Pathogenesis of cardiovascular disease in people with HIV infection

HIV-associated CVD is the result of complex interactions between traditional cardiovascular risk factors, side effects of ART and the chronic inflammation and immune activation associated with long term HIV infection. It is important to acknowledge, however, that the assessment and management of traditional CVD risk factors according to standard guidelines is still the cornerstone of clinical practice and should be prioritised among people with HIV infection on at least an annual basis.

Cigarette Smoking

Cigarette smoking is directly responsible for a significant proportion of the CVD experienced by people with HIV infection, partially because smoking rates are well above that reported in the general population but also because of an interaction between smoking and HIV-associated inflammation, which magnifies the negative effects of smoking above the already significant levels experienced in the general population [6]. It is estimated that more than 40% of heart attacks experienced by people with HIV infection could be prevented through smoking cessation [6]. Importantly, it is also clear that smoking cessation is associated with a substantial reduction in CVD risk (estimated at ~30%) that is evident within 2-3 years [7]. Smoking cessation interventions remain challenging among the relatively high proportion of people with HIV infection who are smokers (~30% in a recent Australian study) [8], particularly those with ongoing recreational drug and alcohol use. However, the diagnosis and treatment of HIV infection provide opportunities for successful intervention as a part of a discussion around improving general health [8]. Smoking cessation can also be supported by treatments such as varenicline, which has been shown to be safe and effective in people with HIV infection in a placebo-controlled trial [9].

Dyslipidemia

Dyslipidemia is also  an important driver of CVD, with elevations in total and low-density lipoprotein (LDL) cholesterol and deficiencies in high-density lipoprotein (HDL) cholesterol strongly related to the development of CVD [10]. HIV infection, independent of the effects of ART, can lead to significant alterations in cholesterol levels, predominantly characterized by decreases in total, LDL and HDL cholesterol with mild elevations in triglycerides [11]. In general, initiation of ART leads to further increases in triglycerides, along with increases in total and LDL cholesterol, sometimes to well above pre-seroconversion levels, but only modest, if any, correction in HDL [12]. Different antiretroviral drugs (as discussed below) have differing impacts on lipid levels which may be important when choosing an antiretroviral regimen for an individual at high risk of CVD.

In terms of evaluating cholesterol levels and considering the benefits of cholesterol lowering therapy in people with HIV infection, current evidence indicates that standard population-based calculators of CVD risk are still appropriate and valuable tools [13]. In this context, attempts have been made to develop CVD risk calculators that incorporate HIV-specific risk factors, although these have not been shown to improve the prediction of CVD events in a large population-based study [14], and have the potential to underestimate risk compared to standard tools such as the Framingham Risk Score.

From a practical perspective, there is a concern that true CVD risk in people with HIV infection is prone to underestimation when using all available risk calculators [15], which may reflect the 1.5-fold increased risk that has been attributed to HIV infection per se as noted above. The benefits of statin therapy in primary CVD prevention among people with HIV infection, and particularly among those with estimated CVD risk <15%, will ultimately be addressed by large-scale studies such as the REPRIEVE trial that has enrolled >7,500 participants across 11 countries [16]. In the meantime, it is reasonable to offer statin therapy to all people with HIV infection for whom this is recommended under current Australian guidelines [17 ], including those with an estimated 5-year CVD risk of >15%, as well as those with moderate risk (10-15%) and limited reversible risk factors.

Hypertension

CVD develops in 63.3% of hypertensive individuals (defined as a blood pressure ≥ 140/90) compared with 46.1% of normotensive controls. It also accelerates the development of atherosclerosis with cardiovascular events occurring on average 5.0 years earlier than would otherwise be expected [18]. Hypertension is common in people with HIV infection, with rates of 13-45% reported [19,20]. It is likely that both HIV infection and ART are contributing to the development of hypertension in this group but determining the relative contribution of ART, chronic inflammation due to HIV infection, traditional cardiovascular risk factors, and the interactions between these factors, is very difficult to determine. In people with HIV infection, elevated rates of CVD have been detected in those with pre-hypertension (systolic blood pressure 120-130 mmHg or diastolic blood pressure 80-89 mmHg) as well as hypertension (blood pressure ≥140/100 mmHg) (HR 1.81 [95%CI: 1.22-2.68] and 2.76 [95%CI: 1.9-4.02] respectively) [21], meaning that even small increases in blood pressure can have significant consequences.

