Ceasing Anticoagulants for Interventional Pain Procedures

A troublesome issue in the practice of interventional pain medicine has been how to manage patients who need to undergo an invasive procedure but who are taking anticoagulants. Between five and 20 years ago, various professional societies issued guidelines for ceasing anticoagulants in preparation for various spinal procedures [1–5]. The implication was that not ceasing anticoagulants ran the risk of hemorrhagic complications from these procedures. However, with respect to several procedures for spinal pain, this sense of risk was based on only theoretical considerations, or based on experience with certain procedures being extrapolated to other procedures. There was no empirical evidence of that risk. The advent of anticoagulant guidelines raised concerns among many physicians who performed invasive procedures for spinal pain. They felt that their procedures were relatively innocuous; in addition, ceasing anticoagulants introduced diverse complexities, such as different detailed regimens for ceasing every anticoagulant that patients might be taking, but if they did not comply with the guidelines, they could be held culpable. These concerns prompted a small number of pioneering studies. A review of their results found that many spine procedures appeared to be safe, but ceasing anticoagulants introduced the risk of medical complications [6]. The present issue of the Journal features two new studies on this topic, which extend the base of evidence. The first study [7] is a typical hospital-based study in which the authors present their experience with a modest number of procedures (275). When stratified for different procedures, this small sample yields sample sizes that are too small for drawing robust conclusions, as evidenced by the wide confidence intervals of the zero prevalence of hemorrhagic complaints for each procedure. The authors recognized this limitation but explained that their objective was to add numbers to the literature so that a meta-analysis might be performed in the future. The second study [8] is an update of a self-funded study in which the lead author—assisted by his staff and partners in a private practice—collected data over a period of 13 years on >11,000 spinal procedures. For several types of procedure, this did not yield sufficiently large sample sizes, but robust data emerged for two procedures (lumbar transforaminal injections and lumbar facet injections) and two anticoagulants (warfarin and clopidogrel). For these two procedures and agents, the sample sizes were sufficiently large to show, with statistical significance, that continuing anticoagulants was safer than ceasing anticoagulants. However, the latter study [8] also reported how exhausting it is for a single private practice to collect data on multiple procedures and multiple anticoagulants. It explains that a single practice is unlikely ever to be able to collect enough data for every procedure and every anticoagulant to provide statistically robust data. If interventional pain physicians want data that show how safe their procedures are, the former—and smaller—of these two studies provides a model. That is not to say that the Journal now invites a flood of replica small studies to be published. Rather, if a consortium of investigators could pool their experience in a valid manner, they could publish a massive, multicenter report, in the manner of a registry. Thereby, they could promptly provide the data set that Ehsanian et al. [7] want for their meta-analysis, and thereby give Endres et al. [8] a rest from their 13-year project.