2021 年 85 巻 6 号 p. 837-846
Background: The aim of this study is to evaluate clinical outcomes after percutaneous coronary intervention (PCI) in patients with cancer.
Methods and Results: Cancer screening was recommended before PCI in consecutive 1,303 patients who underwent their first PCI. By using cancer screening, cancer was diagnosed in 29 patients (2.2%). In total, 185 patients had present or a history of cancer. Patients with cancer more often suffered from non-cardiac death than those without (4.4% vs. 1.5%, P=0.006), and patients with cancer requiring ongoing therapy (n=18) more often suffered from major bleeding compared with those with recently (≤12 months) diagnosed cancer who do not have ongoing therapy (n=59) (16.7% vs. 3.4%, P=0.049). During the 1-year follow up, 25 patients (2.0%) were diagnosed as having cancer, in which 48.0% of bleeding events led to a cancer diagnosis. Patients with high bleeding risk according to the Academic Research Consortium for high bleeding risk (ARC-HBR) were associated with a greater 1-year major bleeding risk than those without high bleeding risk in patients with (7.9% vs. 0.0%, P=0.02) and without cancer (7.1% vs. 2.5%, P<0.001), respectively.
Conclusions: Cancer was diagnosed in 2.2% of 1,303 unselected patients before PCI by cancer screening and in 2.0% within 1-year after PCI. Cancer was associated with a greater risk of non-cardiac death, whereas ongoing active cancer was associated with greater risk of major bleeding. ARC-HBR criteria successfully identified high-bleeding risk patients, irrespective of the presence or absence of cancer.
Cardiovascular diseases and cancer are predominant causes of death accounting for 31.4% and 15.8% of mortality, respectively.1 Patients with a history of cardiovascular disease and cancer survivors are reported to be increasing, especially in the aged society.2,3 Due to the recent advances in cancer treatment leading to longer life expectancy after cancer diagnosis, only a few cancer survivors suffered from cardiovascular disease,4 whereas primary and secondary preventions for cardiovascular disease might reduce the risk of cardiovascular mortality and increase the potential risk of malignancy. Furthermore, cardiovascular risk factors, age and smoking habits in particular, are also risk factors for cancer,5–7 whereas radiation and chemotherapy for cancer might accelerate atherosclerosis progression.8
Editorial p 847
Bleeding after percutaneous coronary intervention (PCI) was reported to be an independent predictor for late adverse events,9 and bleeding is also a major complication for malignancy.10 Several bleeding risk scores have been proposed and validated as useful tools to estimate the bleeding risks after PCI,11–13 although it remained unclear whether these risk stratification tools were applicable for cancer patients. Recently, consensus on high bleeding risk has been published by the Academic Research Consortium for High Bleeding Risk (ARC-HBR);14 however, to the best of our knowledge, the criteria have not been entirely validated, especially for cancer patients.
In light of that described above, we sought to evaluate the clinical outcomes in patients with or without cancer who underwent their first PCI, and to assess the predictive value of risk stratification tools for bleeding in cancer and non-cancer patients.
From 24 February 2016 to 4 December 2017, a consecutive total of 1,303 patients who had no prior history of any coronary revascularization underwent their first PCI at Kokura Memorial Hospital. According to the institutional policy, patients were encouraged to undergo cancer screening that was aimed at reducing the risk of bleeding prior to the PCI procedure, in cases of patients’ agreement and not when they required emergency procedures. Fecal immunochemical test (FIT) and upper gastrointestinal endoscopy (UGE) were recommended in for all patients, whereas lower gastrointestinal endoscopy (LGE) was recommended in cases of positive FITs. If a cancer was identified prior to the PCI procedure, priorities of cases for PCI or treatments of cancer were decided on consensus among cardiologists and oncologists.
To evaluate the effect of cancer on 1-year clinical outcomes, we divided patients into 2 groups: cancer group and non-cancer group, as shown in Figure 1. The non-cancer group was defined as those patients who have not been diagnosed as cancer. As an exploratory analysis, the cancer group were subdivided into an active cancer group and a non-active cancer group; and for the active cancer group, we further divided the patients into the ongoing active cancer group (under ongoing active cancer treatment such as surgery, radiotherapy, chemotherapy, or immunotherapy) and the recently active cancer group (diagnosed within 12 months before PCI). The specific time point of 12 months was used, according to the ARC-HBR criteria.14
Study flow chart.
The study protocol was approved by the institutional review board of Kokura Memorial Hospital, and the study was performed in accordance with the Declaration of Helsinki. Due to retrospective enrollment, written informed consent was waived for patients with an opt-out approach.
