Meta-analysis finds increased bleeding with antithrombotic therapy during breast biopsy, but clinically relevant events are rare

This systematic review and meta-analysis examined the safety of continuing antithrombotic therapy in patients undergoing image-guided breast biopsy. The analysis pooled data from 8 observational studies, encompassing a total of 11,524 patients. The population consisted of 1,154 patients who continued their antithrombotic therapy during the procedure and 10,370 control patients who were not on such therapy. The specific setting for the included studies was not reported, nor was the duration of patient follow-up after the biopsy.

The intervention was the continuation of various antithrombotic medications during the periprocedural period of an image-guided breast biopsy. The comparator group consisted of patients undergoing the same procedure who were not on antithrombotic therapy. The meta-analysis did not specify the exact types, doses, or regimens of the antithrombotic drugs, noting significant variability across the included studies. Similarly, the specific biopsy techniques (e.g., core needle, vacuum-assisted) were not standardized and varied among the studies.

The primary outcome was the occurrence of any bleeding event. The pooled analysis found a statistically significant increase in bleeding frequency among patients on antithrombotic therapy. Specifically, 203 out of 1,154 patients (17.9%) in the antithrombotic group experienced a bleeding event, compared to 1,110 out of 10,370 patients (10.7%) in the control group. The pooled odds ratio was 1.89 (P < 0.001), indicating an association between therapy continuation and higher odds of bleeding.

A key secondary outcome was the rate of clinically relevant bleeding events. The data for this outcome were sparse and inconsistently reported across studies. In the one study that provided data, only 3 out of an unspecified total number of patients (0.23%) experienced a clinically relevant bleeding event. An overall pooled effect size for clinically relevant bleeding was not reported due to data fragmentation and variability in how this outcome was defined across the included primary studies.

Safety and tolerability findings centered entirely on bleeding as the adverse event of interest. The analysis confirmed that any bleeding event was more common with continued antithrombotic use. However, the vast majority of these events were minor. The reported rate of serious adverse events, defined as clinically relevant bleeding, was very low at 0.23% in the single study reporting it. Data on therapy discontinuations due to bleeding or other tolerability issues were not reported.

These results add to a growing body of evidence, primarily from observational studies in various procedural contexts, suggesting that many minor procedures can be performed safely without interrupting antithrombotic therapy. Prior landmark studies in other fields (e.g., dermatology, dentistry) have similarly found that the increased risk of minor bleeding often does not translate to a significant increase in major, clinically important hemorrhage, supporting a trend toward periprocedural continuation for many patients.

Several important methodological limitations burden the certainty of these findings. The primary limitation is the unstandardized reporting and significant data fragmentation across the eight included studies. There was notable variability in the specific antithrombotic drugs used, the techniques employed for the image-guided biopsy, and, crucially, the definitions used for 'clinically relevant bleeding.' The underlying data are observational, preventing the establishment of causality, and the small number of studies limits the robustness of the meta-analysis.

The clinical implication is that for patients on antithrombotic therapy requiring an image-guided breast biopsy, the procedure can likely be performed without routine suspension of therapy. The evidence suggests that while the overall frequency of bleeding is higher, the absolute increase is modest and the events are predominantly minor. The very low observed rate of clinically relevant bleeding supports a practice of continuing therapy, which avoids the thrombotic risks associated with interruption. This must be balanced against individual patient factors, such as the specific medication and bleeding risk.

Significant questions remain unanswered. The analysis could not differentiate risks between specific classes of antithrombotic agents (e.g., direct oral anticoagulants vs. vitamin K antagonists vs. antiplatelets). The optimal management for patients on dual antiplatelet therapy or combined anticoagulant/antiplatelet regimens is unclear. Furthermore, the lack of a standardized, consensus definition for 'clinically relevant bleeding' in this context makes it difficult to precisely quantify the risk of meaningful harm. More prospective, standardized data are needed to refine risk stratification.