Cardiovascular disease and cancer are the two leading causes of death in Portugal and worldwide.1,2 Advances in cancer treatment in recent decades have led to improvements in life expectancy for these patients, but also to an increase in the burden of cardiovascular disease, with the risk of developing an acute coronary syndrome (ACS) at least double that of the general population.3–5 Observational studies have shown that cancer at various stages may be observed in up to 15% of patients with ACS.6–8

The association between cancer and ACS is complex and multifactorial. There are several shared risk factors in the etiopathogenesis of coronary artery disease and cancer, such as older age, male gender, smoking, diabetes and obesity.9 Malignancy itself is associated with a hypercoagulable and prothrombotic state which promotes oxidative stress and progression of atherosclerosis.10 Also, cancer treatment (chemotherapy and radiotherapy) have additional cardiotoxic effects, increasing the risk of arterial thrombosis and vasospasm.11

Cancer patients are furthermore at twofold higher risk of bleeding,12 and it is well known that bleeding in ACS is associated with higher risk of mortality.13 Management of ACS in this population is challenging due to older age, frailty, presence of anemia, thrombocytopenia and coagulopathy, nutritional deficiencies, tumor and metastasis bleeding risk, vascular effects of chemotherapy and radiation therapy, delayed stent endothelization, prevalent use of anticoagulation for venous thromboembolism or atrial fibrillation, more frequent renal and hepatic dysfunction and probability of interruption of antiplatelet therapy in the event of urgent surgery, biopsy or re-initiation of cancer therapy.7,8,14,15 The new European non-ST-elevation myocardial infarction (NSTEMI) guidelines recognized these issues and considered the presence of active cancer as a major criterion at the time of percutaneous coronary interventions (PCI) for high bleeding risk.16 Concerns about bleeding risk may contribute to suboptimal use of invasive diagnostic and treatment strategies and evidence-based recommended medication in this population, which in turn appears to increase both short- and long-term mortality.14,17,18

Randomized controlled trials assessing the efficacy and safety of ACS treatment have excluded cancer patients and observational studies have shown varying results in terms of outcomes.19–21 The aim of the present study was to assess the safety and efficacy of ACS treatment strategies in cancer patients, analyzing data from a multicenter registry of contemporary patients treated in a European country.

At total of 18 845 consecutive ACS patients were included in the present study, of whom 934 (5%) had a diagnosis of cancer. Baseline characteristics of both populations are presented in Table 1.

Compared to patients without malignancy, cancer patients were older, fewer were male and fewer were active smokers, but they had more cardiovascular risk factors such as hypertension and diabetes, and more previous history of myocardial infarction, PCI, heart failure, atrial fibrillation and bleeding events. The cancer group also had more comorbidities, including chronic kidney disease, peripheral vascular disease, chronic obstructive pulmonary disease and dementia. Mean hemoglobin was significantly lower in cancer patients.

In terms of ACS, both groups most frequently had NSTEMI or unstable angina (54% in cancer patients vs. 51%), followed by STEMI (41% vs. 45%) and undetermined myocardial infarction (5% vs. 4%). Cancer patients less often presented with chest pain and more often with symptoms and signs of heart failure (Killip class >1).

During hospitalization, cancer patients had lower LVEF and less often underwent ICA and PCI, even if they had more severe coronary anatomy. When PCI was performed, there were no significant differences in rates of balloon angioplasty between groups (5.8% vs. 5.3%), while the cancer group had higher rates of bare-metal stent implantation (23.3% vs. 18.6%, odds ratio [OR] 1.33, p=0.0003).

Regarding the endpoints, cancer patients had more events: major bleeding (2.9% vs. 1.5%, p<0.001), in-hospital mortality (5.8% vs. 3.4%, p<0.001) and the combined endpoint of in-hospital mortality, reinfarction and ischemic stroke (7.4% vs. 4.9%, p<0.001) (Table 2). Cancer was an independent predictor of major bleeding events (OR 1.97; 95% confidence interval [CI] 1.36-2.88, p<0.001) and cancer patients with the primary endpoint (n=27) had significantly higher in-hospital mortality (22.2% vs. 5.3%, OR 5.1; 95% CI 1.92-13.26, p=0.003), lower baseline hemoglobin levels (11.9±2.6 g/dl vs. 12.9±2.1 g/dl, p=0.032) and a trend to lower platelet count (16.7% vs. 13.3%, p=0.051). Aspirin, P2Y12 inhibitors (ticagrelor and prasugrel) and dual antiplatelet therapy (DAPT) were less frequently prescribed in these patients. Four patients (15%) with the primary endpoint were on oral anticoagulants: three with vitamin K antagonists (VKAs) (9% of VKA users) and one with a non-vitamin K antagonist oral anticoagulant (NOAC) (8% of NOAC users).

In the cancer group (n=934), the primary endpoint was associated with a previous bleeding event, hemoglobin at admission, atrial fibrillation, higher creatinine level during hospitalization and oral anticoagulation. In the multivariate analysis, previous bleeding events and atrial fibrillation were independent predictors of the primary endpoint (Table 3).