Clinical Outcomes and Risk ScoresWe obtained clinical follow-up information from the medical records, patients’ referral document and/or telephone interviews with patients or their families. The primary endpoint of the study was net adverse clinical and cerebral events (NACCE) at 1 year. NACCE was defined as a composite of all-cause death, myocardial infarction (MI), stroke, or major bleeding. The secondary endpoints included major bleeding and a thrombotic composite of MI or definite stent thrombosis. According to the Bleeding Academic Research Consortium (BARC) definition, major bleeding was defined as BARC 3 and 5, whereas minor bleeding was defined as BARC 2.15 MI was defined by the fourth universal definition of MI.16 We also assessed all-cause death, MI, stroke, coronary artery bypass grafting, target vessel revascularization, target lesion revascularization, definite stent thrombosis, minor bleeding, and new diagnosis of cancer after the PCI procedure. The cause of death was classified as cardiovascular or non-cardiovascular according to the Academic Research Consortium (ARC) criteria.17 Target vessel revascularization, target lesion revascularization and definite stent thrombosis were defined according to the ARC criteria.17
High bleeding risk was regarded as being present for at least 1 major or 2 minor criteria of the ARC-HBR criteria,14 PRECISE DAPT score ≥25,12 PARIS bleeding score ≥8,11 or Coronary REvascularization Demonstrating Outcome study in Kyoto (CREDO Kyoto) score ≥3.13
All baseline variables and clinical outcomes were adjudicated by the 2 assessors who reviewed the original source of documents.
Statistical AnalysisContinuous variables are expressed as mean±standard deviation or median (interquartile range) and were compared using the Student’s t-test or Wilcoxon rank sum test. Categorical variables are presented as values and percentages and were compared using Fisher’s exact test. Kaplan-Meier curves were used to estimate the cumulative incidences of clinical events and the logrank test was used to assess betweeen group differences. Cox proportional hazard models were constructed to evaluate the effect of high bleeding risk according to the risk stratification tools on clinical outcomes and its interaction with the presence or absense of cancer. A 2-sided P value of <0.05 was considered to be significant for all tests. All analyses were performed using R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria).
The study flow chart is shown in Figure 1. Among a total of 1,303 consecutive patients, a FIT was used to examine 560 patients (42.9%), a UGE was performed in 315 patients (24.1%) and a LGE was performed in 147 patients (11.2%). Compared with stable patients (n=809, 62.1%), patients presenting with acute coronary syndrome (n=494, 37.9%) less often underwent cancer screening prior to PCI (FIT, 11.1% vs. 62.4%, P<0.001; UGE, 4.5% vs. 36.2%, P<0.001; LGE, 2.2% vs. 16.8%, P<0.001). In terms of cancer screening prior to PCI, we newly diagnosed cancer in 29 patients (2.2%), of which 25 patients were treated without requiring additional cancer treatment (e.g., complete surgical resection with no evidence for residual disease) and the remaining 4 patients underwent active cancer treatment at the time of PCI. In patients who underwent cancer screening and were not diagnosed with cancer, we treated 7 patients with gastric ulcers and 46 patients with colon polyps prior to PCI to minimize the risk of bleeding.
After cancer screening, 1,118 (85.8%) patients had not been diagnosed with cancer prior to PCI (non-cancer group). In 185 patients (14.2%) in the cancer group, 77 patients (5.9%) were classified as the active cancer group, including 18 patients (1.4%) who were undergoing active cancer treatment (ongoing active cancer group) and 59 patients (4.5%) who had a history of cancer ≤12 months prior to PCI without ongoing active cancer treatment (recently active cancer group), whereas the remaining 108 patients (8.3%) had a history of cancer >12 months prior to PCI (non-active cancer group). The distribution of cancer types is listed in Supplementary Table 1. Colorectal cancer was the most common (30.8%) type of cancer, followed by gastric (13.5%), prostate (13.0%), and breast cancer (8.1%).
Baseline CharacteristicsBaseline characteristics are summarized in Table 1. Compared with patients in the non-cancer group, those in the cancer group were older (75.5±9.1 vs. 70.5±11.8, P<0.001) and more often a past smoker (P=0.03) and likely to have liver cirrhosis (4.3% vs. 0.1%, P<0.001) and anemia (severe: 25.9% vs. 15.0%, P<0.001; mild 13.5% vs. 7.6%, P=0.01). Regarding procedural characteristics, patients in the cancer group less often had multivessel disease (19.5% vs. 29.0%, P=0.009) and more often underwent bare metal stent implantation (4.9% vs. 2.0%, P=0.03). No significant difference between the groups was observed in terms of medication status at discharge. Most of patients received dual antiplatelet therapy, whereas 74.9% received proton pump inhibitors.
