The Case

A 58-year-old man, scheduled for aortoiliac artery bypass graft, had an epidural catheter placed for postoperative pain management. Surgery proceeded uneventfully under general anesthesia. During the closure of the surgical incision, the surgery fellow drew 12 mL of 0.25% bupivacaine (a local anesthetic) into a 12 mL labeled syringe of which 4 mL was injected into the epidural space. The fellow subsequently placed the bupivacaine syringe with the remaining 8 mL of the drug in the pocket of his scrub suit.

Anticipating the conclusion of surgery, the fellow prepared a second 12 mL syringe drawing up 4 mL of neostigmine (4 mg) and 4 mL of glycopyrrolate (0.8 mg) totaling 8 mL and labeled it, adding this syringe to the same pocket of his scrubs. At the conclusion of the surgical procedure, the attending asked the fellow to reverse the neuromuscular block with neostigmine and glycopyrrolate. The fellow pulled a syringe out of his pocket, assumed it contained the neostigmine and glycopyrrolate, and injected 6 mL and then placed the syringe back in his pocket.

After 3 minutes, the patient still appeared weak (i.e., with residual neuromuscular blockade), and the attending requested administration of an additional 1 mg of neostigmine. When the fellow retrieved a syringe from his pocket, he recognized that he had previously pulled the wrong syringe and inadvertently administered bupivacaine, the anesthetic, rather than neostigmine/glycopyrrolate. Once the correct medication was administered, neuromuscular blockade was reversed. Although the patient was not harmed by the erroneous intravenous (IV) administration of bupivacaine, he potentially could have been.

The Commentary

At first glance, the problem represented by this case seems trivially obvious. A physician, choosing between only two syringes, both known to contain high-risk drugs, didn't bother to read the label before casually injecting the wrong one by the wrong route. Only luck saved the patient from the well-known cardiotoxic effects of IV bupivacaine.

Anesthesiologists have known for more than 30 years that drug administration errors are among the most common adverse events in anesthesia (1), and the "syringe swap" is the most common of those errors.(2) Yet they still occur with unsettling regularity. And it's not just the occasionally incompetent clinician who commits these errors: in anonymous surveys, a majority of anesthesiologists admit to having made at least one significant medication error.(2) Thus, it seems clear that there are systemic factors inherent to the anesthesia system of drug administration that contribute to the persistence of the problem. Blaming individuals for character defects when things go wrong, the so-called person-centered approach, denies the reality that errors are inevitable, and directs attention away from potentially correctable latent system problems.

Despite the ubiquity of drug errors in anesthesia, there is little reliable evidence regarding the actual rates of these errors. Methodological difficulties continue to confound studies of drug errors. There is no agreement, for example, on the definition of the term "error," even at the most basic level. In some studies, an error is counted only if there is identifiable patient harm (3); in others an error is any discrepancy between the intended drug and the one actually delivered.(4) By the first definition, the present case, which resulted only in a slight delay in the reversal of neuromuscular blockade, might be judged as having no error! Moreover, the methods for identifying errors vary widely, and often are not clearly specified. It is well known that direct observation, chart review, and self-generated incident reports find error rates that differ by more than two orders of magnitude.(5) Direct observation is generally considered to be the most accurate method, although the most expensive and invasive.

Although there is no consensus concerning measurement of anesthetic error rate, and published studies are difficult to compare, the risk of a medication error is probably on the order of 1 error per 100 anesthetics.(6) It should be noted that an "anesthetic" consists of the administration of multiple drugs, by multiple routes over an extended time period. Of these errors, the rate of injury, and more rarely, death, is probably also approximately 1%.(2) Thus, the risk of injury or death from medication error per se is approximately 1 in 10,000 anesthetic cases.

