In this retrospective cross-sectional analysis of patient records at a large tertiary hospital in Cape Town, South Africa, the 12-month period prevalence of HIV in those hospitalised with ACS was 3.4% (95% CI 2.3–4.8). HIV-positive patients were younger, predominantly male, had a lower prevalence of medical comorbidities and a lower prevalence of a family history of CAD, even after adjusting for age and sex. HIV-positive patients experienced more STEMIs and single-vessel disease. The LAD was the most commonly affected coronary artery, translating into more LAD-territory myocardial infarctions. The majority of the HIV-positive patients in our study were immune-suppressed (CD4+ counts < 500 cells/mm3) and virally-suppressed (VL < 200 copies/mL). Most HIV-positive patients were on ART at the time of ACS event, with the majority of them on NNRTI-based ART. Immediate and long-term management did not differ significantly between HIV-positive and HIV-negative patients. Our findings contribute data on the prevalence and presentation of HIV in ACS from South Africa with the largest ART program in the world and a rising rate of non-communicable disease-attributed morbidity and mortality [14].
The relative absence of traditional risk factors in the HIV-positive sub-group suggests additional pathogenic mechanisms in these patients; one being the higher rates of sub-clinical dyslipidaemia seen in HIV-positive individuals. HIV autopsy studies have shown evidence of premature CAD in HIV-positive patients even before initiation of ART, as a result of complex dyslipidaemic patterns: reduced total serum cholesterol, reduced high-density lipoprotein (HDL), reduced apolipoprotein B, and increased low-density lipoprotein (LDL) [4]. Despite a lower prevalence of diagnosed dyslipidaemia in our HIV-positive sub-group, we do not have data on their lipid profiles at the time of presentation for ACS, which may have revealed a higher prevalence of sub-clinical dyslipidaemia. The fact that only five (17%) HIV-positive patients in our study were receiving PIs, most strongly implicated in the causation of dyslipidaemia, suggests that ART cannot be the only risk factor for dyslipidaemia [2]. Studies showing the beneficial role of early ART-initiation in the prevention of CAD in HIV-positive individuals, as well as the increased risk of CAD and other comorbidities associated with ART-interruption further support this notion [15]. Therefore, it is plausible that it is the combined effect of HIV infection and certain ART drugs that results in complex sub-clinical dyslipidaemic patterns in HIV-positive patients, thereby increasing their risk for CAD and resultant ACS.
An additional suggested pathogenic mechanism for CAD in HIV-positive individuals is immune-system dysfunction, which can be directly measured by the number of CD4+ lymphocytes in the body. Lichtenstein et al. showed that a CD4+ count of less than 500 cells/mm3 is an independent risk factor for CAD, with comparative attributable risk of approximately 20% - a figure similar to several other traditional CAD risk factors [16]. The majority (81%) of patients in our study had CD4+ counts less than 500cells/mm3 which suggests an increased risk of CAD in these patients. Current detectable viraemia has been found to be a further risk factor for CAD, due to its contribution to an environment of persistent chronic inflammation [15]. Even virally suppressed HIV-positive patients have higher levels of inflammatory markers than those without HIV; thereby predisposing them to CAD [8].
The predominant presentation of single-vessel CAD and STEMIs in the HIV-positive patients in our study (even after adjusting for age and sex) is most likely due to the unique histological characteristics of coronary plaques in these patients. Virtual histology intravascular ultrasound analysis of HIV-positive patients affected by ACS has shown a high prevalence of unstable plaque morphology that is rich in necrotic tissue, less calcific, and has a thicker fibrous cap compared to that seen in traditional CAD [4]. Non-calcified plaques are more likely to rupture than calcific or mixed plaques, putting these patients at higher risk of single-vessel STEMI than their HIV-negative counterparts [17]. This higher plaque vulnerability is thought to be linked to the chronic inflammatory process of being infected with the HI virus itself. Furthermore, Moran et al. found that HIV-positive patients with single-vessel disease had higher Gensini Scores [18] than HIV-negative controls with single-vessel disease, indicating more severe vessel stenosis in the former [17]. The presence of more extensive, more vulnerable non-calcific, fibro-fatty plaque could also explain the higher prevalence and earlier onset of ACS in the HIV-positive population.