| Entire cohort (n=1,303) |
Cancer (n=185) |
Non-cancer (n=1,118) |
P values | |
|---|---|---|---|---|
| Patient characteristics | ||||
| Age, years | 71.2±11.6 | 75.5±9.1 | 70.5±11.8 | <0.001 |
| ≥75 years | 546 (41.9) | 103 (55.7) | 443 (39.6) | <0.001 |
| Male | 880 (67.5) | 128 (69.2) | 752 (67.3) | 0.67 |
| Body mass index | 23.6±3.7 | 23.1±4.7 | 23.7±3.5 | 0.11 |
| Acute coronary syndrome at presentation | 494 (37.9) | 50 (27.0) | 444 (39.7) | 0.001 |
| Prior myocardial infarction | 39 (3.0) | 4 (2.2) | 35 (3.1) | 0.64 |
| Hypertension | 1,038 (79.7) | 156 (84.3) | 882 (78.9) | 0.09 |
| Diabetes mellitus | 502 (38.5) | 74 (40.0) | 428 (38.3) | 0.68 |
| On insulin therapy | 74 (5.7) | 13 (7.0) | 61 (5.5) | 0.39 |
| Hyperlipidemia | 826 (63.4) | 120 (64.9) | 706 (63.1) | 0.68 |
| Smoking status | 0.03 | |||
| Current smoker | 213 (16.3) | 21 (11.4) | 192 (17.2) | |
| Past smoker | 560 (43.0) | 94 (50.8) | 466 (41.7) | |
| Renal failure | ||||
| eGFR <30 mL/min/1.73 m2 or dialysis | 133 (10.2) | 20 (10.8) | 113 (10.1) | 0.79 |
| eGFR 30–59 mL/min/1.73 m2 | 419 (32.2) | 66 (35.7) | 353 (31.6) | 0.27 |
| Chronic obstructive pulmonary disease | 25 (1.9) | 5 (2.7) | 20 (1.8) | 0.38 |
| Peripheral artery disease | 151 (11.6) | 26 (14.1) | 125 (11.2) | 0.26 |
| Atrial fibrillation | 138 (10.6) | 25 (13.5) | 113 (10.1) | 0.16 |
| Prior heart failure | 110 (8.4) | 21 (11.4) | 89 (8.0) | 0.15 |
| Liver cirrhosis with portal hypertension | 9 (0.7) | 8 (4.3) | 1 (0.1) | <0.001 |
| Prior spontaneous major bleeding† | ||||
| In the past 6 months or at any time if recurrent | 10 (0.8) | 4 (2.2) | 6 (0.5) | 0.04 |
| In the past 6–12 months and not recurrent | 1 (0.1) | 0 (0) | 1 (0.1) | 1.00 |
| Prior spontaneous intracranial hemorrhage | 20 (1.5) | 4 (2.2) | 16 (1.4) | 0.51 |
| Prior ischemic stroke | ||||
| Moderate or severe stroke within 6 months‡ | 8 (0.6) | 2 (1.1) | 6 (0.5) | 0.32 |
| Mild stroke or any stroke >6 months | 116 (8.9) | 18 (9.7) | 98 (8.8) | 0.68 |
| Recent major surgery or major trauma within 30 days | 4 (0.3) | 3 (1.6) | 1 (0.1) | 0.01 |
| Non-deferrable major surgery on dual antiplatelet therapy | 1 (0.1) | 0 (0) | 1 (0.1) | 1.00 |
| White blood cell count (×106/L) | 72.1±29.6 | 67.7±36.9 | 72.8±28.2 | 0.07 |
| Anemia | ||||
| Hemoglobin <11 g/dL | 216 (16.6) | 48 (25.9) | 168 (15.0) | <0.001 |
| Hemoglobin men: 11–12.9 g/dL, women: 11–11.9 g/dL | 110 (8.4) | 25 (13.5) | 85 (7.6) | 0.01 |
| Thrombocytopenia (platelet <100×109/L) | 14 (1.1) | 4 (2.2) | 10 (0.9) | 0.12 |
| ARC-HBR high bleeding risk | 650 (49.9) | 144 (77.8) | 506 (45.3) | <0.001 |
| PRECISE DAPT score | 24 (17–35) | 28 (21–40) | 23 (16–34) | <0.001 |
| ≥25 | 603 (46.3) | 108 (58.4) | 495 (44.3) | <0.001 |
| PARIS bleeding score | 5 (4–8) | 6 (5–8) | 5 (4–7) | <0.001 |
| ≥8 | 335 (25.7) | 61 (33.0) | 274 (24.5) | 0.02 |
| CREDO Kyoto score | 0 (0–1) | 1 (1–3) | 0 (0–1) | <0.001 |
| ≥3 | 145 (11.1) | 51 (27.6) | 94 (8.4) | <0.001 |
| Procedural characteristics | ||||
| Multivessel disease | 360 (27.6) | 36 (19.5) | 324 (29.0) | 0.008 |
| Left main disease | 126 (9.7) | 18 (9.7) | 108 (9.7) | 1.00 |
| Total stent length, mm | 44.8±30.4 | 41.5±27.5 | 45.4±30.8 | 0.09 |
| Minimum stent diameter, mm | 2.9±0.5 | 2.8±0.5 | 2.9±0.5 | 0.19 |
| Newer generation drug-eluting stent | 1,192 (91.5) | 165 (89.2) | 1,027 (91.9) | 0.25 |
