Meta-analysis: colonoscopic post-polypectomy bleeding in patients on continued clopidogrel therapy

Authors


  • As part of AP&T's peer-review process, a technical check of this meta-analysis was performed by Dr P. Collins.

Correspondence to:

Dr M. Farkouh, 4N-474 585 University Avenue, Toronto M5G 2N2, ON, Canada.

E-mail: michael.farkouh@uhn.ca

Summary

Background

Current guidelines recommend the cessation of clopidogrel therapy 5 days and 7–10 days prior to colonoscopic polypectomy. Recent studies have advocated for continued clopidogrel as post-polypectomy bleeding (PPB) rates have been similar to those in the general population not on antithrombotic therapy.

Aim

To assess colonoscopic post-polypectomy bleeding in patients on continued clopidogrel therapy.

Methods

A literature search was conducted for studies that investigated PPB in patients on continued clopidogrel therapy. The primary outcome of interest was the pooled relative risk ratio (RR) of colonoscopic PPB in patients on continued clopidogrel therapy vs. controls. Secondary outcomes were a comparison of immediate and delayed colonoscopy PPB in patients on continued clopidogrel therapy vs. controls.

Results

Five observational studies included 574 subjects on continued clopidogrel therapy and 6169 control subjects. The pooled RR for PPB on continued clopidogrel therapy was 2.54 (95% CI 1.68–3.84, < 0.00001). For immediate PPB there was a nonsignificant pooled RR of 1.76 (95% CI 0.90–3.46, P = 0.10), and delayed PPB there was a significant pooled RR of 4.66 (95% CI 2.37–9.17, P < 0.00001).

Conclusions

The results of this meta-analysis suggest that continued clopidogrel increases the risk of delayed but not immediate post-polypectomy bleeding. Clopidogrel interruption in individuals with coronary artery disease predisposes to serious acute ischaemic events. In high-risk patients, endoscopists should be cognisant of these risks and consider deferring elective colonoscopy and polypectomy until it is considered safe to interrupt clopidogrel therapy.

Introduction

With the successful implementation of population-based screening programs for colorectal cancer in many jurisdictions, rates of colonoscopy and subsequent polypectomy rates are rising. Bleeding is the most common complication of colonoscopic polypectomy with rates ranging from 0.3% to 6%.[1] Immediate post-polypectomy bleeding (PPB) is observed at the time of the procedure, and delayed PPB can occur from hours to 30 days post procedure.[2, 3] Contributing risk factors for PPB include patient-related factors such as advanced age, hypertension, coronary artery disease, diabetes mellitus, chronic renal failure and chronic obstructive pulmonary disease. Procedure-related factors include size of the polyp removed, number of polyps removed, location and mode of removal.[4-7]

Although colorectal cancer is a leading cause of malignancy-related mortality it is overshadowed by the growing burden of cardiovascular disease. Clopidogrel or thienopyridine therapy is a cornerstone for treatment of myocardial ischaemia, and secondary prevention of cardiovascular and cerebrovascular events. The most frequent indication for clopidogrel prescription is prevention of coronary artery stent thrombosis following percutaneous coronary intervention (PCI). A meta-analysis by Eisenberg et al. demonstrated discontinuation of antithrombotic therapy is the dominant risk factor for coronary artery stent thrombosis in patients with drug-eluting stents, with events occurring as early as 5–10 days after cessation of clopidogrel therapy.[8]

Current guidelines recommend the cessation of clopidogrel therapy 5 days and 7–10 days prior to colonoscopic polypectomy to reduce the risk of procedure-related bleeding.[9, 10] This recommendation is based upon consensus expert opinion due to the lack of randomised control trials. Despite these recommendations, a survey of gastroenterologists in the United States showed that only 48% of respondents discontinued clopidogrel accordingly prior to colonoscopic polypectomy.[11] Recent studies by Feagins and Singh et al. have advocated the continued use of clopidogrel therapy for colonoscopic polypectomy as their studies have shown PPB rates similar to those reported in the general population.[4, 8, 12] We performed a meta-analysis of studies assessing the risk of clinically significant immediate and delayed PPB in patients receiving continued clopidogrel therapy.

