A retrospective, single-center study was performed at the Second Affiliated Hospital of Harbin Medical University, which was approved by the ethics committee of Harbin Medical University. The study procedures were conducted in compliance with the principles of the Declaration of Helsinki, and patient information was collected anonymously. All AMI patients from June 2016 to December 2019 were included in the study. Myocardial infarction (MI) was defined according to the fourth universal definition of MI (17). The population included in the final analysis consisted of 6,607 AMI patients. All detailed clinical data of those patients were collected from electronic medical records, including age, sex, type of malignancy, cardiovascular risk factors [smoking status, hypertension, hyperlipidemia, diabetes mellitus, and previous coronary heart disease (CHD)], treatment, and outcomes. During a 3-year follow-up period, patients were surveyed semi-annually via telephone about major adverse events using a standardized questionnaire.

Our primary outcome was defined as cardiovascular mortality during follow-up. Secondary outcomes included all-cause mortality, major adverse cardiovascular and cerebrovascular events (MACCE), non-fatal MI, non-fatal stroke, and revascularization. MACCE is composed of cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke.

For all statistical tests, a two-tailed P-value < 0.05 indicated statistical significance, and data analyses were performed using R version 3.6.2 software (R Institute Inc.). Continuous variables are presented as the means ± standard deviations (SDs) if normally distributed or presented as medians with interquartile ranges (IQRs) if non-normally distributed. discrete variables are presented as frequencies (percentages), and missing data were excluded from the summary statistic calculations. To evaluate the differences in baseline characteristics between unmatched groups, Student's t-test was used for nearly normally distributed continuous variables, the Wilcoxon rank-sum test was used for non-normally distributed continuous and ordinal discrete variables, and Categorical data have been compared using the χ² or Fisher's exact test. Furthermore, to make the two groups comparable with regard to the vast majority of baseline characteristics, pairwise matching was performed via a greedy matching algorithm to match each pair of reference patients and patients with cancer according to the following restrictions: (1) age within 1 year, (2) sex, (3) hyperlipidemia status, (4) smoking status, and (5) diabetes status. The control group allowed a variable number of reference matches and a maximum of 4 matches per patient with cancer. Except for unpaired patients, each patient pair was used once in the further analyses. Comparisons between reference patients and patients with cancer were tested via the same test for baseline characteristics and outcomes. To evaluate the incremental relative risk increase among subgroups in the heterogeneity analysis, models were fit with an indicator for any history of cancer and with another indicator for the subgroup. Forest plots were drawn to analyze the heterogeneity of the effect of coexisting cancer on the event risk between subgroups.

A total of 6,607 AMI patients were included between June 2016 and December 2019. Among those patients, 150 (2.3%) had been diagnosed with cancer. According to the order of frequencies, the most prevalent malignancies were lung (31, 20.7%), colorectum (21, 14.0%), stomach (19, 12.7%), and breast (15, 10.0%) cancers (Supplementary Table S1).

The characteristics of the overall cohort and matched cohort are summarized in Table 1. Before matching, the group of patients with cancer was older (70.3 ± 10.0 vs. 63.9 ± 11.5 years, P < 0.001) and had higher proportions of patients with hyperlipidemia (31.3 vs. 22.7%, P = 0.018) and diabetes (38.3 vs. 24.2%, P < 0.001). The group of patients with cancer had a lower proportion of current smokers (31.1 vs. 48.7%, P < 0.001) but higher proportions of patients with comorbidities. Furthermore, the group of patients with cancer had higher proportions of patients with previous CHD (44.9 vs. 25.7%, P < 0.001), previous MI (21.8 vs. 11.0%, P < 0.001) and previous PCI (16.3 vs. 7.3%, P < 0.001) than the group without cancer. During hospitalization, patients with cancer tended to receive clopidogrel (52.0 vs. 40.0%, P = 0.004) rather than ticagrelor (45.6 vs. 58.2%, P = 0.003) given an aspirin background. In addition, matching was possible for 542 pairs of reference patients and patients with cancer, and those patients constituted our matched study groups. After controlling for these heterogeneous covariates, such as age, sex, diabetes, smoking habits, and hyperlipidemia, the baseline characteristics were similar between the groups after matching with the exception of higher proportions of patients with previous CHD (44.9 vs. 26.8%, P < 0.001), previous MI (21.8 vs. 11.1%, P = 0.001) and previous PCI (16.3 vs. 5.9%, P < 0.001) in the group of patients with cancer than in the matched controls. Moreover, patients with cancer were less likely to undergo in-hospital PCI (61.3 vs. 70.0%, P = 0.055).

With regard to the long-term outcomes, patients with cancer had a significantly higher cumulative incidence of all-cause mortality (22.7 vs. 9.8%; adjusted HR, 2.40; 95% CI, 1.52–3.79; P < 0.001) (Supplementary Figure S1) and cardiovascular mortality (14.0 vs. 8.3%; adjusted HR, 1.934; 95% CI, 1.11–3.39; P = 0.021) (Figure 1; Supplementary Table S2). MACCE were also significantly higher in the patients with cancer than in the matched non-cancer group (16.0 vs. 10.3%; adjusted HR, 1.98; 95% CI, 1.21–3.26; P = 0.007). Moreover, there was no significant difference in MI (2.7 vs. 1.7%; adjusted HR, 1.64; 95% CI, 0.50–5.41; P = 0.419), stroke (0.7 vs. 0.9%; adjusted HR, 0.84; 95% CI, 0.10–7.34; P = 0.876), and revascularization (1.3 vs. 4.6%; adjusted HR, 0.259; 95% CI, 0.061–1.097; P = 0.067) between patients with or without cancer. Cardiovascular mortality tended to be similar across all pre-specified subgroups (Figure 2), as was all-cause mortality and MACCE (Supplementary Figures S2, S3).