| First generation drug-eluting stent | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Bare metal stent | 31 (2.4) | 9 (4.9) | 22 (2.0) | 0.03 |
| Drug coated stent | 9 (0.7) | 2 (1.1) | 7 (0.6) | 0.37 |
| Medication at discharge | ||||
| P2Y12 inhibitors | 1,290 (99.0) | 182 (98.4) | 1,108 (99.1) | 0.41 |
| Aspirin | 1,284 (98.5) | 180 (97.3) | 1,104 (98.7) | 0.17 |
| Oral anticoagulant | 166 (12.7) | 32 (17.3) | 134 (12.0) | 0.06 |
| Non-steroidal anti-inflammatory drugs or steroids | 74 (5.7) | 14 (7.6) | 60 (5.4) | 0.23 |
| Proton pump inhibitor | 976 (74.9) | 132 (71.4) | 844 (75.5) | 0.23 |
| H2 blocker | 76 (5.8) | 11 (5.9) | 65 (5.8) | 0.87 |
| Statin | 611 (46.9) | 86 (46.5) | 525 (47.0) | 0.94 |
Continuous variables are presented as mean±standard deviation or median (interquartile range). Categorical variables are given as n (%). eGFR, estimated glomerular filtration rate. Chronic bleeding diathesis, prior traumatic intracranial hemorrhage within 12 months, and brain arteriovenous malformation were not regarded as being present in any patients. Values were missing for complete blood count in 5 patients, for eGFR in 5 patients, for PRECISE DAPT score in 7 patients, for PARIS bleeding score in 7 patients, and for CREDO Kyoto score in 6 patients. High bleeding risk was regarded as absent, if risk scores were not calculated. †Bleeding requiring hospitalization or transfusion. ‡National Institutes of Health Stroke Scale score ≥5.
Among patients in the cancer group, patients in the active cancer group tended to be younger (73.9±10.0 years vs. 76.6±8.3 years, P=0.06) and were more likely to be male (77.9% vs. 63.0%, P=0.04) as compared with those in the non-active cancer group (Supplementary Table 2). Performance status, cancer stage, and cancer therapy for active cancers are summarized in Supplementary Table 3 and Supplementary Figure.
Patients in the ongoing active cancer group had significantly lower body mass index (21.3±3.3 vs. 24.2±6.2, P=0.01) and more likely to be anemic (hemoglobin <11 g/dL: 44.4% vs. 15.3%, P=0.02) as compared with those in the recently active cancer group (Supplementary Table 4).
Clinical OutcomesClinical follow-up information was obtained for 96.9% of patients at 1 year. The primary endpoint of NACCE at 1 year was more often observed in patients in the cancer group than in those in the non-cancer group (13.6% vs. 8.8%, P=0.04, Table 2, Figure 2A), mainly driven by tendencies toward greater risks of all-cause death (7.1% vs. 4.1%, P=0.07) and stroke (2.3% vs. 0.8%, P=0.08). Patients in the cancer group more often suffered from non-cardiac death (4.4% vs. 1.5%, P=0.006), whereas the cumulative incidence of cardiac death was comparable between the groups (2.8% vs. 2.5%, P=0.85, Supplementary Table 5). No significant between the group differences were observed in terms of major bleeding (6.1% vs. 4.5%, P=0.37, Figure 2B) and a composite of MI or definite stent thrombosis (0.0% vs. 0.7%, P=0.25, Figure 2C). No patients in the cancer group and 4 patients in the non-cancer group suffered from definite thrombosis (P=0.42). The 1-year rates of discontinuation of at least 1 antiplatelet therapy were comparable between patients with vs. without cancer (50.5% vs. 45.5%, P=0.15, Figure 2D).