Comparison between anesthesia medication errors and drug errors in other settings is difficult because of the unique nature of the anesthesia task. Anesthesiologists prescribe, prepare, administer, and document medications without any of the checks and balances routinely employed on hospital wards and in intensive care units (ICUs). Most of the drugs used by anesthesia practitioners are administered by IV push or IV infusion, and many of them fall into high-risk categories that, in an ICU for example, would require special handling. At least part of the reason that bar coding, pharmacy review, and other safety checks are not used by anesthesiologists is the unstable nature of the anesthetic state. Changes in anesthetic agents and associated adjuvant drugs occur at a rate far too high for available safety systems to keep up.

Without reliable knowledge of the relative frequency of the different types of medication error, and without randomized controlled trials to support evidence-based recommendations, it is difficult to make recommendations for improving anesthesia drug practice.(7) Nevertheless, a recent consensus conference convened by the Anesthesia Patient Safety Foundation (8) has issued a set of recommendations intended to improve safety. They include standardization of drug concentrations, barcode identification of medications, computerized documentation of administration, standardized anesthesia carts, and the addition of a clinical pharmacist to the operating room (OR) team. Finally, the committee strongly advocated establishing a "just culture" for reporting errors, cooperation, and education. Of these recommendations, only barcode identification of medications might have prevented the error in this case, and it is not clear that portable barcode readers of the sophistication required have yet been developed.

The present case involved the most frequently occurring drug error in anesthesia incident report analyses: the substitution of one syringe for another, even though the syringes, as in this case, are properly labeled. We are not informed as to other possibly distracting events occurring at the time the error occurred. It seems to have been just as the patient was emerging from general anesthesia, usually a busy time in the OR, with preparation being made to move the patient to the transport gurney, disconnection of monitoring equipment, cleanup of instruments, and many staff members moving in and out of the OR.

Proposals intended to reduce syringe swap errors include improvements in syringe labeling, standardized color coding of labels according to class of drug, barcoded labels accompanied by mandatory reading of the barcode by a device that pronounces the contents of the syringe, use of prefilled barcoded syringes rather than having the anesthetist draw up and label drugs on a case-by-case basis, connection of the barcode reader to the hospital's pharmacy computer system that would report a proposed dose and warn of any potential contraindications, and even having an assistant stand by during the case to perform a second assessment of every drug before it is given.(8-10)

Of these suggestions, improved labeling has been widely though not universally adopted. The use of standardized pharmacy- or commercially prepared barcoded prefilled syringes is just beginning to be adopted after significant delays associated with determining the stability and sterility of drugs packaged in syringes. Mass-produced prefilled syringes would essentially eliminate the problem of mislabeled syringes, and probably would be worth the extra cost on that ground alone. Barcode readers are just beginning to be introduced; barcode readers linked to central pharmacy computers are not yet commercially available. There is no evidence clearly supporting the effectiveness of any of these measures, but all seem reasonable. Providing a pharmacy assistant for every anesthesiologist would be extremely expensive; therefore it would need clearcut evidence of its effectiveness, which is at present a subject of some controversy.(10)

The current case raises another issue: the drug accidentally injected IV, bupivacaine, had been intended for delivery via the epidural route. Many different tubing devices used in patient care, including epidural catheters and ordinary syringes share the common Luer connector system. Accidental misconnection between incompatible systems has been known for years, with sometimes fatal consequences. The problem of cross-connection of anesthesia gasses was recognized more than 50 years ago, and has been almost eliminated by the mandated use of incompatible connectors for the different gasses. Similar incompatible connector designs for tubing that should never be cross-connected (i.e., an epidural anesthetic connected to an IV line) have been designed and clinical testing is ongoing at present. It is expected that commercial devices will be available sometime in 2013.(11)

Finally, it has been suggested that the mere act of having the syringe in a pocket somehow contributed to this mishap. I have made a survey of convenience of anesthesiologists of my acquaintance, of many different ages and subspecialties. I asked three questions: Do you ever carry drugs in your pocket? Does every anesthesiologist you know carry drugs in a pocket? Do you think it's safe to carry drugs in a pocket? The answer to all three questions was unanimously "yes." A review of the anesthesiology literature also showed no papers even addressing the issue. Most anesthesiologists believe that during the transfer of relatively unstable patients from the OR to the recovery room, safety factors require the immediate availability of preloaded syringes of suitable drugs. These patients are just emerging from anesthesia, and are notoriously subject to several abrupt physiological derangements, from hypotension to loss of airway, demanding a near-instantaneous response. Moreover, many of the anesthesiologists questioned pointed out that the syringe has to be somewhere; if not in the pocket, then where else? It's not possible to transfer the entire anesthesia workspace with a patient for the short trip to recovery.

