Lumbar Puncture on an Anticoagulated Patient in the Emergency Department: Is it safe?

lumbar punctureThe lumbar puncture (LP) procedure is commonly performed in the Emergency Department (ED). While minor complications of LP such as post-procedure headache or back pain occur somewhat regularly, significant complications such as post-procedural spinal hematomas, are rare.1 Despite their low incidence, these spinal hematomas are associated with a significant amount of morbidity for the patient and increased medicolegal risk for the provider.

While post-LP bleeding is usually self-limited and relatively inconsequential, the incidence, severity, and consequence of bleeding may be higher for patients on anticoagulation therapy or anti-platelet agents. To date there has been limited discussion in the Emergency Medicine literature evaluating the potential risk of performing a LP in a patient who is anticoagulated. The following are 5 papers addressing the risks.

Risk for patients taking systemic anticoagulants: The evidence

A 1981 study compared 342 inpatients who were anticoagulated with heparin or warfarin following LP with 342 patients who underwent a LP without subsequent anticoagulation.2 Indications for LP included meningitis, multiple sclerosis, and cerebral ischemia. All LPs were performed using a 20 gauge needle, and patients who had red blood cells present on microscopic cerebrospinal fluid (CSF) examination were categorized as having a traumatic tap..

Major bleeding complications, defined as paraparesis or severe back pain lasting >48 hours, occurred more often in patients on anticoagulants (6.7% vs 1.8%, P<0.001). Patients who underwent a “traumatic tap” had a higher incidence of major complications (76% vs 2%, p<0.001). Of the patients who were categorized as having major complications, only 5 anticoagulated patients developed paraparesis from hematomas requiring operative intervention. The remainder of the patients with major complications had back pain without any confirmatory studies. While the authors concluded that systemic anticoagulants increased the risk of significant post-LP bleeding complications, the overall risk of patients with bleeding complications needing an emergent intervention was low (~0.3%).Paper: Nontraumatic acute spinal subdural hematoma: report of five cases and review of the literature

In a retrospective series of including 106 patients3, the authors reviewed cases of acute non-traumatic spinal subdural hematoma. In 50 (47%) of the cases, the patients underwent a LP prior to developing a hematoma. Over half (n=57, 54%) of the patients were noted to have a defective hemostatic mechanism or to be using systemic anticoagulants. The authors concluded that both LP and abnormal coagulation were associated with the development of spinal hematomas; however, it is not clear from the data how many of these hematomas were the result of an anticoagulated patient undergoing a LP.

In a 10-year retrospective review of 1,710,000 spinal and epidural nerve blocks performed by anesthesiologists4, the overall rate of spinal hematoma was ~1:50,000. In certain subsets such as females undergoing knee arthroplasty, the rate of spinal hematoma was as high as 1:3600.  In patients who developed a spinal hematoma only 1/3 had a documented coagulopathy or were taking anticoagulants.  Procedural “difficulty” (including multiple needle punctures) was documented in 30% of cases with significant bleeding. The authors concluded that select subsets of patients exhibit a higher risk of hematoma formation after a spinal anesthetic procedure, yet this conclusion is not completely supported by the data. While patients on anticoagulants may develop post-LP bleeding, the majority of the patients in this data set with bleeding complications were not anticoagulated. Despite including a large number of patients, this data is unable to show a causative link between anticoagulants and post-LP bleeding and at best shows a weak association.

To date there is no available data regarding the risk of post-LP bleeding in patients taking novel oral anticoagulants (NOACs.) As previous studies have shown similar risks of bleeding when comparing traditional anticoagulants to NOACs, it would be reasonable to assume that these agents pose a similar risk following LP.

Risk for patients taking antiplatelet agents: The evidence

Horlocker et al.5 studied a group of 386 patients with a history of antiplatelet medications and undergoing spinal anesthesia. Regular-strength aspirin was the most common anti-platelet medication (39%). Following spinal or epidural anesthesia for elective orthopedic procedures, there were no noted reports of spinal hematoma, despite a report of blood in the catheter or needle during insertion in 22% of cases. In a similar study by the same authors6 among 1,035 NSAID users undergoing outpatient epidural steroid injections, there were no reported spinal hematomas. The authors concluded that there was no evidence among NSAIDS users (including full-strength aspirin) showing an increased risk of significant spinal bleeding following epidural injection. The data regarding the risk posed by clopidogrel is limited to an isolated case report of a patient who developed a post-procedural epidural hematoma while taking clopidogrel, diclofenac, and aspirin.

Overall the data is limited

To date there are limited high-quality studies examining the risk of post-LP bleeding in anticoagulated patients. Only one study specifically examined post-procedure bleeding after lumbar punctures showing a possible increase in the risk of complications in anticoagulated patients. Despite this potential risk, the overall risk of patients developing a complication that needed an intervention was low. The available data from the inpatient and operating room settings shows a possible association between anticoagulants and post-procedural bleeding but suggests that overall patients are at low risk of developing a significant bleeding complication. 7

While there are fundamental differences between the operating room and ED settings, the most comprehensive guidelines regarding spinal puncture originate from the anesthesia literature.

  • Recommendations for anticoagulated patients requiring elective spinal anesthesia include discontinuing anticoagulant use for a duration of twice the half-life of the drug and, when possible, normalizing blood coagulation profiles prior to the procedure.
  • For patients with impaired renal function, who are thought to be at an increased risk of coagulopathy, some sources call for a delay of up to 4 days before performing a spinal puncture. 8
  • When invasive spinal procedures are performed in anticoagulated patient, current recommendations call for 24-hours of close post-procedure neurologic monitoring. 9

Unfortunately for ED providers, these guidelines suggest the need for caution without fully appreciating the emergent nature and technical nuances of LP performed in the ED. While not directly applicable, ED providers need to balance these conservative recommendations with the overall low rate of significant complications following LP.

