The aim of the study was to assess the separate contributions of smoking, diabetes and hypertension to acute coronary syndrome (ACS) in HIV-infected adults relative to uninfected adults.
The aim of the study was to assess the separate contributions of smoking, diabetes and hypertension to acute coronary syndrome (ACS) in HIV-infected adults relative to uninfected adults.
Two parallel case–control studies were carried out. In the first study, HIV-positive adults diagnosed with ACS between 1997 and 2009 (HIV+/ACS) were matched for age, gender and known duration of HIV infection with HIV-positive adults without ACS (HIV+/noACS), each individual in the HIV+/ACS group being matched with three individuals in the HIV+/noACS group. In the second study, each individual in the HIV+/ACS group in the first study was matched for age, gender and calendar date of ACS diagnosis with three HIV-negative individuals diagnosed with ACS between 1997 and 2009 (HIV–/ACS). Each individual in the HIV–/ACS group was then matched for age and gender with an HIV-negative adult without ACS (HIV–/noACS). After matching, the ratio of numbers of individuals in the HIV+/ACS, HIV+/noACS, HIV–/ACS and HIV–/noACS groups was therefore 1 : 3 : 3 : 3, respectively. We performed logistic regression analyses to identify risk factors for ACS in each case–control study and calculated population attributable risks (PARs) for smoking, diabetes and hypertension in HIV-positive and HIV-negative individuals.
There were 57 subjects in the HIV+/ACS group, 173 in the HIV+/noACS group, 168 in the HIV–/ACS group, and 171 in the HIV–/noACS group. Independent risk factors for ACS were smoking [odds ratio (OR) 4.091; 95% confidence interval (CI) 2.086–8.438; P < 0.0001] and a family history of cardiovascular disease (OR 7.676; 95% CI 1.976–32.168; P = 0.0003) in HIV-positive subjects, and smoking (OR 4.310; 95% CI 2.425–7.853; P < 0.0001), diabetes (OR 5.778; 95% CI 2.393–15.422; P = 0.0002) and hypertension (OR 6.589; 95% CI 3.554–12.700; P < 0.0001) in HIV-negative subjects. PARs for smoking, diabetes and hypertension were 54.35 and 30.58, 6.57 and 17.24, and 9.07 and 38.81% in HIV-positive and HIV-negative individuals, respectively.
The contribution of smoking to ACS in HIV-positive adults was generally greater than the contributions of diabetes and hypertension, and was almost twice as high as that in HIV-negative adults. Development of effective smoking cessation strategies should be prioritized to prevent cardiovascular disease in HIV-positive adults.
HIV-infected patients may expect to live longer now than ever before, with the widespread availability of effective antiretroviral therapy, but their life expectancy is still lower than that of the general population . In this context, cardiovascular disease has emerged as an increasing cause of morbidity [2-4] and mortality [5-7] in HIV-infected patients. A high prevalence of tobacco, alcohol and illicit drug consumption [8, 9], immunodeficiency , and immune activation and inflammation caused by HIV replication [11, 12] are contributing factors that may explain the higher than expected incidence of cardiovascular disease in HIV-infected persons. Effective antiretroviral therapy is able to ameliorate immunodeficiency and to decrease immune activation and inflammation, but it cannot entirely resolve the problems associated with HIV infection [13, 14]. In addition, some antiretroviral drugs may themselves contribute to cardiovascular disease by causing metabolic abnormalities and possibly through other mechanisms that are not yet completely understood [4, 15].
Specific sections addressing the prevention of cardiovascular disease have been developed in major guidelines for the management of HIV infection [16-18]. In addition to earlier initiation of antiretroviral therapy, the updated 2011 version of the European AIDS Clinical Society guidelines recommends the promotion of healthy lifestyle measures and adequate management of diabetes, dyslipidaemia and hypertension . In general, recommendations for HIV-infected patients follow those for the general population, assuming that similar responses to the management of traditional cardiovascular risk factors will result in similar benefits in terms of decreasing the risk of cardiovascular disease.