Methods

A systematic search was conducted to retrieve studies that investigated the risk of colonoscopic PPB in patients on continued clopidogrel therapy. We identified potential English-language sources from the Pubmed, Medline and EMBASE databases from 1950 to March 2013. Keywords used were ‘clopidogrel, thienopyridine’ and ‘colonoscopy, polypectomy’. In addition, reference lists of any studies meeting inclusion criteria were reviewed to identify additional relevant publications by manual search.

Inclusion/exclusion criteria

Studies were included if they met the following criteria: (i) subjects included patients on continued clopidogrel therapy who underwent colonoscopic polypectomy; (ii) a control group included patients who were not taking clopidogrel who underwent colonoscopic polypectomy; (iii) subjects were not on concomitant anticoagulant therapy such as vitamin K antagonists (warfarin) or heparin; (iv) immediate PPB was defined as observed bleeding at the time of the procedure, delayed PPB was defined as observed bleeding that occurred within 30 days of the procedure; (v) studies reported PPB as clinically significant observed bleeding requiring hospitalisation, red cell transfusion or endoscopic re-evaluation; and (vi) published as a full article in the English. Eligibility assessment and data extraction were carried out independently by two investigators (SG and NN) with discrepancies resolved by consensus in consultation with the senior author.

Outcomes of interest

The primary outcome of interest was the pooled relative risk ratio (RR) of colonoscopic PPB in patients on continued clopidogrel therapy vs. controls. Secondary outcomes were a comparison of immediate and delayed colonoscopy PPB in patients on continued clopidogrel therapy vs. controls.

Study quality and data extraction

Quality assessment was carried out independently by two investigators (SG and NN) using the Newcastle-Ottawa quality assessment scale. We assessed eligible studies by three criteria: the selection of the study groups (0–4 points), the comparability of the groups (0–2 points) and the ascertainment of either the exposure or outcome of interest (0–3 points), with a total score of 9. A score ≥5 was adequate for inclusion in the meta-analysis. Discrepancies in interpretation of data and inclusion of studies were resolved by in consultation with the senior author.

Statistical analysis

Meta-analysis was conducted by combining the risk ratios of individual studies into a pooled risk ratio using a random-effects model. Relative risk ratios are reported with 95% confidence intervals. We tested for heterogeneity using the chi-squared test and the I2 test. The I2 test describes the percentage of variability in effect estimates that is due to heterogeneity rather than chance. A value of 25% suggests low variability, 50% suggests moderate variability and 75% suggests high variability between studies.[13] Funnel plots were constructed to assess for publication bias. Analyses were performed with RevMan 5.1 [Review Manager (RevMan) [Computer program] Version 5.1. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011].

Results

Search results

Our search strategy yielded 151 studies, of which 127 were excluded on review of the title and abstract. A further 19 studies were excluded after careful review of the full text. Where only partial information on the outcomes of interest was reported in the article, the authors were contacted and asked to provide additional information. Using a questionnaire, Waters et al. assessed the risk of cardiovascular events and PPB in patients who continued/discontinued clopidogrel prior to polypectomy.[14] The authors could not provide specific results for inclusion in our meta-analysis. Two abstracts were excluded by Feagins and colleagues, as the subjects were part of larger data sets of studies included in this meta-analysis.[15, 16]

Overall, five studies were eligible for inclusion in the systematic review. Among these, three studies allowed calculation of risk ratios for immediate PPB and five studies allowed calculation of risk ratios for delayed PPB. Figures 1-3 summarises outcomes for each individual study, while Table 1 summarises the pooled results of the meta-analysis.

Table 1. Summary of pooled analysis
 Clopidogrel groupControl groupRelative risk ratioLower 95% CIUpper 95% CIP valueI2%
Immediate PPB (%)22/431 (5.10)66/3920 (1.68)1.760.903.460.1030
Delayed PPB (%)15/565 (2.65)37/6158 (0.60)4.662.379.17<0.000010
Total PPB (%)37/574 (6.45)103/6169 (1.67)2.541.683.84<0.000012
Figure 1.

