Prediction of Cardiovascular Risk Comparing Algorithmic Models and Cardiac Risk Factors in Plwha on Art With Overwiegth/ Obesity

Objective: The aim of this study was in seropositive subjects from Northern Mexico undergoing cART. Methods: This study included 186 PLWH under cART. The variables analyzed were CD4+ count, viral load, lipid prole, glucose, insulin resistance, anthropometric measures, family history of hypertension and cardiovascular disease, years of treatment and cART scheme. In this study, we used two well-established algorithmic models D:A:D (5-year period) and Framingham (10-year period) for assessing cardiovascular risk. Results: In our study, 51.3% of the PLWH had arterial hypertension; most of the subjects were diagnosed with overweight, hypertriglyceridemia and metabolic syndrome, which are factors that increase the risk of cardiovascular disease. After the assessment with the Framingham model, risk was considered as low, while for the D:A:D model was moderate for this population. Conclusions: PLWH receiving cART present factors that may increase the risk of early heart disease including hypercholesterolemia, hypertriglyceridemia, smoking and age. Although risk was low/moderate after the assessment, it is important to consider other factors such as age of the subjects, overweight/obesity, smoking or coinfections, in addition to years of exposure to cART, which could increase the rate of heart disease.


Introduction
The successful use of combination antiretroviral therapy (cART) has resulted in an increase in life expectancy for people living with human immunode ciency virus (HIV) infection (1,2). Over the last years, there has been a shift in HIV-related health complications, changing from opportunistic infections related to the was evaluated based on the WHO criteria (triglycerides ≥150 mg/dL, blood pressure 130/85 mmHg and insulin resistance) (30). Blood pressure was measured using an electronic device (Equaline BW-323A, Taiwan) and normality was de ned with values of 110-120 mmHg of systolic pressure and 60-80 mmHg of diastolic pressure.

Biochemical measures
Fasting peripheral blood samples (4 mL) were taken by venipuncture in a tube with EDTA and (5 mL) in another tube without anticoagulant. The lipid pro le characterization was performed by conventional clinical laboratory methods. Biochemical parameters were measured by the Vitros® 250 Chemistry System.

Cardiovascular Risk Assessment
The cardiovascular risk of each participant was calculated using two logarithmic models: D:A:D (5-year risk) and Framingham (10-year risk). Both models estimate the risk of cardiovascular disease by combining information on age, sex, systolic blood pressure, total cholesterol, HDL, diabetes and smoking. While the equation of D:A:D includes the use of antiretroviral therapy (31,32), Framingham's equation estimates the risk including antihypertensive therapy, but not antiretroviral therapy (32,33). In addition, for the D:A:D model we used two versions: D:A:D reduced (including age, gender, systolic blood pressure, smoking, family history of CVD, diabetes, total cholesterol, high-density lipoprotein, CD4 cell count) and D:A:D full model (including age, gender, systolic blood pressure, smoking, family history of CVD, diabetes, total cholesterol, high-density lipoprotein, CD4 cell count, cumulative exposure to protease and reverse transcriptase nucleoside-inhibitors and current use of abacavir). Finally, the risk for participants according to D:A:D was considered as low (<1%), moderate (1-5%), high (5-10%), and very high (>10%); the risk according to Framingham was low (<10%), moderate (10-20%), high (21-30%) and very high (> 30%) (31).

Statistical analysis
Data was analyzed using the statistical software package SPSS v21.0 (IBM Corp, Chicago IL, USA). Sociodemographic and clinical quantitative variables were analyzed using central tendency and dispersion measurements (i.e., means and standard deviations). In addition, Mann-Whitney U test was used to compare non-normally distributed quantitative variables. Odds ratios and multinomial logistic regression were used to measure association strength. Finally, p-values <0.05 were considered as statistically signi cant.