A critical unanswered question regarding the assessment, prevention and management of cardiovascular disease in HIV-infected patients is the degree to which traditional risk factors such as smoking, diabetes and hypertension increase cardiovascular risk in the HIV-infected population. This is an important question because HIV-infected patients are at risk of cardiovascular disease at a younger age than the general population, with HIV infection, antiretroviral therapy, and other risk factors and comorbid conditions modifying the effects of a given risk factor. Although smoking, diabetes and hypertension have consistently been shown to be involved in the development of cardiovascular disease in both HIV-uninfected and HIV-infected adults, the prevalence of these factors may vary between HIV-infected and HIV- uninfected adults, and, if this is the case, different interventions may require to be prioritized in HIV-infected patients. The contributions of smoking, diabetes and hypertension to myocardial infarction may also depend on additional factors such as the geographical origin of the population. Smoking, diabetes and hypertension accounted for 36, 10 and 18%, respectively, of attributable risk of myocardial infarction in the INTERHEART study carried out in 52 countries representing all inhabited continents , but these factors accounted for 55, 26 and 50%, respectively, of the attributable risk in young Catalonian adults . Population attributable risk (PAR) is the portion of the incidence of a disease in the population (exposed and unexposed) that is attributable to exposure. In other words, PAR resulting from a certain risk factor represents the reduction in incidence that would be expected if exposure to this factor were completely eliminated. Smoking [21, 22], diabetes [23, 24] and hypertension [25, 26] have been more commonly reported in HIV-infected patients than in the general population. Therefore, it could be argued that their absolute contributions to myocardial infarction are higher in HIV-infected patients than in the general population. However, HIV-infected patients have additional contributions from other risk factors, including HIV infection and antiretroviral therapy, which might ultimately reduce the relative contributions of smoking, diabetes and hypertension in this population. We aimed to determine the extents to which smoking, diabetes and hypertension in HIV-infected patients contribute to acute coronary syndrome (ACS) in terms of PAR relative to non-HIV-infected adults from the same geographical area.
We designed two parallel case–control studies including HIV-infected (HIV+) and uninfected (HIV–) adults, respectively. For each participant, clinical information was required on smoking, diabetes and hypertension prior to or on the date of the ACS event for cases and the date of censorship for controls. Current smoking was defined as active smoking within at least 6 months prior to the date of the ACS event or censorship. Diabetes was defined as having been clinically diagnosed with diabetes and having received any anti-diabetic therapy, or having had plasma glucose ≥ 200 mg/dL or confirmed fasting plasma glucose ≥ 126 mg/dL within at least 6 months prior to the date of the ACS event or censorship . Hypertension was defined as having been clinically diagnosed with hypertension and having received any anti-hypertensive therapy, or having had confirmed blood pressure ≥ 140 (systolic) or 90 (diastolic) mmHg within at least 6 months prior to the date of the ACS event or censorship .
In addition to smoking, diabetes and hypertension, collection of other available clinical or laboratory data with a potential impact on cardiovascular risk was also attempted. For both HIV-positive and HIV-negative participants, we were able to collect data on age, gender, family history of cardiovascular disease, and plasma total cholesterol. Hypercholesterolaemia was defined as having been clinically diagnosed with hypercholesterolaemia and having received any cholesterol-lowering therapy, or having had confirmed plasma total cholesterol > 240 mg/dL within at least 6 months prior to the date of the ACS event or censorship . For HIV-positive patients only, we also collected data on the following variables: years from diagnosis of HIV infection, AIDS (defined using the 1993 classification of the Centers for Disease Control and Prevention) , chronic hepatitis C virus (HCV) infection (defined as positive HCV serology), current illicit drug abuse (defined as any consumption at least 6 months prior to the date of the ACS event or censorship), peak plasma viral load, nadir CD4 cell count, and current (e.g. within 6 months prior to the date of the ACS event or censorship) therapy with thymidine nucleoside reverse transcriptase inhibitors, abacavir or protease inhibitors.
ACS was defined according to the criteria of The Joint European Society of Cardiology and the American College of Cardiology Committee for the Redefinition of Myocardial Infarction . The study was approved by the Ethics Committee at each participating centre.
For the HIV-positive case–control study, we identified HIV-positive adults diagnosed with ACS between 1997 and 2009 (HIV+/ACS) from hospital records. For each subject in the HIV+/ACS group, we selected three HIV-positive patients without ACS from HIV databases, matched for age (± 3 years), gender and known duration of HIV infection (± 3 years) (HIV+/noACS).
For the HIV-negative case–control study, we identified patients diagnosed with ACS between 1997 and 2009 with no known diagnosis of HIV infection at the time of the ACS event (HIV–/ACS) and, for each individual in the HIV+/ACS group, we randomly selected three HIV–/ACS individuals matched for age (± 3 years), gender and calendar date of ACS diagnosis (± 3 years). Each of these HIV–/ACS individuals was matched for age (± 3 years) and gender with a healthy adult volunteer (HIV–/noACS) selected from Hospital Clínic Primary Care Centre databases. After matching, the ratio of numbers of individuals in the HIV+/ACS, HIV+/noACS, HIV–/ACS and HIV–/noACS groups was therefore 1 : 3 : 3 : 3, respectively.