| Entire cohort (n=1,303) | Cancer (n=185) | Non-cancer (n=1,118) | Logrank P values |
||||
|---|---|---|---|---|---|---|---|
| 6 months | 12 months | 6 months | 12 months | 6 months | 12 months | ||
| Net adverse clinical and cerebral events |
95 (7.3) | 122 (9.4) | 19 (10.3) | 25 (13.6) | 76 (6.8) | 97 (8.8) | 0.04 |
| Thrombotic composite events | 15 (1.2) | 21 (1.7) | 2 (1.1) | 4 (2.3) | 13 (1.2) | 17 (1.6) | 0.51 |
| All-cause death | 39 (3.0) | 58 (4.5) | 10 (5.4) | 13 (7.1) | 29 (2.6) | 45 (4.1) | 0.07 |
| Cardiac death | 24 (1.9) | 33 (2.6) | 4 (2.2) | 5 (2.8) | 20 (1.8) | 28 (2.5) | 0.85 |
| Non-cardiac death | 15 (1.2) | 25 (2.0) | 6 (3.3) | 8 (4.4) | 9 (0.8) | 17 (1.6) | 0.009 |
| Myocardial infarction | 6 (0.5) | 8 (0.6) | 0 (0.0) | 0 (0.0) | 6 (0.5) | 8 (0.7) | 0.25 |
| Stroke | 14 (1.1) | 19 (1.5) | 2 (1.1) | 4 (2.3) | 12 (1.1) | 15 (1.4) | 0.38 |
| Ischemic stroke | 9 (0.7) | 13 (1.0) | 2 (1.1) | 4 (2.3) | 7 (0.6) | 9 (0.8) | 0.08 |
| Hemorrhagic stroke | 5 (0.4) | 6 (0.5) | 0 (0.0) | 0 (0.0) | 5 (0.4) | 6 (0.5) | 0.32 |
| Coronary artery bypass grafting | 5 (0.4) | 12 (1.0) | 0 (0.0) | 0 (0.0) | 5 (0.5) | 12 (1.1) | 0.16 |
| Target lesion revascularization | 5 (0.4) | 26 (2.1) | 0 (0.0) | 4 (2.3) | 5 (0.5) | 22 (2.1) | 0.82 |
| Target vessel revascularization | 20 (1.6) | 56 (4.5) | 4 (2.2) | 10 (5.7) | 16 (1.5) | 46 (4.3) | 0.38 |
| Definite stent thrombosis | 3 (0.2) | 4 (0.3) | 0 (0) | 0 (0) | 3 (0.3) | 4 (0.4) | 0.42 |
| Bleeding | |||||||
| BARC 5 | 3 (0.2) | 5 (0.4) | 0 (0.0) | 1 (0.6) | 3 (0.3) | 4 (0.4) | 0.70 |
| BARC 4 | 4 (0.3) | 9 (0.7) | 0 (0.0) | 0 (0.0) | 4 (0.4) | 9 (0.8) | 0.23 |
| BARC 3 | 49 (3.8) | 56 (4.4) | 8 (4.4) | 10 (5.5) | 41 (3.7) | 46 (4.2) | 0.41 |
| BARC 2 | 64 (5.0) | 82 (6.5) | 11 (6.1) | 14 (7.9) | 53 (4.8) | 68 (6.2) | 0.41 |
| Major bleeding (BARC 3 or 5) | 52 (4.0) | 61 (4.8) | 8 (4.4) | 11 (6.1) | 44 (4.0) | 50 (4.5) | 0.37 |
| Periprocedural bleeding | 7 (0.5) | 7 (0.5) | 1 (0.5) | 1 (0.5) | 6 (0.5) | 6 (0.5) | 0.99 |
| Non-periprocedural bleeding | 45 (3.5) | 54 (4.2) | 7 (3.8) | 10 (5.6) | 38 (3.5) | 44 (4.0) | 0.34 |
| Malignancy related | 5 (0.4) | 7 (0.6) | 3 (1.7) | 4 (2.3) | 2 (0.2) | 3 (0.3) | <0.001 |
| Iatrogenic | 21 (1.6) | 24 (1.9) | 4 (2.2) | 5 (2.8) | 17 (1.5) | 19 (1.7) | 0.34 |
| Spontaneous | 29 (2.3) | 35 (2.8) | 4 (2.2) | 6 (3.4) | 25 (2.3) | 29 (2.7) | 0.59 |
| Trauma | 2 (0.2) | 2 (0.2) | 0 (0.0) | 0 (0.0) | 2 (0.2) | 2 (0.2) | 0.57 |
| Minor bleeding (BARC 2, 3, or 5) |
109 (8.4) | 136 (10.6) | 18 (9.8) | 24 (13.3) | 91 (8.2) | 112 (10.2) | 0.21 |
| Periprocedural bleeding | 32 (2.5) | 32 (2.5) | 6 (3.2) | 6 (3.2) | 26 (2.3) | 26 (2.3) | 0.45 |
| Non-periprocedural bleeding | 77 (6.6) | 104 (8.8) | 12 (6.9) | 18 (10.5) | 65 (6.1) | 86 (8.1) | 0.32 |
| Malignancy related | 9 (0.7) | 13 (1.1) | 4 (2.4) | 5 (3.0) | 5 (0.5) | 8 (0.8) | 0.01 |
| Iatrogenic | 57 (4.5) | 68 (5.4) | 9 (5.0) | 10 (5.6) | 48 (4.4) | 58 (5.4) | 0.87 |
| Spontaneous | 49 (4.0) | 63 (5.2) | 9 (5.2) | 14 (8.3) | 40 (3.8) | 49 (4.7) | 0.056 |
| Trauma | 3 (0.2) | 5 (0.4) | 0 (0.0) | 0 (0.0) | 3 (0.3) | 5 (0.5) | 0.37 |
| Newly diagnosed malignancy after PCI |
13 (1.0) | 25 (2.0) | 2 (1.1) | 6 (3.5) | 11 (1.0) | 19 (1.8) | 0.15 |
Data are presented as number of patients with event (cumulative incidence). BARC, Bleeding Academic Research Consortium; PCI, percutaneous coronary intervention.
Cumulative incidences of (A) NACCE, (B) major bleeding, (C) myocardial infarction or definite stent thrombosis, and (D) discontinuation of ≥1 antiplatelet therapy between patients with vs. without cancer. NACCE, net adverse clinical and cerebral events.