At present there are no evidence-based guidelines that have been developed specifically for the management of hypertension among people with HIV infection, although the available data would indicate that this is an important risk factor for CVD and stroke that is increasing in prevalence among those that  are receiving ART [22]. In this context, the implementation of standard management guidelines [23] is appropriate and warrants attention as part of overall CVD risk management.

Gender

In the general population, CVD mortality is 1.4 times higher in men than in women [24], yet in HIV positive populations women appear to be disproportionately at increased risk of CVD. While absolute CVD rates remain low in women, HIV infection increases the baseline risk twice as much in women with HIV infection than it does in men with HIV infection [25,26], perhaps as the result of proportionally greater increases in immune activation [27].

Effects of ART on cardiovascular risk

It is important to highlight that CVD and all-cause mortality is overall dramatically decreased by ART, and that any side effects in terms of dyslipidemia or hyper-coagulation are clearly outweighed by the dramatic reduction in AIDS and non-AIDS related mortality [28].  However, given that a 20 year old who acquires HIV infection is likely to be on ART for 50-60 years, it is important to understand how different agents may be impacting on risk and select a regimen with the smallest side effect imprint whenever possible.

Different types of antiretroviral drugs can increase CVD risk in differing ways. Many ‘first generation’ HIV protease inhibitors (particularly indinavir and lopinavir) that were used in Australia in the early ‘HAART era’ were associated with significant metabolic complications including dyslipidaemia and elevated CVD risk. A recent analysis of contemporary HIV treatment involving 49,709 participants in the D:A:D observational cohort study has identified increased relative risk of CVD associated with cumulative exposure to darunavir after adjustment for other risk factors (59% increase per 5 years of additional use, confidence interval 33-91%). No association between CVD risk and atazanavir therapy was identified [29], in keeping with other evidence that atazanavir has a relatively favourable effect on CVD risk in comparison with other HIV protease inhibitors, as well as with other HIV medications [30]. It is therefore clear that there is no generalised HIV protease inhibitor ‘class effect’ on CVD risk, and that each medication needs to be considered individually.

The use of non-nucleoside reverse transcriptase inhibitors (NNRTIs) has not been associated with elevated CVD risk to date, including a recent updated meta-analysis [31]. Within this class, there is some evidence for more favourable lipid outcomes associated with rilpivirine compared to efavirenz treatment, both in treatment-naïve individuals [32], and in switching studies [33]. However, there is no evidence that efavirenz treatment is associated with elevated CVD risk in large population-based analyses [34].

With regard to nucleoside reverse transcriptase inhibitor (NRTI) therapy, cohort studies have identified an association between abacavir exposure and CVD events that appears to be related to current treatment rather than cumulative drug exposure. This was first identified in the D:A:D cohort study in 2008 [35] and the association was confirmed in an updated analysis published in 2016 [36]. In these analyses and in a subsequent meta-analysis of 17 epidemiological studies [37], current abacavir treatment was associated with an estimated 1.6-1.9-fold relative increase in underlying CVD risk; although this has not been replicated in an analysis of pooled clinical trial data involving abacavir therapy [38]. The underlying mechanism remains unclear, although there is some evidence that abacavir may be promoting a prothrombotic state and unmasking subclinical atheromas and promoting symptomatic CVD [39-41].  Based on this evidence, it seems prudent to avoid abacavir use, where possible, in individuals who have elevated CVD risk (ie >10-15%) and, in particular, in cases where these risk factors are not modifiable.

To date there has been no suggestion that integrase strand inhibitors (INSTIs) are associated with increased CVD risk or that INSTI treatment contributes to risk of metabolic complications, such as hyperlipidaemia or insulin resistance. There is emerging evidence that their use may be associated with increased fat accumulation and body mass [42]. This is an emerging area of research, with current indications that demographic factors may be important (with higher risk associated with female sex and black race) and that treatment-associated risk may be modified by choice of INSTI as well as NRTI therapy (including tenofovir alafenamide versus tenofovir disoproxil fumarate). Further results are awaited with interest.