While endoscopic resection is the standard method of treatment for colorectal tumors and highly effective in reducing the incidence of colon cancer and its subsequent mortality [
Several factors such as large polyp size and sessile form have been reported to increase the risk of bleeding after endoscopic treatment, but controversies still exist [
We performed a case-control study to evaluate the risk factors including patient and tumor characteristics associated with bleeding after endoscopic treatment.
From January 2006 to October 2012, 1,970 cases with polyps of the colon and rectum, measuring larger than 10 mm, were removed by polypectomy, endoscopic mucosal resection (EMR), or endoscopic submucosal dissection (ESD) at our hospital. All procedures were performed by 5 expert endoscopists. Patients were instructed to contact the endoscopist immediately if they had bloody feces. We identified patients who reported delayed bleeding by examining in patients’ records for the 30-day period after the procedure and checking for hospital visits to ensure that we did not overlook any cases of delayed bleeding in this study. “Cases” were defined as those patients who developed delayed bleeding, which was designated as bleeding that occurred 6 or more hours after endoscopic treatment. For such cases, after admission, second-look colonoscopy was performed to identify the origin of bleeding, and endoscopic hemostasis was performed when active bleeding was found. “Controls” were defined as patients who underwent endoscopic treatment with colonoscopy but did not develop delayed bleeding and were matched for age and gender with cases. Controls were selected in a 3 : 1 ratio compared with cases (Figure
Flow diagram showing the process for analysis in the case-control study.
We investigated patient-factors (resuming anticoagulants (warfarin, heparin, and dabigatran) and antiplatelet agents (aspirin and thienopyridines) within 5 days following endoscopic resection, hypertension, and diabetes mellitus) and tumor-factors (morphology, size, location, and resection technique) by univariate and multivariate logistic regression analysis. If anticoagulants or antiplatelet agents had been prescribed, they were discontinued at least 5 days before the procedure. To reduce the risk of thromboembolic events, patients on warfarin and who were at high thromboembolic risk were switched to a shorter-acting therapy—unfractionated heparin—(i.e., bridge) in the periendoscopic period according to guidelines [
Statistical analysis was performed using SPSS version 19.0 for Windows. The correlation between delayed bleeding and risk variables was assessed by a two-sided Fisher’s exact test. In addition, standard logistic regression methodologies were used to calculate the relative risks as odds ratios (ORs) with 95% confidence intervals (CIs). The median intervals between endoscopic treatment and bleeding in cases with and without anticoagulants were compared with the Mann-Whitney
During the study period, a total of 1,970 cases with colorectal polyps were found and polyps were resected by polypectomy, EMR, or ESD. Of these, 52 cases (2.6%) met the study criteria for delayed bleeding, with 50 requiring endoscopic hemostasis and 2 blood transfusion. During the perioperative period in this study, there were no thrombogenic events. An additional 156 cases were selected, as described above, to serve as controls. Polyps in the case and control group were resected by polypectomy (62 cases), EMR (130 cases), and ESD (16 cases). Baseline characteristics of cases and controls are summarized in Table
Characteristics of study subjects.
Cases |
Controls |
|
---|---|---|
Age, years |
59.5 ± 11.6 | 59.5 ± 11.6 |
Male, sex, |
44 (85) | 132 (85) |
Table
Univariate analysis of patient-factors for delayed bleeding.
Cases (%) | Controls (%) | Odds ratio |
|
|
---|---|---|---|---|
Resuming anticoagulants | 12 (23) | 9 (6) | 4.9 | 0.0003* |
Resuming aspirin | 6 (12) | 11 (7) | 1.7 | 0.31 |
Resuming thienopyridines | 3 (15) | 7 (4) | 1.8 | 0.37 |
Hypertension | 16 (31) | 43 (28) | 1.2 | 0.66 |
Diabetes mellitus | 10 (19) | 22 (14) | 1.5 | 0.37 |
Results of univariate analysis of tumor-factors for delayed bleeding are shown in Table
Univariate analysis of tumor-factors for delayed bleeding.