The effects of smoking, diabetes, hypertension and other available cardiovascular risk factors on ACS in each case–control study were assessed by unconditional logistic regression adjusted for the matching criteria. Odds ratios (ORs) and their corresponding 95% confidence intervals (CIs) were calculated for every risk factor of interest. PARs for smoking, diabetes and hypertension were calculated by unconditional logistic regression within each case–control study . PARs were adjusted for confounders in a similar manner to the corresponding logistic regression models for OR estimates. Statistical analyses were performed with sas version 9.2 (SAS Institute, Cary, NC) and stata, release 9.1 (Stata Corp, College Station, TX). All statistical tests of hypotheses were two-sided.
Although 71 HIV+/ACS patients were identified, 14 (all men) were excluded from the analysis because they did not have sufficient data available for the purpose of this study. Therefore, there were 57 subjects in the HIV+/ACS group, 173 in the HIV+/noACS group, 168 in the HIV–/ACS group and 171 in the HIV–/noACS group. Table 1 shows the characteristics of the HIV-positive and HIV-negative participants. Fifty-seven per cent of participants were men, and their mean age was 53 years. Smoking, diabetes, hypertension, family history of cardiovascular disease, and hypercholesterolaemia were more common in patients with ACS than in those without, in both HIV-positive and HIV-negative participants. The prevalences of smoking, diabetes, hypertension and hypercholesterolaemia are shown in Figure 1. In patients with ACS, the prevalence of smoking in the HIV-positive group was almost double that in the HIV-negative group, the prevalence of diabetes was similar, and the prevalence of hypertension in the HIV-positive group was nearly half that in the HIV-negative group. In participants without ACS, the prevalences of smoking, diabetes and hypertension in the HIV-positive group were double those in the HIV-negative group. The prevalences of hypercholesterolaemia were similar in the HIV+/ACS, HIV+/noACS, HIV–/ACS and HIV–/noACS groups.
|ACS||noACS||P value||ACS||noACS||P value|
|(n = 57)||(n = 173)||(n = 168)||(n = 171)|
|Age (years)||53 ± 11||53 ± 9||NS||54 ± 11||52 ± 13||NS|
|Smoking (%)||72||42||< 0.0001||40||24||0.0226|
|Hypertension (%)||25||20||NS||46||12||< 0.0001|
|Family history of cardiovascular disease (%)||12||3||0.0287||17||11||NS|
Regarding HIV-positive participants, approximately one-third had a previous diagnosis of AIDS and roughly one-quarter had chronic hepatitis C (Table 2). Seven per cent were current users of illicit drugs; 11% of individuals in the HIV+/ACS group admitted use of cocaine compared with 3% of the HIV+/noACS group (P = 0.0591). The mean nadir CD4 count was 200 cells/μL and the mean peak log HIV-1 RNA was 4.8 HIV-1 RNA copies/mL. Seventy per cent of individuals in the HIV+/ACS group had a most recent measurement of plasma HIV RNA below the quantification limit compared with 60% of the HIV+/noACS group (P = 0.3647). Antiretroviral therapy within 6 months prior to the date of the event (cases) or the date of censorship (controls) included thymidine nucleoside reverse transcriptase inhibitors in 40%, abacavir in 20%, and protease inhibitors in 26% of patients. None of the characteristics related specifically to HIV infection showed significant differences between the HIV+/ACS and HIV+/noACS groups.
|(n = 57)||(n = 173)|
|Years from HIV+ diagnosis||12 ± 7||12 ± 7||NS|
|Chronic hepatitis C (%)||28||27||NS|
|Current illicit drug abuse (%)||12||6||NS|
|Consumption of cocaine||6||5||NS|
|Peak log HIV-1 RNA copies/mL||4.8 ± 1.0||4.8 ± 1.1||NS|
|Nadir CD4 count (cells/μL)||224 ± 202||192 ± 160||NS|
|Current antiretroviral therapy:|
|Thymidine NRTIs (%)||37||42||NS|
|Protease inhibitors (%)||26||26||NS|
Considering all HIV-positive participants, smoking (OR 4.091; 95% CI 2.086–8.438; P < 0.0001) and a family history of cardiovascular disease (OR 7.676; 95% CI 1.976–32.168; P = 0.0003) were identified as independent risk factors for ACS in the multivariate analysis, while diabetes (OR 1.540; 95% CI 0.550–4.119; P = 0.3949), hypertension (OR 1.315; 95% CI 0.597–2.895; P = 0.4971) and hypercholesterolaemia (OR 1843; 95% CI 0.978–3.473; P = 0.0585) were not.