Sociodemographic and clinical characteristics
The sample consisted of 186 participants, 145 males (77.4%), and 41 (22.6%) females, with a mean age of 39.1 years (Table 1). In addition, 45.6% of the participants were homosexual, 44.4% heterosexual, and 10% bisexual. At the time of the study, all patients were receiving cART with a mean treatment of 7.18 (±6.15) years and 9.73 (±6.46) years with the infection. Seventy out of the 186 participants (37.6%) were reported as smokers and 94 individuals (50.5%) as former smokers. According to metabolic parameters, 51.3% of the total population sample had a history of hypertension, whereas 5.4% were individuals with T2D. The mean count of CD4+ at the beginning of the treatment was 520.17 (±438.62 cell/mm 3) . Regarding cART treatment, ~65% of the participants were under a treatment scheme of two nucleoside inhibitors plus a non-nucleoside reverse transcriptase inhibitor (Table S1).
Physical examination and biological pro le of the participants The mean of BMI was 25.44±5.3 kg/m 2 . Approximately, 4.3% of the participants were classi ed as underweight, 49.5% with normal weight, 28.5% with overweight, and 17.7% with obesity. One hundred patients (53.8%) had hypertriglyceridemia, 44 (23.7%) hypercholesterolemia, 72 (38.7%) presented HDL levels <40 mg/dL, 69 (37.7%) subjects had insulin resistance and 102 (54.8%) met the criteria to be classi ed with metabolic syndrome. When the groups were compared according to sex, HDL levels were signi cantly higher in women (Table 1).
In the bivariate analysis, HDL showed an association with gender (p=0.04) and BMI (p=0.02). Moreover, total cholesterol was associated with smoking (p=0.048), whereas triglycerides and atherogenic index were associated with the presence of metabolic syndrome (p<0.001) ( Table 2).