Between the active vs. non-active cancer group, no significant difference was observed with regards to any clinical endpoints measured in the present study (Supplementary Table 6). However, patients in the ongoing active cancer group more often suffered from NACCE (27.8% vs. 6.8%, P=0.01) compared with those with recently active cancer, mainly driven by the greater risk of major bleeding (16.7% vs. 3.4%, P=0.049, Supplementary Table 7). All of the major bleeding in the ongoing active cancer group was gastrointestinal cancer-related bleeding.
Within 1-year, a total of 25 patients (2.0%) were newly diagnosed as having cancer. Cancer type included colorectal in 8 patients, gastric in 5 patients, bladder in 4 patients, lung in 4 patients, breast in 1 patient, thyroid in 1 patient, hepatocellular in 1 patient and gallbladder Figure 1 in 1 patient (Supplementary Table 1). In 25 patients with newly diagnosed cancer, 13 patients (52.0%) suffered from any bleeding, including major bleeding in 7 patients (28.0%) and minor bleeding in 8 patients (32.0%). Although bleeding events led to a new diagnosis of cancer in 12 (48.0%) patients (colon cancer in 5 patients, bladder cancer in 4 patients, and gastric cancer in 3 patients) during the 1-year follow up (Supplementary Table 8), all of the 13 patients who were newly diagnosed with gastrointestinal cancer after PCI had not undergone cancer screening before their PCI.
Risk ScoresHigh bleeding risk was more prevalent in patients in the cancer group than those in the non-cancer group, according to the ARC-HBR criteria (at least 1 major or 2 minor criteria, 77.8% vs. 45.3%, P<0.001), PRECISE DAPT score (≥25, 58.4% vs. 44.3%, P<0.001), PARIS bleeding score (≥ 8, 33.0% vs. 24.5%, P=0.02) and CREDO Kyoto score (≥3, 27.6% vs. 8.4%, P<0.001). Compared with patients without a high bleeding risk, those with a high bleeding risk were associated with greater risks of the primary endpoint of NACCE without significant interactions with cancer, according to the ARC-HBR, PRECISE DAPT score, PARIS bleeding score, and CREDO Kyoto bleeding risk score (Figure 3).
One-year cumulative incidences of NACCE, major bleeding, and myocardial infarction or definite stent thrombosis in patients with vs. without high bleeding risk according to (A) ARC-HBR, (B) PRECISE DAPT Score, (C) PARIS bleeding score, and (D) CREDO Kyoto bleeding score. NACCE, net adverse clinical and cerebral events; ARC-HBR, Academic Research Consortium for High Bleeding Risk; CREDO Kyoto, Coronary REvascularization Demonstrating Outcome study in Kyoto.
High bleeding risk according to ARC-HBR criteria was associated with greater 1-year bleeding risk regardless of the presence or absence of cancer (Cancer group: 7.9% vs. 0.0%, P=0.02; Non-cancer group: 7.1% vs. 2.5%, P<0.001; P for interaction 0.15), whereas 1-year cumulative incidence of major bleeding in patients who were classified as having a high bleeding risk according to ARC-HBR criteria, PRECISE DAPT score, PARIS bleeding score, and CREDO Kyoto bleeding score exceeded ≥4% in patients with or without cancer.
No patients in the cancer group suffered from a thrombotic composite of MI or definite stent thrombosis. In patients in the non-cancer group, no significant differences in terms of thrombotic composite events were observed between patients with vs. without high bleeding risk assessed by ARC-HBR criteria (0.8% vs. 0.6%, P=0.70), PRECISE DAPT score (1.0% vs. 0.5%, P=0.27), and PARIS bleeding score (1.2% vs. 0.6%, P=0.36), whereas patients with high bleeding risk according to the CREDO-Kyoto bleeding score more often suffered from a composite of MI or stent thrombosis (3.5% vs. 0.5%, P=0.02).
Regarding major and minor criteria of the ARC-HBR, chronic bleeding diathesis, prior traumatic intracranial hemorrhage within the past 12 months, and presence of a brain arteriovenous malformation were not regarded as present in our cohort (Table 3). Among major or minor criteria, severe renal impairment (12.2% vs. 4.0%, P<0.001), severe anemia (10.8% vs. 3.6%, P<0.001), thrombocytopenia (28.6% vs. 4.5%, P<0.001), recent moderate or severe stroke (25.0% vs. 4.6%, P=0.006), recent major surgery or trauma (25.0% vs. 4.7%, P=0.04), age ≥75 years (6.8% vs. 3.3%, P=0.004), mild anemia (7.4% vs. 3.2%, P=0.03), and long-term use of non-steroidal anti-inflammatory drugs or steroids (9.7% vs. 4.5%, P=0.04) were significantly associated with 1-year major bleeding risk, and their 1-year major bleeding risks exceeded ≥4%. Other remaining criteria were also associated with major bleeding risk ≥4% at 1-year, except for liver cirrhosis with portal hypertension (n=9), non-deferrable major surgery (n=1), and non-recent prior spontaneous major bleeding (n=1).