Cancer was also an independent predictor of in-hospital mortality (OR 1.73; 95% CI 1.30-2.30, p<0.001). Patients who died were older (79±9 vs. 72±11 years, p<0.001) and had more thrombocytopenia (26.5% vs. 12.6%, p=0.006), anemia (minimum hemoglobin 10.8±2 vs. 11.6±2 g/dl, p=0.014), renal impairment (23.1% vs. 10.1%, p=0.003), STEMI (61.1% vs. 39.7%, p=0.002) and atrial fibrillation (20.4% vs. 9.4%, p=0.009) and lower LVEF (38±13% vs. 50±12%, p<0.001). In the univariate analysis on cancer patients, mortality was associated with lower prescription of neurohormonal blockers and antiplatelet therapy – no antiplatelet treatment (OR 10.29), single antiplatelet treatment (OR 2.15) – and less frequent ICA (OR 0.23) and PCI (OR 0.49). There was also a trend for less radial access use (p=0.068) and more frequent left main disease as culprit lesion (12.5 vs. 3.1%, p=0.043). We found no association between anticoagulation treatment (oral or parenteral) or type of percutaneous treatment (balloon vs. stent implantation) and in-hospital mortality. In the multivariate analysis, STEMI, LVEF<40%, thrombocytopenia, chronic kidney disease, no ACE inhibitor therapy and no use of ICA were independent predictors of in-hospital mortality (Table 3).

Sixty-nine (7.4%) cancer patients reached the secondary combined endpoint (in-hospital mortality, reinfarction and ischemic stroke). This group less often received antiplatelet therapy (none [4.3 vs. 0.6%], single [17.4 vs. 9.5%] or dual [78.3 vs. 89.9%]) and neurohormonal blocker therapy (53.6 vs. 85.8%, p<0.001) and less frequently underwent ICA (54% vs. 80.5%, p<0.001), and there was a trend to less PCI (49% vs. 60%, p=0.074). We found no differences in anticoagulant prescription rate or type of percutaneous treatment in patients who reached the secondary combined endpoint. In the multivariate analysis, STEMI, LVEF <40%, Killip class >I, thrombocytopenia, creatinine >2 mg/dl, no ACE inhibitor therapy and no use of ICA were independent predictors of the combined secondary endpoint (Table 3).

Our study confirmed that cancer patients represent a non-negligible group in the ACS population (5%) and that they have a worse prognosis during hospitalization for ACS, including more major bleeding and ischemic events and higher mortality.

Several factors may be responsible for this increased risk of events. Firstly, and similar to previously published results,23,24 cancer patients were older and had more comorbidities and more severe coronary artery disease. Although the proportion of patients undergoing thoracic radiation therapy is unknown, the high incidence of left main disease in our cohort could be related to previous radiotherapy.25 Presentation may be more atypical, due to radiotherapy- and chemotherapy-related neurotoxicity, which can affect the ability to feel pain and also due to the frequent use of opioids and other analgesics.26 In addition, cancer patients have lower cardiac reserve, related to previous myocardial infarction, PCI and heart failure at presentation.20 Our propensity score analysis is in line with this hypothesis, because homogenization in terms of baseline characteristics blurs the differences in endpoints.

In our overall population, major bleeding events were associated with increased mortality and cancer diagnosis. A previous study also demonstrated that more than half of noncardiac deaths in cancer patients were due to bleeding events.20 Bleeding events were associated with previous predisposition to bleeds and hemoglobin level, use of oral anticoagulation, atrial fibrillation and renal impairment. However, bleeding risk was not significantly higher compared to patients with similar characteristics, specifically previous bleeding history, oral anticoagulation, atrial fibrillation and diagnosis of chronic kidney disease. In the paired populations, only mean hemoglobin level was significantly lower in patients with cancer. However, mean hemoglobin level at admission and during hospitalization was still above 12 g/dl. Other studies have shown that lower hemoglobin at admission could predict severe bleeding.24,27 In a substudy of the BleeMACS registry, hemoglobin <11 g/dl was the strongest predictor of severe spontaneous bleeding at one year after ACS.24 However, in our overall population, a previous bleeding episode was the most powerful multivariate predictor of the primary endpoint during hospitalization.

Moreover, it has been reported that cancer and its treatment contribute to an increase of at least 1.8% in the prevalence of atrial fibrillation,28 due to comorbidities, anticancer treatment, and cancer-associated factors such as dehydration and inflammation.6 Decisions about whether to initiate or maintain anticoagulation therapy, and which anticoagulant to choose, are complex because scores such as CHA2DS2-VASc, HAS-BLED and HEMORR2HAGES do not take into consideration the additional risks conferred by malignancy, BleeMACS being the only score that includes cancer as a risk factor.12,28–30 Also, no randomized trials have been conducted specifically assessing oral anticoagulants use in cancer patients with atrial fibrillation. However, secondary analyses of the novel oral anticoagulant (NOAC) trials in patients with or without a history of cancer or in those who developed cancer after enrollment showed that NOACs are at least as effective and safe as warfarin in this subgroup.31–33 On the other hand, the use of VKAs in the cancer population can be more challenging than NOACs, because of the former's more frequent interactions with chemotherapy, antibiotics, analgesics and other commonly prescribed drugs.34,35 Consequently, the International Society on Thrombosis and Haemostasis recommends that NOACs should be considered as the first choice in cancer patients, especially in those with a favorable prognosis (except for luminal gastrointestinal cancers or active gastrointestinal mucosal abnormalities).28