Cardiovascular Risk Assessment
The average cardiovascular risk according to the Framingham algorithmic model was 2.63% (± 5.80). On the other hand, for the D:A:D algorithmic model (5 year prediction), the average risk in the reduced model was 2.3% (± 2.69) whereas for the full model, estimated risk was 3.11% (± 4.64), indicating a moderate risk (Table 3). In addition, when comparing the three algorithmic models, along with sex, smoking, diabetes and age, D:A:D full model showed to be more speci c for the prediction of a cardiovascular event in a period of ve to 10 years for people who lives with HIV (Table 3).
When comparing cardiovascular risk using the Framingham and D:A:D algorithmic models strati ed by BMI (normal weight, overweight and obese) no statistically signi cant differences were found; however, it can be observed that the risk is high with the D:A:D full model in the normal weight group compared to overweight/obesity (Table 4).
Finally, Table 5 shows a multivariate analysis of risk factors in which only smoking remained signi cant (p=0.002) after adjustment by age and sex for the Framingham model, whereas smoking and age were signi cant for D:A:D reduced model only (p=0.020 and p<0.001, respectively).
In individuals under cART attrition has become less common. Recent studies performed in high and low income countries have reported weight gain regardless of the type of cART used (34)(35)(36)(37)(38)(39). Among the most common co-morbidities among people living with HIV are type 2 diabetes, hypertension, respiratory diseases and liver disease, among others. (40)(41)(42)(43). Cardiovascular disease is a comorbidity of HIV infection which has also been linked to cART exposure in the HIV-seropositive population (44). It is known that comorbidity increases with HIV severity. The greater prevalence of comorbidities among people living with HIV/AIDS (PLWHA) may be attributed to antiretroviral toxicity (e.g. diabetes, vascular disease and liver disease) or caused by the HIV infection itself (e.g. vascular, pulmonary and renal diseases) (43,45). The use of antiretroviral therapy has been associated with obesity. Non-nucleoside reverse transcriptase inhibitors (NNRTIs), protease inhibitors (PIs) and integrase inhibitors (IIs) are among the most commonly associated, as they are lipophilic and susceptible to diffusion in adipose tissue, which concentrates antiviral activity and is related to plasma antiviral concentrations. (46, 47).
Cardiovascular risk estimation, based on the Framingham score, was designed in the general population, but its use in people living with HIV is not well de ned(28). Based on the prospective multicentre D:A:D study, which included 11 cohorts of HIV-positive patients treated in 212 clinics in the US, Europe, Argentina and Australia, algorithms were developed speci cally for this population. The DAD score was rst published in 2010, and took into account CD4+ cell count, abacavir use, and time of exposure to protease inhibitors and nucleoside reverse transcriptase inhibitors, in addition to classical cardiovascular risk factors. In order to simplify patient risk as well as take into account antiretroviral therapy, a change in the D: A: D score was established and published in 2016, assessing the same clinical outcomes over 5 years, but without using the classes and time of exposure to cART (27,48).
In our study, we evaluated different metabolic alterations in HIV-seropositive Mexican subjects, which were under cART treatment. We observed that prevalence of dyslipidemias (hypercholesterolemia, hypertriglyceridemia, HDL <40 mg/dL) was in ~38.6% of the individuals. Also this prevalence was higher in a population of China, which was 75.4% (49). It is known that lipid abnormalities prevail in HIV-seropositive people with cART and may contribute to increased risk of cardiovascular disease (50).
Mexico has been characterized by having a high prevalence of obesity in the HIV-seronegative population, in which ~75% of population has been classi ed as overweight/obesity based on BMI criteria (51). In addition, the study conducted by Xiaonli et al., 2019 in China showed that the male population presented a higher prevalence of overweight compared to the female population. Based on the WHO criteria for general population, we found that ~54.8% of the participants in this study presented metabolic syndrome. In contrast, in a study conducted by Sear et al., 2019 the prevalence of metabolic syndrome was 34% in an HIV-seropositive population from Southern United States (52). This percentage was also higher than that from a population of Bologna, Italy, where the prevalence of metabolic syndrome was 20.9% (53). Globally, it is estimated that the metabolic syndrome in the HIV-seropositive population is 29.6% (according to the ATPIII criteria), which is similar to that of a population not infected with HIV (54). Approximately 46.2% of our studied population was in overweight/obesity status, although still lower that an HIV-seropositive population from Peru which presented 52.70% (55). In a study in several populations from Latin America (Venezuela, Brazil, Colombia, Peru and Ecuador), in which PLWHA showed a high prevalence of metabolic syndrome, and 44% of the participants were under IP treatment (56). In the present study, we found a prevalence of 54.8% of the population with metabolic syndromes, with 65% of the participants receiving 2NTRI + 1 NNTRI in their treatment scheme.
In a study performed in a Ghanaian population in 2015, where the levels of transaminases in an HIV-positive population under treatment were compared by sex, values were found within normal parameters (57). However, in our study, when comparing the glutamic oxaloacetic transaminase (GOT)/glutamic-pyruvic transaminase (GTP) values by sex, a signi cant difference was found, in which males presented the highest levels. This could lead to the hypothesis that, because the male population has a higher BMI, GOT/GTP values could be different from normal parameters and could be involved in the development of nonalcoholic fatty liver disease (NAFLD)/non-alcoholic steatohepatitis (NASH). NALDF It is a liver disease that has been associated with obesity, diabetes and metabolic syndrome (58). Liver disease-related morbidity in people living with HIV remains high, even with advances in hepatitis treatment. NALDF has a prevalence of 50%. Factors such as insulin resistance, mitochondrial dysfunction and dyslipidaemia appear to be major factors for the increase in NALDF and NASH in people living with HIV. HIV infection combined with antiretroviral therapy increases the risk of developing NALDF. The use of rst generation nucleoside reverse transcriptase inhibitors and protease inhibitors is also associated with the development of NAFLD/NASH. (59).