The mild systolic dysfunction seen in both groups is in keeping with the echocardiographic changes expected after a myocardial infarction. We would expect more advanced diastolic dysfunction in our HIV-negative sub-group as a result of the increased prevalence of hypertension in this group. A study done in Cameroon comparing the left atrial remodelling in hypertensives compared to healthy participants showed the hypertensive patients to have a larger left atrial diameter, surface area and volume, indicating an altered diastolic function in these patients [19]. Although statistically nonsignificant, it is plausible that the 11 µmol/L higher mean serum creatinine in the HIV-negative sub-group can be explained by the increased prevalence of medical comorbidities in this group, resulting in a higher incidence of target-organ damage which may manifest as sub-clinical/clinical chronic kidney disease. Serum creatinine is reported in many studies as a prognostic marker for overall cardiac mortality. In a study by Matts et al. it was found that each nine µmol/L (0.1 mg/dL) rise in baseline serum creatinine had a 36% increased relative risk of future overall mortality and a 47% increased relative risk for future atherosclerotic CAD mortality (no confounding factors present) [20]. This potentially translates into a 44% increased relative risk of overall future mortality and a 57% increased relative risk for future atherosclerotic CAD mortality in the HIV-negative patients in our study. Such a finding in our study cannot be over-interpreted due to our small sample size.
As there is no current evidence for a change in immediate or long-term management of CAD and ACS based on HIV status, all patients in our study were treated similarly. Simvastatin coadministered with PIs are expected in increase simvastatin concentrations markedly due to CYP3A inhibition [21]. However, we noted that 80% of patients on PIs incorrectly received simvastatin. The more frequent prescription of calcium-channel blockers and thiazide diuretics in the HIV-negative sub-group is most likely as a result of the increased prevalence of hypertension in this population.
Overall, our findings were consistent with the literature. There is limited global data on the prevalence of HIV in people presenting with ACS, but our findings are similar to the 2.4% HIV prevalence found in the CAD sub-group (consisting of 581 patients) of all de novo cases of heart disease presenting to a tertiary hospital in Soweto, South Africa [22]. These values are significantly lower than both the 13.1% estimated national HIV-prevalence in South Africa in 2018 and the 7% prevalence of HIV in a population of people with ACS in a HIC such as Spain [11]. This may be due to the under-reporting of HIV in ACS in South Africa as evidenced by the large portion of untested patients in our study (Fig. 1) [1]. Furthermore, it may be due to other confounding factors; one being age: HIV-prevalence is highest in young adults whereas ACS primarily affects the elderly. Despite this, data from a study conducted in Québec, Eastern Canada, showed the incidence of ACS to be 3.88 in the HIV-positive cohort compared to 2.21 in the HIV-negative cohort per 1000 patient-years, irrespective of exposure to ART [23]. A study conducted in Boston, Massachusetts, showed increased ACS rates per 1000 person-years in HIV-positive vs HIV-negative patients (11.13 vs 6.98), even after adjusting for age, sex, race, hypertension, diabetes and dyslipidaemia [12]. The demographic profile and CAD risk factor profile of the patients in our study was consistent with that seen in the literature. Furthermore, the type of ACS and number of involved coronary arteries of the HIV-positive patients in our study matched that seen in the literature. Various studies from the USA and France showed angiographic findings of fewer involved vessels and a greater burden of inflammatory plaque in the HIV-positive cases as compared to the HIV-uninfected cases [7]. Our findings of a predilection for LAD artery involvement in the HIV-positive patients are supported by those of Vachiat et al. who showed the LAD to be the most common culprit artery (60%) in HIV-positive patients with ACS [24]. The same proportion of HIV-positive and HIV-negative patients in our study underwent percutaneous coronary intervention (PCI), but we have no follow-up data on long-term success rates of such an intervention. A Spanish study showed a lower long-term success-rate in HIV-positive patients who underwent PCI compared to HIV-negative patients (75% versus 85% success-rate respectively) [25].
An interesting finding from our study was that despite a nonsignificant increased unadjusted prevalence of smoking seen in HIV-positive compared to HIV-negative patients, this finding was reversed when adjusting for age and sex. This finding could perhaps explain the heterogeneity in smoking data when comparing HIV-positive and HIV-negative patients with ACS. The majority of studies show a positive correlation between HIV-infection and smoking; however, when matching on age and sex, Dwyer et al found no difference in rates of smoking between HIV-positive and HIV-negative sub-groups [1, 26].
Our study has a number of limitations. First, the retrospective, observational design limits its ability to control for unmeasured confounders. Second, our sample size is limited which did not allow us to perform sub-group analyses. Third, we may have underestimated the HIV-prevalence within the ACS population due to the relative infrequent reporting of HIV-results. It is not routine for everyone presenting to a healthcare facility in South Africa to be tested for HIV, meaning that the actual prevalence of HIV within the ACS population might be higher than reported in our study. Last, we were confronted with missing data. CD4+ counts and viral loads were not available for all HIV-positive patients, and when available, were not always taken during the same admission as for the ACS event.