences in population characteristics between the two groups. In the United States, Önen et al. [22] estimated greater prescription of anti-hypertensive and lipid-lowering medications for PLWH than their general population counterparts. Among a number of possible reasons for undertreatment of CV risk factors in PLWH in Spain is the lower efficacy thresholds in PLWH of drugs used for treating metabolic disorders, particularly dyslipidemia [31]. Most of these drugs also have potential interactions with some antiretroviral drugs, particularly booted PI and non-nucleoside reverse transcriptase inhibitors (NNRTI); this may result in the use of suboptimal doses or slow titration that ultimately have an impact on treatment efficacy. All of these considerations support the need for a more thorough monitoring of CV risk factors in PLWH.
Lower HDL-c levels and lower prevalence of hypertension in PLWH were described in the French study by Savès et al. [24]; in contrast, we did not observe a lower prevalence of hypertension in our study. Up to three times less PLWH diagnosed with hypertension and/or diabetes were receiving treatment to manage these chronic conditions in our cohort compared to the general population. The treatment of dyslipidemia was also comparatively lower in PLWH, which can be explained by the lower concentrations of total cholesterol and LDL-c concentrations in PLWH than in the Spanish general population.
Many of these previous studies did not stratify PLWH characteristics by sex. Our study is one of the few to estimate CV risk factor prevalence separately in women and in men. Unlike men, WLWH had a higher glucose concentration and systolic blood pressure compared to women from the general population. These differences support the need for stratification by sex in future studies in order to better describe each gender's characteristics and health necessities.
Our findings suggest that previous non-standardized CV risk-factor prevalence comparisons, such as those discussed above, of PLWH with the general population may have produced biased results due to the large difference of age in both populations. Our analysis pointed to this observation as we observed differences between crude and standardized estimates in both WLWH and MLWH. Standardization controls for the effect of age in different populations and allows for fair comparisons [24,32,33]. We used the direct method with European standard population weights. The results obtained after standardization and stratification by sex yield comparable values for both continuous and categorical factors.

Limitations
This study has three main limitations. First, its observational and cross-sectional design makes it strictly descriptive. Second, the treatment of dyslipidemia has been computed on the overall populations because the criteria for diagnosis did not match in the DARIOS and VACH cohorts. Third, in the VACH cohort, we have excluded a large number of patients because they were out of the age range or because they had missing values in CV risk factors. This fact can only influence our results against our hypothesis since patients with worse management were excluded.

Conclusions
After appropriate standardization to correct for the effect of age, triglyceride concentration and the prevalence of smoking and diabetes were higher in both sexes of PLWH compared to the general population. In addition, glucose concentration and systolic blood pressure were higher in WLWH than in women from the general population. Spanish PLWH were undertreated for diabetes, hypertension, and dyslipidemia compared to the general population. More thorough intervention to address these modifiable risk factors is needed and should become a part of routine HIV care.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.