Cases & Commentaries
Check the Anesthesia Machine
A 62-year-old man with weight of 134 kg (body mass index [BMI] of 40) and history of hypertension, diabetes, sleep apnea, claustrophobia, and 3-vessel coronary artery disease was scheduled for elective coronary artery bypass surgery. Once in the operating room, the resident placed an arterial line in the patient's left arm under local anesthesia. Because the patient was unable to tolerate the facemask oxygen (due to claustrophobia), the attending anesthesiologist gave him the anesthetic circuit, which had oxygen flowing at 10 liters per minute and asked him to keep it in his mouth and breathe through it (like snorkeling) to achieve pre-oxygenation. The anesthesiologist then injected 2 mg of midazolam to sedate the patient.
After about 5 minutes, the anesthesia team noticed that the patient was unresponsive, with shallow breathing. Breathing was assisted with facemask ventilation and the airway was secured with endotracheal tube (after propofol, fentanyl, and rocuronium rapid induction). Once anesthesia was induced, the anesthesiologist tried to turn on the anesthetic agent and noticed that the desflurane vaporizer was set to 12%. It was then discovered that the patient had inadvertently received 12% desflurane (a general anesthetic) instead of oxygen alone during pre-oxygenation. The logbook on the anesthesia machine showed that the machine had been checked that morning, but the resident had failed to notice the open desflurane vaporizer. The patient did not experience any obvious harm from the uncontrolled inhalation induction.
This case can be called a medication error that occurred at the beginning of an anesthetic. The term medication error in anesthesia usually brings to mind images of insulin overdoses, incorrect heparin concentration administrations, or infusion of a bag of potassium-containing normal saline instead of plain normal saline. With anesthetics, when volatile anesthetics are inhaled, administering oxygen that inadvertently contains a high concentration of desflurane instead of plain oxygen is also a medication error.
Medication error is among the most common type of errors in anesthesia; they have been studied for more than 30 years. For most medications to reach a patient, a caregiver must order it, pharmacy must review and dispense it, and a nurse must administer it—a triple-check system, albeit an imperfect one. In the operating room, the anesthesiologist fulfills all of these roles. Moreover, while most guidelines to minimize drug errors recommend that medications be prepared in an environment free from distractions (1), anesthetic drugs are decided on, obtained, and administered in conditions that are prone to distraction. This case is a good illustration. Distracting factors included the patient's comorbid disease, claustrophobia, inadequate intravenous access, morbid obesity, and the induction itself. The more frail a patient, the more likely are both distractions and harm from medication errors.
Despite an increasing focus on medication errors in anesthesia, the rate of such errors is not well established. Reported medication error rates differ greatly based on both the detection method (e.g., self-reporting, direct observation) and the definition of medication error used. For example, must a medication error include identifiable harm? Does incorrect timing constitute an error? In fact, some definitions would not have included the present case. Although the studies are difficult to compare, current estimates put medication error rates at 1 in 100 anesthetics.(2) Injury and death from such errors are estimated to occur with 1% of errors.(3) By extrapolation, then, the risk of injury or death from medication errors may be approximately 1 in 10,000 anesthetics. We expect that the relatively new program for error reporting in anesthesia by the Anesthesia Quality Institute (4) will soon yield better information about how often such events occur.
The dial of a vaporizer controls the concentration of anesthetic vapor (volatile anesthetic) that is mixed with "fresh gas" (usually oxygen or air) and then delivered to the patient. Each type of volatile anesthetic has an agent-specific vaporizer. Before intravenous induction, the dial should always be set to zero. The pre-induction checklist should include this check. In this case, the desflurane concentration was somehow set to 12%, which led to the inadvertent inhalational induction of general anesthesia. (Note that "inhalational inductions" are occasionally intentionally performed with sevoflurane, but almost never with desflurane because of its pungency and the notable airway irritation that it produces.) After administration of intravenous induction agents and securing the airway, the anesthesiologist went to turn on the vaporizer to maintain anesthesia (as the administered intravenous agents would typically only last 5–10 minutes). At this point, the error was detected.
This patient and anesthesiologist were fortunate that the error was promptly detected, because serious harm and even death could have occurred. A high concentration of desflurane could cause severe cardiac depression, airway reactivity, aspiration (before securing the airway), obstruction, and/or loss of airway prior to intubation—any of which could lead to a fatal outcome, even if recognized and treated expeditiously. The American Society of Anesthesiologists Closed Claims Project recently presented an update on patient injuries from anesthesia delivery equipment that resulted in settled malpractice cases.(5) Of the 39 gas delivery equipment claims since 1990, anesthesia vaporizers were involved in 14 (35%). In vaporizer claims, the most common harm was awareness or patient movement during surgery due to light anesthesia. Reasons included failure to turn on the vaporizer due to lack of familiarity with equipment or memory lapse, failure to notice the vaporizer was empty, incorrect mounting of vaporizer, vaporizer malfunction, and vaporizer leak. Three claims involved volatile anesthetic agent overdose, one of which resulted in severe brain damage.