Among 150 patients with cancer, 52 had a historical cancer diagnosis beyond 5 years before AMI, 59 had a historical cancer diagnosis within 5 years before AMI, and the other 39 had a current cancer diagnosis after AMI. The incidences of all-cause mortality, cardiovascular mortality and MACCE were significantly higher among patients with a historical cancer diagnosis within 5 years than among those without cancer (adjusted HR, 3.38; 95% CI, 1.88–6.04; P < 0.001; adjusted HR, 2.59; 95% CI, 1.25–5.35; P = 0.010; and adjusted HR, 2.66; 95% CI, 1.39–5.11; P = 0.003, respectively) (Table 2). A similar pattern was observed for all-cause mortality among patients with a current cancer diagnosis (adjusted HR, 2.71; 95% CI, 1.25–5.88; P = 0.012).

We found that patients with cancer are less likely to undergo PCI treatment during hospitalization than those without cancer, and they were also less likely to undergo revascularization during follow-up. According to previous data, patients with active cancer have ~2- and 3-fold higher risks of 90 days for readmission with AMI or major bleeding after PCI, respectively, than patients without cancer (18). Thus, clinicians are often wary of performing invasive therapies in patients with cancer. However, data from large retrospective studies showed that PCI results in significantly lower risks of in-hospital all-cause mortality and MACCE than conservative treatment, irrespective of whether the patient had a cancer diagnosis, and PCI did not increase the risk of in-hospital complications, including massive bleeding and stroke (19). To date, there has been no large randomized trial to assess the benefits and risks of invasive and conservative approaches to treating AMI in patients with cancer, and such patients are often excluded from clinical trials. The current guidelines recommend that percutaneous revascularization should be considered even in cancer patients with an expected survival duration of <1 year (20). Balloon angioplasty without stents are recommended to limit the duration of antiplatelet therapy. If stents need to be used, those with fast reendothelialization rates may be a better choice.

The coexistence of high risks of ischemia and major bleeding presents a challenge for clinicians when treating AMI patients with cancer with regard to antiplatelet therapy. When faced with this dilemma, clinicians prefer conservative approaches with regard to aspirin-based DAPT. A less potent P2Y12 inhibitor, namely, clopidogrel rather than ticagrelor, was administered to AMI patients with cancer, but there is a lack of reliable evidence to confirm the greater benefits of clopidogrel among such high-risk patients.

A previous study that included 6,563,255 AMI patients revealed that patients with cancer, irrespective of the cancer type, had higher risks of in-hospital mortality, MACCE, and stroke than those without cancer (21). Inflammation plays a vital role in the progression of both cancer and atherosclerotic lesions (including CHD) (22). Although the mechanism underlying this association is unclear, we propose that local malignancies might increase vascular wall inflammation by releasing inflammatory cytokines and that this circulatory inflammation might subsequently lead to progressive coronary atherosclerosis. In addition, cardiotoxicity can be a major complication of cancer treatment, radiotherapy is recognized as a cardiovascular risk factor among patients with cancer, and many anticancer drugs (anthracyclines, vinca alkaloid anti-metabolites, and biologics) are known to be closely associated with acute early and late cardiovascular adverse events. Perhaps because of the overlap of common risk factors for cancer and CHD and the susceptibility to atherosclerosis caused by oncology treatments (such as radiation therapy or tyrosine kinase inhibitors), patients with cancer tend to exhibit a relatively higher cardiovascular risk. In particular, there was no significant difference in cardiovascular mortality and MACCE for 1 year, but we found that there was no significant difference in cardiovascular mortality and MACCE for 1 year (Supplementary Table S3), and the 3-year incidences of all-cause mortality, cardiovascular mortality and MACCE were significantly higher among patients with cancer than among those without cancer (Supplementary Table S2). These problems highlight the fact that cardiovascular diseases become more important during the long-term survival of patients with cancer. Advances in screening, big data, targeted and immune therapies, and significant new knowledge of cancer biology are changing the prevention, detection, diagnosis, treatment and survival of cancer. However, the current treatments are still mostly based on extrapolation from non-cancer patient data, and there remain some gaps in achieving the goal of personalized treatment for AMI patients with cancer.

Furthermore, subgroup analysis was performed according to the time between the diagnosis of cancer and the occurrence of AMI. The results showed that the incidences of all-cause death, cardiovascular death and MACCE in the group with a historical cancer diagnosis within 5 years were significantly higher than in those without cancer, and the risks in that subgroup were the highest among all subgroups. This connection is not accidental, and a large-scale study from Sweden also found that patients with cancer had the highest risk of CHD in the first 6 months after diagnosis (23). Another previous study reported similar results: the risks of in-hospital mortality and MACCEs were higher by at least 50% among AMI patients with a current cancer diagnosis than among those without cancer, whereas they were not higher among patients with a historical cancer diagnosis (21). Our findings also underscore the importance of vigilance in cardiovascular risk monitoring after cancer treatment. It is critical to continue assessing the risk of potential cardiovascular events among patients with cancer, and future randomized trials are needed to evaluate the effectiveness of such surveillance.

(a) We acknowledge all limitations inherent to a retrospective, single-center study, which restrict the generalization of our findings and the inference of causality. (b) The overall cancer population was relatively small, and the subgroups related to cardiovascular safety concerns were potentially underpowered. In addition, the patients with cancer were a heterogeneous population with different cancer types and stages, and the sample size was too small to evaluate each cancer type separately. (c) Although the data for AMI patients were abundant, the lack of complete cancer history and cancer types may be considered a limitation of this study. The missing data on cancer metastasis, stages, and cancer treatment limits the further understanding of the differences in outcomes between AMI patients with cancer and those without cancer.