Considering all HIV-negative participants, smoking (OR 4.310; 95% CI 2.425–7.853; P < 0.0001), diabetes (OR 5.778; 95% CI 2.393–15.422; P = 0.0002) and hypertension (OR 6.589; 95% CI 3.554–12.700; P < 0.0001) were identified as independent risk factors for ACS in the multivariate analysis, while hypercholesterolaemia (OR 1.329; 95% CI 0.852–2.073; P = 0.2104) and a family history of cardiovascular disease (OR 1.269; 95% CI 0.663–2.428; P = 0.4718) were not.
Results obtained using the other logistic regression model were highly consistent. Considering only patients with ACS (Table 3), smoking significantly increased the risk of ACS 4-fold in both HIV-positive and HIV-negative patients. In addition, diabetes and hypertension significantly increased the risk 5-fold and 6-fold, respectively, in HIV-negative patients, but these factors did not significantly increase the risk in HIV-positive patients (Table 3).
|Risk factor||No. Case-patients||Proportions of case-patients (pi)||aOR||aOR 95% CI||p-value||PAR (%)||PAR 95% CI|
|Combination of factors||68.75||57.94–76.77|
|Combination of factors||60||32.85–76.24|
The calculated PARs resulting from smoking, diabetes and hypertension in HIV-positive and HIV-negative patients with ACS are shown in Table 3. The combination of these three factors accounted for approximately two-thirds of PAR in both HIV-positive and HIV-negative patients. In contrast, PARs resulting from diabetes and hypertension were 3 and 4 times lower, respectively, in HIV-positive than in HIV-negative patients. However, their individual contributions were different in HIV-positive and HIV-negative patients. The PAR resulting from smoking in HIV-positive patients was nearly double that in HIV-negative patients. In HIV-positive patients, the PAR resulting from smoking was several times higher than that resulting from diabetes or hypertension, and accounted for most of the PAR resulting from the combination of these three factors. In HIV-negative patients, PARs resulting from hypertension, smoking and diabetes were more similar among each PAR value compared with the others and the contribution of each factor was substantially lower than the PAR resulting from the combination of the three factors.
The most important finding of our study is that we were able to detect differences between HIV-positive and HIV-negative adults in the PARs for developing ACS resulting from smoking, diabetes and hypertension. Smoking was the greatest contributor to ACS in HIV-positive patients, explaining 54% of the PAR compared with 60% of the PAR explained by the combination of the three factors. Smoking has been recognized as one of the major contributors to cardiovascular disease in the general population  and consequently active smoking is included (and has an important relative weight in comparison with other factors) in most scores estimating cardiovascular risk. In general, HIV-positive adults have a higher prevalence of smoking than HIV-negative adults, and the reasons for this are probably multifactorial. Smoking rate and characteristics in HIV-positive adults have been associated with factors already described in the general population, such as male sex and smoking environment, but also with factors specific or more common to the HIV-infected population, such as disclosure of HIV status and reported experience of disclosure rejection, and higher rates of alcohol and illicit substance use . In HIV-positive adults, major smoking-related health risks include not only cardiovascular disease but also non-AIDS neoplasia, bacterial pneumonia, and overall mortality . On the plus side, smoking is a modifiable cardiovascular risk factor. Of all interventions targeting modifiable cardiovascular risk factors in the general population, stopping smoking is probably the intervention with the greatest impact on the risk of cardiovascular disease . The Data Collection on Adverse events of Anti-HIV Drugs (D:A:D) study found that the risk of myocardial infarction and cardiovascular disease decreased with each passing year of having stopped smoking, and the risk almost halved after 3 years . Smoking cessation programmes following a similar design as in the general population have been developed [37, 38], with a success rate of approximately 25% at 1 year. Unfortunately, smoking cessation interventions for HIV-positive adults are not easy to incorporate into routine clinical practice. Specific approaches with the aims of improving the incorporation of smoking cessation strategies by HIV doctors into clinical practice  and obtaining better responses given the unique needs of HIV-positive adults  have been suggested. Our study confirms that the contribution of smoking to ACS in HIV-positive adults is even higher than that in the HIV-negative population, and consequently the need to stop smoking should be prioritized in HIV-positive adults.