| Present | Absent | Logrank P values |
|||
|---|---|---|---|---|---|
| N (%) | No. of events (%) |
N (%) | No. of events (%) |
||
| Major criteria | |||||
| Anticipated use of long-term oral anticoagulation | 166 (12.7) | 10 (6.2) | 1,137 (87.3) | 51 (4.6) | 0.32 |
| Estimated glomerular filtration rate <30 mL/min/1.73 m2 or dialysis | 133 (10.2) | 15 (12.2) | 1,170 (89.8) | 46 (4.0) | <0.001 |
| Severe anemia (hemoglobin <11 g/dL) | 216 (16.6) | 22 (10.8) | 1,087 (83.4) | 39 (3.6) | <0.001 |
| Prior spontaneous major bleeding in the past 6 months or at any time if recurrent† |
10 (0.8) | 1 (10.0) | 1,293 (99.2) | 60 (4.7) | 0.41 |
| Thrombocytopenia (platelet <100×109/L) | 14 (1.1) | 4 (28.6) | 1,289 (98.9) | 57 (4.5) | <0.001 |
| Chronic bleeding diathesis | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Liver cirrhosis with portal hypertension | 9 (0.7) | 0 (0.0) | 1,294 (99.3) | 61 (4.8) | NA |
| Active malignancy (excluding non-melanoma skin cancer) within the past 12 months |
76 (5.8) | 5 (6.6) | 1,227 (94.2) | 56 (4.7) | 0.42 |
| Prior spontaneous intracranial hemorrhage | 10 (0.8) | 1 (10.0) | 1,293 (99.2) | 60 (4.7) | 0.41 |
| Prior traumatic intracranial hemorrhage within the past 12 months | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Presence of a brain arteriovenous malformation | 0 (0) | 0 (0) | 0 (0) | 0 (0) | NA |
| Moderate or severe ischemic stroke within the past 6 months‡ | 8 (0.6) | 2 (25.0) | 1,295 (99.4) | 59 (4.6) | 0.006 |
| Non-deferrable major surgery on dual antiplatelet therapy | 1 (0.1) | 0 (0.0) | 1,302 (99.9) | 61 (4.8) | 0.90 |
| Recent major surgery or major trauma within 30 days before PCI | 4 (0.3) | 1 (25.0) | 1,299 (99.7) | 60 (4.7) | 0.04 |
| Minor criteria | |||||
| Age ≥75 years | 546 (41.9) | 36 (6.8) | 757 (58.1) | 25 (3.3) | 0.004 |
| Estimated glomerular filtration rate 30–59 mL/min/1.73 m2§ | 419 (32.2) | 20 (4.9) | 751 (57.6) | 26 (3.5) | 0.25 |
| Mild anemia (hemoglobin men: 11–12.9 g/dL, women 11–11.9 g/dL)§ |
110 (8.4) | 8 (7.4) | 977 (75) | 31 (3.2) | 0.03 |
| Prior spontaneous major bleeding in the past 6–12 months and not recurrent† |
1 (0.1) | 0 (0.0) | 1,302 (99.9) | 61 (4.8) | NA |
| Long-term use of non-steroidal anti-inflammatory drugs or steroids | 74 (5.7) | 7 (9.7) | 1,229 (94.3) | 54 (4.5) | 0.04 |
| Mild ischemic stroke or any ischemic stroke >6 months | 116 (8.9) | 6 (5.4) | 1,187 (91.1) | 55 (4.7) | 0.75 |
NA, not applicable. †Bleeding requiring hospitalization or transfusion. ‡National Institutes of Health Stroke Scale score ≥5. §Excluding patients who met major criteria.
The salient findings of this study were as follows:
1. By cancer screening using FIT, UGE and LGE, we newly diagnosed cancer in 2.2% of patients prior to PCI. During the follow up of 1-year, 2.0% of patients were newly diagnosed as having cancer; 48.0% of bleeding events led to a cancer diagnosis. In total, we newly diagnosed cancer in 4.1% of patients before and within 1-year after PCI.
2. The primary endpoint of NACCE was more often observed in patients with cancer than those without, driven by a greater risk of non-cardiac death and a tendency towards greater risk of stroke. Patients with cancer requiring ongoing therapy more often suffered from NACCE than those with recently (≤12 months) diagnosed cancer without ongoing therapy, mainly driven by a greater risk of major bleeding.
3. Patients with high bleeding risk according to ARC-HBR criteria were associated with greater bleeding risk than those without, irrespective of the presence or absence of cancer. Respective ARC-HBR criterion successfully identified patients with 1-year major bleeding risk ≥4% after PCI, except for those comprising a small proportion of patients (<1%).