In this case, two technical errors occurred that are beyond the scope of this piece. First, the neuromuscular blockade should have been shown, quantitatively, to have been reversed before the patient was moved from the OR. Second, there was no need to carry bupivacaine along to the postanesthesia care unit; need for more bupivacaine is never an emergency. However, the primary point here is that syringe swaps do occur. We know from other studies of medication safety that, even with better labeling, such a system will still be prone to errors.

Take-Home Points

This case illustrates several points about drug safety in anesthesia:

• Anesthesia drug errors are ubiquitous. They happen to almost every anesthesiologist sooner or later.
• The frequency of drug errors in anesthesia is not well known, but it is probably much higher than has been reported. It will be difficult to determine the effects of proposed improvements in drug administration systems without much more robust measurements of medication administration error rates.
• Reduction in drug errors will not occur without careful attention to the system within which anesthesia drugs are delivered. It is likely that many of the errors result from poorly designed drug delivery systems.
• Anesthesiology personnel should be aware that they are, themselves, somewhat unreliable elements in a system that is not particularly resilient. As a result, it is imperative to maintain the highest possible level of alertness to possible drug errors.
• Standardization and simplification of anesthesia medication regimens may contribute to improvements in drug safety in anesthesia.

John C. Kulli, MD Medical Director

Perioperative Services

University of California, San Francisco

References

1. Cooper JB, Newbower RS, Long DC, McPeek B. Preventable anesthesia mishaps: a study of human factors. Anesthesiology. 1978;49:399-406. [go to PubMed]

2. Orser BA, Chen RJ, Yee DA. Medication errors in anesthetic practice: a survey of 687 practitioners. Can J Anaesth. 2001;48:139-146. [go to PubMed]

3. Ferner RE. The epidemiology of medication errors: the methodological difficulties. Br J Clin Pharmacol. 2009;67:614-620. [go to PubMed]

4. Wheeler SJ, Wheeler DW. Medication errors in anaesthesia and critical care. Anaesthesia. 2005;60:257-273. [go to PubMed]

5. Flynn EA, Barker KN, Pepper GA, Bates DW, Mikeal RL. Comparison of methods for detecting medication errors in 36 hospitals and skilled-nursing facilities. Am J Health-Syst Pharm. 2002;59:436-446. [go to PubMed]

6. Webster CS, Merry AF, Larsson L, McGrath KA, Weller, J. The frequency and nature of drug administration error during anaesthesia. Anaesth Intensive Care. 2001;29:494-500. [go to PubMed]

7. Jensen LS, Merry AF, Webster CS, Weller J, Larsson L. Evidence-based strategies for preventing drug administration errors during anaesthesia. Anaesthesia. 2004;59:493-504. [go to PubMed]

8. Eichhorn JH. APSF hosts medication safety conference: consensus group defines challenges and opportunities for improved practice. APSF Newsletter. 2010;25:1-20. [Available at]

9. Poon EG, Keohane CA, Yoon CS, et al. Effect of bar-code technology on the safety of medication administration. N Engl J Med. 2010;362:1698-1707. [go to PubMed]

10. Evley R, Russell J, Mathew D, Hall R, Gemmell L, Mahajan RP. Confirming the drugs administered during anaesthesia: a feasibility study in the pilot National Health Service sites, UK. Br J Anaesth. 2010;105:289-296. [go to PubMed]

11. Preventing catheter/tubing misconnections: Much needed help is on the way. ISMP Medication Safety Alert! Acute Care Edition. July 15, 2010;15:1-2. [Available at]