Cases (%) | Controls (%) | Odds ratio |
|
|
---|---|---|---|---|
Morphology |
36 (69) | 95 (61) | 1.4 | 0.28 |
Polyp size ( |
16 (31) | 40 (26) | 1.3 | 0.47 |
Polyp location |
32 (62) | 99 (63) | 0.9 | 0.80 |
Resection technique |
37 (71) | 109 (70) | 1.1 | 0.86 |
Left side: rectum-descending colon.
Right side: transverse colon-cecum.
Multivariate analysis of factors influencing the bleeding revealed that a significant factor was resuming anticoagulants within 5 days after endoscopic resection (OR 10.2; 95% CI 2.7–38.3;
Multivariate analysis of risk factors associated with delayed bleeding.
Odds ratio | 95% CI |
|
|
---|---|---|---|
Resuming anticoagulants | 10.2 | 2.7–38.3 | 0.0006* |
Resuming aspirin | 0.6 | 0.1–2.4 | 0.43 |
Resuming thienopyridine | 0.9 | 0.1–2.6 | 0.40 |
Resection technique | 1.2 | 0.5–2.6 | 0.71 |
Tumor size | 1.5 | 0.7–3.1 | 0.29 |
Figure
Median days from endoscopic treatment to bleeding.
Anticoagulants | + | − |
|
|
|||
Median number of days |
4 (2–7) | 2 (0–18) | 0.04* |
Days from treatment to bleeding in cases who were and were not taking anticoagulants. Delayed bleeding occurred within 7 days in most cases (87%).
The use of anticoagulants and antiplatelet medications for various cardiovascular and hematologic conditions has become increasingly widespread over the past decade [
This study revealed that resuming anticoagulants within 5 days after endoscopic resection is an independent risk factor and is strongly associated with severe delayed bleeding, whereas resuming antiplatelet agents within 5 days has no association. It can be hypothesized that these results are associated with the mechanism of hemostasis: anticoagulants work on the secondary hemostasis process, such as the manufacturing of fibrin, while antiplatelet agents work on the primary hemostasis process, such as the cohesion of platelets. The former process is stronger, so anticoagulants, which prevent the secondary hemostasis process, tend to enable easy bleeding and thus pose a greater risk for delayed bleeding.
Our practice for resumption of anticoagulants following endoscopic treatment is to restart warfarin and/or heparin within 5 days. Anticoagulants were resumed within 3 days in most cases who had been taking anticoagulants. Our data suggests that resuming anticoagulants within 3 days of endoscopic treatment may result in an increase in the risk of delayed bleeding. We propose taking precautionary hemostatic measures after endoscopic treatment, as this may be effective in preventing delayed bleeding in these cases.
In our study, antiplatelet agents were resumed within 5 days after treatment, and there was no incremental increase in the risk of bleeding with antiplatelet agents, including aspirin. It may be appropriate, however, to consider that all subjects in this study were of Japanese ethnicity. In a study conducted by Lee et al. [
Based on our results, in cases involving thienopyridines, we can restart this medicine within 5 days after treatment with greater confidence regardless of thromboembolic risk. In contrast, in cases with low thromboembolic risk, as determined by United States guidelines [
Although our data showed no correlation between polyp location and risk of bleeding, Buddingh et al. [
When bleeding occurs, it is usually recognized within 7 days [
Our study had a number of limitations. First, this study was conducted retrospectively in the single institution. A further study to prospectively assess the long-term outcome will be necessary to overcome this limitation. Second, most data in our study were collected before the setting of new Japanese guidelines for gastroenterological endoscopy in patients undergoing antithrombotic treatment [
In summary, resuming anticoagulants within 5 days after endoscopic treatment was associated with delayed bleeding whereas resuming thienopyridines within 5 days after endoscopic treatment was not. In addition, patients should be monitored for a minimum of 7 days following endoscopic treatment.
No conflict of interests is declared.