In our study, oral anticoagulation – mainly with VKAs – was independently associated with major bleeding events. However, we found no difference between the proportion of patients on VKA and NOAC therapy (9% vs. 8%, p=0.94) who suffered major bleeding. Hence, NOAC therapy appears to be as safe as VKAs in our population. Regarding parental anticoagulation, in our registry it is not specified for how long and at what dosage it was prescribed. We can assume that, since atrial fibrillation was also associated with major bleeding, cancer patients with this arrhythmia probably received more aggressive anticoagulation therapy, including parenteral anticoagulants. Decisions on triple therapy (anticoagulation and DAPT) in cancer patients with ACS and atrial fibrillation should be individualized and it should mostly be reserved for patients with good functional status and favorable prognosis, with low bleeding risk and without major risk factors (such as previous bleeding or anemia).6,34 Since we do not know the degree of anticoagulation of patients on parenteral therapy, we did not perform a subanalysis regarding triple therapy.

More patients in the cancer groups had the secondary combined endpoint, mainly with increased mortality. However, mortality risk and ischemic events were not significantly higher than in patients with similar characteristics after propensity score matching. As expected, in our population predictors of the secondary endpoints were STEMI, Killip class >I, reduced left ventricular ejection fraction and renal failure. Other identified predictors that merit particular consideration were the presence of thrombocytopenia, no prescription of neurohormonal blockers and antiplatelet therapy, and also no invasive procedures (ICA and PCI). A platelet count <100 000/mm3 is estimated to be present in 10-25% of patients with cancer.34 Thrombocytopenia was not an independent predictor of bleeding in our population, in line with a study of cancer patients with ACS and chronic thrombocytopenia,36 but it was consistently associated with mortality and ischemic events. In patients with thrombocytopenia, the risk of bleeding varies and may depend on the underlying cause of thrombocytopenia; clinical experience suggests that platelet function rather than platelet count is the determinant factor.36 Another study has postulated the ‘platelet paradox’ of coronary thrombosis in thrombocytopenic patients, in whom aspirin therapy was associated with significantly improved seven-day survival after ACS without a significant increase in bleeding complications.37 Previous studies also demonstrated that concerns about bleeding may prompt interventional cardiologists to defer invasive approaches or PCI and prescription of aspirin, more potent platelet inhibition with ticagrelor or prasugrel or DAPT, leading to higher mortality.3,15,17,26,38–41 On the other hand, when the invasive strategy is not postponed, cancer patients could have the same long-term cardiac mortality as the general population, although with higher noncardiac mortality due to cancer itself.20 Our results after propensity score matching are in agreement with these findings: the matched cancer population had similar rates of aspirin prescription, DAPT, ICA and PCI, and no difference in mortality was observed between patients with and without cancer.

Expert consensus recommends the use of stents, preferably newer-generation drug-eluting stents (which may have lower rates of stent thrombosis), avoiding bifurcation and overlapping stents in patients with cancer.42 Our data showed that in cancer patients undergoing PCI, drug-eluting stents were more often used, without significant differences in outcomes according to the type of stent. The use of intravascular ultrasound or optical coherence tomography is also recommended to ensure adequate stent expansion, apposition and lack of edge dissection,42 but this information was not retrieved in our registry. A minority of our cancer patients underwent balloon angioplasty only (5.8%), which may be preferred, for example, if PCI is necessary in patients awaiting cancer surgery.42

Cancer patients were also less likely to be prescribed optimal medical therapy, such as ACE inhibitors, beta-blockers or statins. In our study, patients who did not receive ACE inhibitors more often presented with hypotension and more severely impaired renal function, which may explain their worse short-term prognosis. However, long-term data in cancer patients show that optimal medical therapy in this population can reduce cardiovascular mortality at one year after ACS.6,18

Our findings indicate that all patients with a current or prior history of cancer should be treated as a high-risk group using an integrative approach, addressing both cardiac and noncardiac aspects.

The management of ACS in patients with cancer is particularly challenging due to higher bleeding and mortality risks. In our study, major bleeding events were mainly related to previous bleeding predisposition and atrial fibrillation, associated with anticoagulation strategy. On the other hand, in-hospital mortality, reinfarction and ischemic stroke were associated with lower use of ICA and antiplatelet and neurohormonal blocker therapy.

After propensity score matching in terms of baseline characteristics and use of invasive procedures, major bleeding, mortality and ischemic risks were not significantly higher in cancer patients. These findings indicate that the presence of cancer should not limit the effective and safe treatment of ACS, but rather lead to a particularly rigorous assessment of bleeding and thrombotic risk, regarding both pharmacological and interventional decisions.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors have no conflicts of interest to declare.