In this study, HDL levels in the HIV-seropositive male population were lower compared to the female population. HDL low levels (<40 mg/dL) in men have been also reported in one HIV-seropositive Brazilian population (60) and one HIV-seropositive population from Medellin, Colombia (61). It has been also described that the activity of the cholesterol ester transfer protein (CETP), which transfers HDL-C cholesterol esters to proteins containing apolipoprotein B (62, 63), is high in HIV infection, thus, its activity is inversely correlated with serum HDL levels (63, 64). This may explain why these HDL-C low levels are present in HIV infected patients (63).
In this work, it was also observed that the relationship between lipid pro le and risk variables (such as smoking) may in uence the development of hypercholesterolemia, which may increase the risk of developing heart disease. A previous report in a Japanese population showed that this association (smoking and hypercholesterolemia) may in uence the development of ischemic cardiovascular accident and coronary artery disease (65). Also, it is well known that the use of tobacco may increase cholesterol levels in individuals (66, 67) and becomes a risk factor for the development of myocardial infarction in the general population since it is also associated with an increase in the lipid pro le (68).
The presence of metabolic syndrome observed in our HIV-seropositive population was considerably high (54.8%). However, a study conducted in India showed higher percentages (91%) for these two conditions in HIV-seropositive subjects (69). Metabolic syndrome is an independent predictive factor of cardiovascular disease (CVD) in HIV-infected individuals, however, there is also a strong association between the increasing number of metabolic syndrome components and the risk of CVD, emphasizing the importance of identi cation and management of all CVD factors in HIV-seropositive population under cART treatment (70). According to the literature, the presence of metabolic syndrome is positively correlated with atherogenic index, being known as a good predictor for the presence of a cardiovascular event. Thus, the atherogenic index may be an important factor that affects the risk of cardiovascular disease among people infected with HIV (71).
Among the analyzed cardiovascular risk factors in the D:A:D models, one of the most signi cant associations was smoking. The prevalence of smoking in HIV-seropositive population is similar compared to the general population (72)(73)(74). The risk factors that were identi ed for the development of a cardiovascular disease through the D:A:D model are the male gender, as well as age (75). Different factors such as hypercholesterolemia have been associated within the increased risk of cardiovascular disease (HR 1.21 CI 1.16-1.27) (76). When performing the associations, hypercholesterolemia was found as a risk factor; however, when adjusting for age, no association was found for this variable.
Another factor that has been increased in people living with HIV/AIDS in cART is triglyceride levels (hypertriglyceridemia) where high levels have been associated with cardiovascular risk (75).
In this study, when performing the associations, hypertriglyceridemia was identi ed as a risk factor; however, when adjusting for age it was not possible to establish as a possible cardiovascular risk factor. By analyzing the years under treatment (≥10 years) of patients living with HIV, it was observed that this could be a risk factor in the study population when analyzed with D:A:D full model: however, When performing the multinomial analysis, it could not be established as a risk factor. Kumar et al., 2016, in an HIV-positive HIV population, found that in uence the increase in cardiovascular risk and the presence of lipodystrophy; therefore, patients with HIV in cART are advised to undergo an early detection of cardiovascular risk and take appropriate measures to prevent progression to cardiac risk (77). In HIV-positive individuals with preserved immunity, immediate antiretroviral therapy has been shown to increase total cholesterol and low-density lipoproteins, as well as continuous increases in high-density lipoprotein cholesterol.. These opposite effects suggest that, in the short term, the effect of early cART on the risk factors of traditional cardiovascular disease may be clinically insigni cant (78). In this study, it was observed that the use of 2INTR + 2INNTR was associated with increased cardiovascular risk with the D:A:D full model; however, when adjusting for age, this association  (83). The use of these tools to identify patients with a cardiovascular risk pro le seems to be a simple way to monitor long-term CVD risks in the population of HIVinfected patients (84) and to de ne whose patients will need further assessment.

Conclusion
In this study with PLWH on ART, it was found that the factors that increase the risk and onset of cardiovascular disease at an early age are hypercholesterolemia, hypertriglyceridemia, metabolic syndrome, insulin resistance, smoking and age. In addition, the algorithmic models D: A: D and Framingham algorithmic models for estimating cardiovascular risk show a low to moderate risk; however, because the population is young, the results of cardiovascular risk should be taken with caution; in addition, exposure to ART, as well as years of infection and years under treatment, should be taken into account..

Declarations Competing Interests
The authors have declared that no competing interests exist.
Funding: The research did not receive nancial support from any institution Con icts of interest/Competing interests: The authors have declared that no competing interests exist.
Availability of data and material (data transparency): All data are contained in the manuscript

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementarymaterial.docx