**Title**
Clinical association of anticoagulant-related bleeding events with new-onset cancer in atrial fibrillation (AF): A systematic review and meta-analysis
**Review question(s)**
Clinical association of anticoagulant-related bleeding events with new-onset cancer in atrial fibrillation (AF).
**Searches**
We will search the following databases: MEDLINE (via PubMed), the Cochrane Library, Scopus, grey literature sources and PROSPERO for eligible studies and for systematic reviews which have been already published. Two reviewers will independently extract data. Any disagreement will be resolved by consensus. Authors will be contacted in order to obtain any missing data relevant to the analysis.
**Types of study to be included**
Observational studies (Cohort and Case–control studies). Case-reports studies were excluded.
**Condition or domain being studied**
New-onset cancer detection.
**Participants/ population**
*Inclusion criteria:*
(1) Patients over 18 years old with atrial fibrillation (first diagnosed, paroxysmal, persistent, longstanding persistent, permanent) under any oral anticoagulants (warfarin, dabigatran, rivaroxaban, apixaban, edoxaban, etc.);
(2) Oral anticoagulant administration could have been initiated to patients either in their enrollment or earlier.;
*Exclusion criteria:*
(1) Patients with valvular AF
**Intervention(s), exposure(s)**
A bleeding event (minor or major) during the follow-up of the included studies.
**Comparator(s)/ control**
No bleeding event during the follow-up period of the included studies.
**Context**
The proportion of detected new-onset cancer during the follow-up should have been described both in patients after a bleeding event and in patients that had not bled. Νo restrictions were set about the type of both bleeding events and new-onset cancer.
**Outcome(s)**
Primary: The difference in new-onset cancer detection.
Secondary: The incidence of gastrointestinal new-onset cancer in one and three months after a gastrointestinal bleeding event.
**Data extraction, (selection and coding)**
All studies will be imported into a reference management software (Mendeley). After duplicate removal, two reviewers will independently screen all titles and abstracts and peruse full texts for eligible studies. Any discordance regarding study eligibility will be resolved by consultation with a third review author.
**Risk of bias (quality) assessment**
The quality of the eligible studies was evaluated using the risk of bias in non‐randomized studies of interventions (ROBINS-I) tool of Cochrane. A sensitivity analysis will be performed excluding studies with low quality.
**Strategy for data synthesis**
-Random effects model will be used. We will report effect sizes as pooled odds ratio (OR) and their 95% confidence intervals (CI).
-Heterogeneity will be tested with the Cochran chi-square test and the degree of heterogeneity quantified by the I2 statistics and its 95% CI. An I2 over 50% will be described as high.
-Subgroup analysis: Type of oral anticoagulants [(non-vitamin K oral anticoagulants (NOACs) versus vitamin K antagonists (VKAs)]
-Publication bias will be formally tested with the Egger’s test.
-Sensitivity analyses will be performed if there is statistically significant heterogeneity.
**Subject index terms**
“atrial fibrillation”, “bleeding”, “cancer” and “oral anticoagulants” in both free text and Medical Subject Headings (MeSH) format.
**Organisational affiliation of the review**
Aristotle University of Thessaloniki, Greece