In our cohort, approximately half of the patients underwent cancer screening with the use of a FIT, UGE and LGE, and we diagnosed cancer in 2.2% of patients prior to PCI. During the 1-year follow up, a total of 25 patients (2.0%) were diagnosed as having cancer, while bleeding events led to a new diagnosis of cancer in 12 (48.0%) patients. In total, we newly diagnosed cancer in 4.1% of patients before and within 1-year after PCI. The annual number of newly diagnosed cancer cases was 977,393 (0.8%) out of a 127 million Japanese population, according to the data from the Ministry of Health, Labor and Welfare in 2017.18 The expected percentage of newly diagnosed cancer patients in the general population matched with our cohort for age and sex was 2.0%, which is approximately half of our observations (4.1%). Moreover, since cancer screening was not performed in more than half of the patients, the number of patients with newly diagnosed cancer in our cohort was likely to be underestimated. Cancer screening was not performed in all of the 13 patients who were newly diagnosed as having digestive cancer, indicating that systematic cancer screening might identify more cancer patients than in our cohort. As the number of patients with coronary artery disease and those with cancer are increasing, especially in the aged population,2,3 and cardiovascular and cancer risk factors are often overlapped,5–7 coronary artery disease and cancer might incidentally coexist. As cancer screening was proved to be cost-effective even for average-risk persons aged >45 years,19 systematic screening for cancer prior to PCI is likely to be beneficial to reduce the risk of bleeding followed by lowering the risk of mortality.
Patients with cancer more often suffered from NACCE; its major drivers were greater risk of non-cardiac death (4.4% vs. 1.5%, P=0.006) and tendency towards greater risk of stroke (2.3% vs. 0.8%, P=0.08). Although cancer was a well-known risk factor for the subsequent non-cardiac mortality, as also shown in our cohort, it was intriguing that no significant between group difference was observed in the incidence of major bleeding. In our further exploratory analysis, although cumulative incidence of major bleeding was similar between the active vs. non-active cancer group (6.5% vs. 5.8%, P=0.80), the risk of major bleeding was significantly higher in patients with ongoing active cancer than in those with recently (≤12 months) active cancer (16.7% vs. 3.4%, P=0.049). Taken together, patients who were undergoing active cancer treatment such as surgery, radiotherapy, chemotherapy, or immunotherapy were associated with a greater risk of major bleeding after PCI, whereas those who had a history of cancer at any timepoint were not, possibly because of successful treatments without residual potential bleeding substrates. In the substudy of the Cardiovascular Outcomes for People Using Anticoagulation Strategies (COMPASS) trial, gastrointestinal or genitourinary bleeding was associated with higher rates of new cancer diagnosis, which is in line with our observations.20 When bleeding events occurred after PCI, holistic assessments are needed to identify cancers, especially for gastrointestinal and urinary cancers.
The ARC-HBR criteria included active cancer and CREDO Kyoto score, and included any cancer in their components. In addition to those criteria, high bleeding risk, according to the PRECISE DAPT score and PARIS bleeding score, was more prevalent in patients in the cancer group than in those in the non-cancer group. Although those risk stratification tools indicated that patients in the cancer group were associated with greater bleeding risk, as compared with those in the non-cancer group, no significant between group difference was observed in the incidence of major bleeding. Possible explanation for this discrepancy was that cancer, per se, might not be a risk factor of major bleeding, but actively ongoing cancer requiring maintenance therapy could confer a greater risk of bleeding. Larger-scale studies with a sufficient number of patients are mandatory to test this hypothesis.
Recently, ARC-HBR criteria have been proposed to standardize the definition of high bleeding risk, which was arbitrarily defined as major bleeding ≥4% at 1 year. Except for rare features such as liver cirrhosis with portal hypertension, non-deferrable major surgery, and non-recent prior spontaneous major bleeding, we validated each criterion as to confer major bleeding risk of ≥4% at 1 year. Furthermore, ARC-HBR criteria proved to be a useful tool to identify high bleeding risk in both cancer and non-cancer patients.
Study LimitationsThere are some limitations to this study. First, it is uncertain whether the results are generalizable due to a single-center design with relatively small sample size. Second, as we collected data retrospectively, there would be an underreporting bias, at least to some extent. However, all the data were systematically assessed from the electronic database automatically using regular expressions and confirmed manually. Finally, the clinical outcomes were assessed at 1 year. Further long-term follow-up studies are needed to fully elucidate the effect of cancer on clinical outcomes.
Cancer was diagnosed in 2.2% of 1,303 unselected patients before PCI by cancer screening and in 2.0% within 1-year after PCI. Cancer was associated with a greater risk of non-cardiac death, whereas ongoing active cancer was associated with greater risk of major bleeding. The ARC-HBR criteria successfully identified patients with 1-year major bleeding risk ≥4% after PCI, irrespective of the presence or absence of cancer.
This study did not receive any funding.
The authors have no conflicts of interest to declare.
The present study was approved by the Clinical research center, Kokura Memorial Hospital (Reference number: 20102301).
The deidentified participant data will not be shared.
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-20-1119