The inadvertent administration of desflurane or similar anesthetics is normally detected in one of the following ways: use of a pre-induction checklist (see below for more on checklists), smell (this may not have been noticeable since the patient was "snorkeling" instead of using a mask), patient discomfort with noxious gas (this could have been blunted by the midazolam or because the patient didn't know that what he was experiencing was not normal), and agent monitoring (available on many anesthesia machines but not always used). The holes of the Swiss cheese (6) aligned in this event, allowing all of these mechanisms to fail. The final defense that led to detection appears to have been fortuitous.
We do not really know all that happened in this case. How did the vaporizer get turned on? Perhaps it was done during an anesthesia machine check, during which some advise that the vaporizer be turned on briefly to check the low pressure system of the anesthesia machine for leaks. Or, perhaps an anesthesia technician turned it on when refilling it, erroneously thinking that the dial had to be opened for filling. Trying to ascertain which of these errors was at play is important, since each would lead to a different intervention to prevent a potential future accident.
In the future, perhaps anesthesia machines will use electronic vaporizing techniques that have a default setting of "off" for all vaporizers pre-induction. This type of system was demonstrated (7) in 1978 but is not so easy to design into a failsafe system and thus is not typical of vaporizers, particularly in the United States (some such designs are more widely used in Europe). However, this type of forcing function only addresses one specific error, while overdoses of inhalational anesthetics could be administered in other ways, thus bypassing such a safety system.
Agent (volatile gas) concentration monitoring may have alerted the anesthesiologist to this event more quickly, but the use of alarms for high concentrations are not ubiquitous, and, even when they are used, anesthesiologists don't always pay full attention to the monitors. An oxygen analyzer would have indicated a lower-than-expected concentration of oxygen (i.e., maximally 88% instead of 100%, given 12% desflurane), which could have been an earlier tipoff. But, the distraction of the unexpected change of patient responsiveness would have led the anesthesiologist to focus on a quick intubation, rather than a scan of the machine and physiologic information.
Perhaps the most effective prevention measure for this type of error is the use of a preoperative machine checklist. Interestingly, while checklists of various kinds are widely discussed today (8), it was the anesthesia machine checklist that was the very first suggested in health care.(9) Many machine checklists in existence would have detected this error.(10,11) Unfortunately, checklists still are not used in every case and shortcuts are common, likely related to production pressure and the low yield of each individual step. Yet, we believe that checking all vaporizers before every anesthetic should be as instinctive as putting on a seatbelt. While the idea of a more sophisticated monitoring system is seductive, a simple checklist might have prevented this error.
- Medication error is among the most common type of error in anesthesia, with incidence estimates at 1 in 100 anesthetics.
- Medication error etiology is multifactorial and includes a distracting environment and production pressure that limit safety check compliance.
- A preoperative checklist that includes check of the vaporizer and is performed in an environment free from distractions is perhaps the most effective means to prevent this type of medication error.
Daniel Saddawi-Konefka, MD Instructor in Anesthesia Department of Anesthesia, Critical Care and Pain Medicine Massachusetts General Hospital Harvard Medical School Boston, MA
Jeffrey B. Cooper, PhD Professor of Anesthesia Department of Anesthesia, Critical Care and Pain Medicine Massachusetts General Hospital Harvard Medical School Boston, MA
1. Preventing and reporting drug administration errors. NursingTimes.net. August 16, 2005. [Available at]
2. 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]
3. 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]
4. Anesthesia Quality Institute. Park Ridge, IL; 2013. [Available at]
5. Mehta SP, Eisenkraft JB, Posner KL, Domino KB. Patient injuries from anesthesia gas delivery equipment: a closed claims update. Anesthesiology. 2013;119:788-795. [go to PubMed]
6. Reason J. Managing the Risks of Organizational Accidents. Aldershot, England: Ashgate Publishing Limited; 1997. ISBN: 9781840141054.
7. Cooper JB, Newbower RS, Moore JW, Trautman ED. A new anesthesia delivery system. Anesthesiology. 1978;49:310-318. [go to PubMed]
8. Gawande A. The Checklist Manifesto: How to Get Things Right. New York, NY: Metropolitan Books; 2009. ISBN: 9780805091748.
9. Cooper JB, Newbower RS, Kitz RJ. An analysis of major errors and equipment failures in anesthesia management: considerations for prevention and detection. Anesthesiology. 1984;60:34-42. [go to PubMed]
10. 2008 American Society of Anesthesiologists Recommendations for Pre-Anesthesia Checkout. [Available at]
11. Anesthesia Apparatus Checkout Procedure. [Available at]