Although diabetes and hypertension were more prevalent in HIV-positive than in HIV-negative adults in participants both with and without ACS, our study suggests that their contribution to ACS (as defined by PAR) in HIV-positive individuals was actually smaller than in HIV-negative individuals. How should these data be interpreted? Participants in our study were matched for age, and the mean age of included subjects was 53 years. This unexpected result could be explained by the relatively young mean age of our patients with ACS. The prevalences of diabetes and hypertension increase with age, and so similar increases might be expected for their ACS-related PARs . Thus, with increasing age, differences in the PARs resulting from diabetes and hypertension between HIV-positive and HIV-negative adults may become smaller, although this explanation remains speculative. Management of diabetes and hypertension in HIV-positive adults is largely based on recommendations for the general population . Although there is a paucity of data concerning complications of HIV-associated diabetes and hypertension, HIV physicians should nevertheless pursue optimal management of these conditions in HIV-positive patients through more aggressive screening and targeted prevention and treatment strategies with hard cardiovascular endpoints.
Our study has some important limitations. The absolute number of HIV-positive patients with documented ACS was low despite the study being a collaborative initiative between two major centres covering a period of more than 10 years. This may be a result in part of the low incidence of ACS in the HIV-positive population. We excluded some HIV-infected patients because they had insufficient data for the purpose of this study. Although we made every effort not to miss any HIV-positive patient diagnosed with ACS, we cannot completely rule out the possibility that some were missed. Despite these limitations, the estimated incidence of myocardial infarction in our cohort would have been 1.75 cases per 1000 patient-years, which is not different from that reported in major cohorts such as the French Hospital Database on HIV ANRS Cohort CO4 (1.24 cases per 1000 patient-years) , although it is lower than that for the D:A:D study (3.3 cases per 1000 person-years) . In addition, 42% of the HIV+/ACS group in our study were women, a percentage that is twice as high as that reported in a recent meta-analysis . As a consequence of the retrospective nature of our analysis, all HIV-infected patients who experienced myocardial infarction in our cohort were not necessarily included in this study. In fact, all 14 patients excluded because of the unavailability of data were men. Nevertheless, our study was designed to control for age and gender, so no biases from these variables should be expected. As in any other retrospective study, we had no information available on a number of variables of potential interest for ACS in both HIV-positive and HIV-negative participants. One of these was the use of cocaine, as this factor has been recently associated with the risk of ACS in our area , particularly in persons younger than 30 years (25%) relative to those aged 45–50 years (5.5%). In our study, the mean age of participants was 53 years, and 11% of the HIV+/ACS group admitted the use of cocaine. This prevalence was higher than that in the HIV+/noACS group (3%) (P = 0.0591), but we had no data on cocaine use in non-HIV-infected persons. Because HIV-positive patients commonly had regular follow-up data, some variables were available in HIV-positive but not HIV-negative participants.
Our study also has some notable strengths. It is the first study, to our knowledge, to assess the PARs of common traditional cardiovascular risk factors in the HIV-positive population. We compared, as accurately as possible, the PARs of those factors between HIV-positive and HIV-negative adults, matching for age and gender in both HIV-positive and HIV-negative participants, and the known duration of HIV infection in HIV-positive participants and calendar date of ACS diagnosis in HIV-negative participants. Moreover, we took particular care to use similar working definitions of risk factors and outcomes in both HIV-positive and HIV-negative populations.
In conclusion, in our study, the contribution of smoking to ACS in HIV-positive adults was almost twice that in HIV-negative adults, and the contribution of smoking in the HIV-positive population was greater than those of diabetes and hypertension. If our results are confirmed, a substantial reduction in the incidence of ACS in HIV-positive adults should be expected if the contribution of smoking can be eliminated. Future efforts to decrease the burden of cardiovascular disease should focus on developing widespread and effective smoking cessation strategies in HIV-positive adults.
This work was supported in part by research grants from Red Temática Cooperativa de Investigación en SIDA (RIS G03/173), Ministerio de Sanidad, Política Social e Igualdad, Spain.
Author contributions: MC-S and EM designed the study, helped with analysis of the data and drafted the manuscript. RP helped to design the study, interpret the results, and draft and revise the manuscript. IP performed statistical analyses and led interpretation of the results. MGM, MJ, ML, JLB, MM-R, MS, JM, JMG and PD helped with collection and interpretation of data and with revision of the manuscript.