Continued clopidogrel and immediate and delayed post-polypectomy bleed. Events = immediate and delayed post-polypectomy bleed.

Figure 2.

Continued clopidogrel and immediate post-polypectomy bleed. Events = immediate post-polypectomy bleed.

Figure 3.

Continued clopidogrel and delayed post-polypectomy bleed. Events = delayed post-polypectomy bleed.

Characteristics of included studies

Characteristics of all eligible studies are outlined in Table S1. All studies included patients referred for elective colonoscopy. Feagins et al.[12] continued ASA therapy in both the continued clopidogrel group and the control group, and stopped clopidogrel 5 days prior to polypectomy in the control group. 78.8% of clopidogrel users were on concomitant ASA compared to 27.9% of controls (P = 0.64). Rodino et al.[17] included patients on both low-dose acetylsalicylic acid (ASA) and clopidogrel with recent coronary stenting for myocardial infarction in the past one year, and all subjects remained on continued ASA therapy during the time of polypectomy. Singh et al.[4] continued ASA therapy in all patients, with 54% of patients in the continued clopidogrel group and 32% of controls receiving concomitant ASA therapy. Feagins and Iqbal[18] continued ASA therapy in all patients, with 87.8% of patients in the continued thienopyridine group and 40.1% of controls. Indications for thienopyridine usage included 82.2% with a coronary stent, 3.2% with a peripheral stent, 1.4% with both a coronary and peripheral stent, 8.7% with coronary artery disease (with or without coronary artery bypass graft), 4.6% for cerebral vascular accident or transient ischaemic attack, and 0.5% who were allergic to ASA. The remaining study did not mention specifically concomitant ASA use. Four studies [4, 12, 18, 19] defined delayed PPB as clinically significant observed bleeding within 30 days or 4 weeks of polypectomy, while Rodino et al.[17] defined delayed PPB as clinically significant observed bleeding within 15 days of polypectomy.

We assessed the quality of included studies using the Newcastle-Ottawa quality assessment scales. A score ≥5 was considered adequate quality for inclusion. Overall, the quality of studies was high. Three studies scored 8 points,[4, 12, 18] one study scored 7 points,[19] one study scored 5 points.[17] There was no publication bias detected in comparing immediate and delayed post-polypectomy bleeding in patients on continued clopidogrel therapy, respectively, based on the symmetry of the funnel plots.

Colonoscopic post-polypectomy bleeding

Five observational studies[4, 13, 17, 19] met inclusion criteria with 574 subjects on continued clopidogrel therapy and 6169 controls (Figure 1). Of 574 subjects who continued clopidogrel, 37 (6.45%) experienced PPB vs. 103 of 6169 controls (1.67%). The pooled RR for PPB on continued clopidogrel therapy was 2.54 (95% CI 1.68–3.84, P < 0.00001, I2 = 2%) suggesting an increase in PPB. Sensitivity analysis revealed that findings remained statistically significant after removal of each individual study.

Immediate post-polypectomy bleeding

Three observational studies that assessed immediate PPB [4, 18, 19] met inclusion criteria, with 431 subjects on continued clopidogrel and 3920 controls (Figure 2). Immediate PPB was observed in 22 of 431 on clopidogrel (5.1%) vs. 66 of 3920 (1.68%) controls, with a nonsignificant pooled RR of 1.76 (95% CI 0.90–3.46, P = 0.10, I2 = 30%). PPB sensitivity analysis demonstrated similar results when each individualised study was removed.

Delayed post-polypectomy bleeding

Five observational studies that assessed delayed PPB [4, 13, 17-19] met inclusion criteria with 565 subjects on continued clopidogrel and 6158 controls (Figure 3). Delayed PPB was observed in 15 of the 565 on clopidogrel (2.65%) vs. 37 of 6158 (0.6%) controls, with a significant pooled RR of 4.66 (95% CI 2.37–9.17, P < 0.00001, I2 = 0%). Sensitivity analysis demonstrated that removal of the subjects from Singh et al.[4] increased the RR to 5.62 (95% CI 2.29–13.82, P = 0.0002), and removal of the subjects from Grossman et al.[20] decreased the RR to 3.93 (95% CI 1.74–8.89, P = 0.001). However, the pooled RR remained significant with removal of each individual study.