The Bottom Line

The exact risk of significant bleeding in anticoagulated patients undergoing an emergent LP is unknown and there is not sufficient evidence to suggest that LP should not be performed in the anticoagulated patient. While the overall rate of significant complication is low, there is a high rate of morbidity in patients who develop post-LP bleeding. Given this risk, providers should make the decision to perform an LP on an anticoagulated patient based on factors such as:

  • Age
  • Comorbidities
  • Spinal pathology
  • Reason for anticoagulation
  • Feasibility of reversing anticoagulation
  • Clinical gestalt of the presenting illness

In addition, the overall utility of the test should be balanced with other considerations such as the possibility of empiric treatment or alternative, non-invasive diagnostic tests.

Providers should involve patients in the medical decision making process both before performing a LP in an anticoagulated patient and when determining the appropriate disposition. While some patients with a high post-procedural risk of bleeding may need to be admitted for further monitoring, other patients with a lower risk may be reasonably discharged with a plan for followup, and specific return instructions.

– Pre-publication peer review performed by Andrew Perron, MD and Emily Vance, PharmD –

1.
Boon J, Abrahams P, Meiring J, Welch T. Lumbar puncture: anatomical review of a clinical skill. Clin Anat. 2004;17(7):544-553. [PubMed]
2.
Ruff R, Dougherty J. Complications of lumbar puncture followed by anticoagulation. Stroke. 1981;12(6):879-881. [PubMed]
3.
Domenicucci M, Ramieri A, Ciappetta P, Delfini R. Nontraumatic acute spinal subdural hematoma: report of five cases and review of the literature. J Neurosurg. 1999;91(1 Suppl):65-73. [PubMed]
4.
Moen V, Dahlgren N, Irestedt L. Severe neurological complications after central neuraxial blockades in Sweden 1990-1999. Anesthesiology. 2004;101(4):950-959. [PubMed]
5.
Horlocker T, Wedel D, Schroeder D, et al. Preoperative antiplatelet therapy does not increase the risk of spinal hematoma associated with regional anesthesia. Anesth Analg. 1995;80(2):303-309. [PubMed]
6.
Horlocker T, Bajwa Z, Ashraf Z, et al. Risk assessment of hemorrhagic complications associated with nonsteroidal antiinflammatory medications in ambulatory pain clinic patients undergoing epidural steroid injection. Anesth Analg. 2002;95(6):1691-7, table of contents. [PubMed]
7.
Benzon H, Wong H, Siddiqui T, Ondra S. Caution in performing epidural injections in patients on several antiplatelet drugs. Anesthesiology. 1999;91(5):1558-1559. [PubMed]
8.
Watanabe M, Qureshi A, Siddiqui F. Periprocedural management of patients on dabigatran etexilate treatment. AJNR Am J Neuroradiol. 2011;32(8):E166. [PubMed]
9.
Horlocker T. Regional anaesthesia in the patient receiving antithrombotic and antiplatelet therapy. Br J Anaesth. 2011;107 Suppl 1:i96-106. [PubMed]
Matthew DeLaney, MD

Matthew DeLaney, MD

Assistant Professor of Emergency Medicine
Assistant Medical Director
University of Alabama at Birmingham
Christopher White, MD

Christopher White, MD

Chief Resident
Department of Emergency Medicine
University of Alabama at Birmingham
Henry Wang, MD

Henry Wang, MD

Professor and Vice Chair for Research
Department of Emergency Medicine
University of Alabama at Birmingham
  • Zlatan Coralic

    Thank you to the authors and all of the reviewers for this excellent article. This is not an uncommon scenario in the ED. Minor points I just want to bring up:

    1. Most of the data summarized stems before the introduction of novel oral anticoagulants, or even before enoxaparin or other low weight molecular heparins in some cases. Broad application of this evidence to non-heparin/non-warfarin patients on anticoagulants should be done with caution, as you point out.

    2. Anticoagulants are metabolized differently. While renal function is important with some injectable low molecular weight heparins, or with novel anticoagulants (you reference dabigatran[8]); renal function is not a predictor of drug clearance for warfarin or heparin. Unfortunately, there is no heuristic approach for determining anticoagulation risk with LP based on clearance, and each agent needs to be considered based on individual pharmacokinetic and pharmacodynamic profiles. The two to five-half live rule may not be applicable to warfarin or renally cleared anticoagulants in patients with renal dysfunction.

    3. All anticoagulants, and specifically novel anticoagulants (dabigatran, apixaban, rivaroxaban) carry a black box warning re: LPs and risk of epidural hematomas. To complicate matters, there is no universally accepted correlate lab to test for the level of anticoagulation of novel anticoagulants. There is some bench data (ECT, or TT for dabigatran), but these labs are not readily available in most hospitals in the United States. With the other agents, coags may be completely normal, but patient would still be at risk for bleeding complications. The burdensome decision of the risk/benefit of an LP falls on the bedside clinician with huge patient harm and medico-legal consequences should complications arise (performing a procedure with a clear black box warning). Even if the risk is 0.3%, the benefits of an LP on a patient who is anticoagulated may not outweigh the risks in an ED setting.

    This is a difficult topic with no clear answers. Thank you again for bringing this important discussion to ALiEM!

    • Josh Guttman

      I agree. I can’t really see a scenario where where benefits outweigh risks in the ED, but would love to hear if such scenario truly exists.