Discussion

The results of this meta-analysis suggest that continued clopidogrel increases the risk of delayed but not immediate PPB. There are several limitations to our study. These include most notably the inability to control for covariates of PPB including concomitant medication use such as ASA, NSAIDs, proton pump inhibitors, duration of clopidogrel therapy and other patient-related risk factors such as advanced age, comorbidities (hypertension, coronary artery disease, diabetes mellitus, chronic renal failure and chronic obstructive pulmonary disease) and prior PPB.

Only five observational studies met inclusion criteria for this meta-analysis, which limits patient numbers, and therefore limits the study power to detect statistically significant findings. Specifically, the risk of immediate PPB in patients on continued clopidogrel therapy included only three studies; although the results were not statistically significant, it is likely clinically significant. Our study was unable to account for concomitant ASA attributing to the risk of PPB. In the study by Singh et al.,[4] eight patients in the clopidogrel group who experienced PPB were on ASA compared to 16 controls (P = 0.002). Multivariate analysis to identify independent risk factors associated with PPB found concomitant ASA usage with clopidogrel increased the risk of PPB (OR 3.69, 95% CI 1.60–8.52, P = 0.002). Feagins et al.[12] found that the single patient with a PPB in the clopidogrel group took concomitant ASA compared with five patients with PPB in the control group. In the study by Feagins and Iqbal,[18] all five patients who experienced delayed PPB in the clopidogrel group were on concomitant ASA. Interestingly, univariate analysis demonstrated that ASA usage was not associated with PPB.

It is also unclear whether specific procedure-related factors may have increased the risk of PPB, such as number and location of polyps removed, and technique of removal. We were unable to account for haemostatic techniques to reduce bleeding risk. Feagins et al.[12] used prophylactic haemoclips in 22% of patients on continued clopidogrel therapy compared to 10.3% of the control group (P = 0.0001). Similarly, Singh et al.[4] used prophylactic endoclip placement in 2.2% of patients on continued clopidogrel therapy compared to 0.32% of patients in the control group (P < 0.01). Feagins and Iqbal[18] used prophylactic haemoclips in 19.2% of patients on continued thienopyridine and in 5.7% of controls (P < 0.0001). The other studies do not report differences between the two groups with regards to endoscopic clip placement. We were also not able to assess cardiovascular events among patients in whom clopidogrel therapy was stopped.

Despite these limitations, the results of this review may have important implications for clinical practice. Among patients taking clopidogrel, the rate of delayed PPB was 2.65%. This is similar to population-based estimates of PPB risk in patient not on antithrombotic therapy, but higher than rates reported in previous studies that advocated continuing clopidogrel for polypectomy.[4, 12] The case fatality rate for stent thrombosis has been reported to be as high as 45% with cessation of antithrombotic agents as the strongest risk factor.[20] In contrast, PPB is not usually life threatening and can be controlled with minimally invasive strategies; none of the studies included in this meta-analysis reported a fatality.

Moving forward, gastroenterologists must weigh the risk of bleeding against that of re-infarction or stent thrombosis among patients on clopidogrel undergoing colonoscopy. Clopidogrel interruption in individuals with coronary artery disease predisposes to serious acute ischaemic events. However, continuing clopidogrel increases the risk of delayed PPB. In high-risk patients, endoscopists should be cognisant of these risks and consider deferring elective colonoscopy and polypectomy until it is considered safe to interrupt clopidogrel therapy.

Authorship

Guarantor of the article: Dr Michael Farkouh.

Author contributions: All authors were involved in study design, data collection, analysis and creation of the manuscript. All authors approved the final version of the manuscript.

Acknowledgements

Declaration of personal interests: The authors do not believe that they have any conflicts of interest with regards to this research paper. All authors meet criteria for authorship. This article represents original work and is not under consideration for publication elsewhere.

Declaration of funding interests: None.

Ancillary