To the Editor:—

Patient-controlled analgesia (PCA) machines are used commonly for postoperative pain management, largely because of their tremendous clinical benefits. A typical PCA machine contains an embedded computer that is programmed using a keypad to give, for example, 1 mg morphine every time the patient pushes the analgesic demand push button (pain pendant). (This is a push button on the end of a cable that usually looks and feels much like a nurse call button.) To help prevent excessive drug administration, the on-board computer ignores further patient demands until a “lockout” period has passed. This period usually is set at between 5 and 10 min.

Unfortunately, misadventures have been attributed to inadvertent misprogramming or other problems associated with these devices. 1–8We report herein a case in which intermittent short circuits from frayed wires in the pendant resulted in a drug unintentionally being delivered automatically.

A 62-yr-old retired grain inspector was diagnosed with carcinoma of the prostate after a transurethral resection of prostate (TURP) procedure and subsequently presented to the hospital for radical prostatectomy. His previous medical health had been unremarkable except for a 20-yr history of smoking and remote appendectomy. He took no regular medications and had no allergies. Routine preoperative laboratory studies were unremarkable except for some unifocal ventricular premature beats on the electrocardiograph.

The anesthesia and surgery were uneventful, as was his stay in the postanesthesia care unit, where he was given an Abbott Lifecare 4100 PCA Plus II PCA machine (Abbott Laboratories, Abbott Park, IL) that was programmed to deliver conditionally 1 mg morphine intravenously each time the pain pendant was activated. The lockout period was set at 6 min. After his stay in the postanesthesia care unit, the patient was admitted to the general urology ward, where he spent an uneventful night.

Late the following morning, the nursing staff was alerted by the patient’s wife, who indicated that she was unable to wake her husband. The nurses found the patient to be “unresponsive, diaphoretic, and with labored breathing.” A “code blue” was called. The responding team noted the following vital signs: blood pressure, 152/70 mmHg; heart rate, 62 beats/min; and respiratory rate, 8 breaths/min. The pupils were noted to be miotic. Pulse oximetry was not immediately available. Naloxone, 0.4 mg, was administered intravenously, resulting in a marked clinical improvement. No harm came to the patient. The patient subsequently was placed on close observation, and PCA was discontinued. The PCA machine was sent to the medical engineering department for inspection.

The medical engineering technicians noted that the wires in the pain pendant were frayed and determined that the machine would register analgesic demands without the push button being depressed. They concluded that the frayed wires in the pendant were short-circuiting intermittently, creating a false triggering situation. The defective pendant was replaced.

Numerous clinical problems with PCA technology have been identified. 1–8Some of these problems are caused by human error, with some of these errors being a consequence of a hostile user interface (unforgiving and unintuitive). 5,9,10 

However, in our situation, machine failure, not human error, was the source of the problem. We note that a partial solution for this problem (beyond mere periodic inspection and testing of the push button) would be for the manufacturer to add a software safety self-check routine in the embedded microcomputer in the PCA pump to check the push button signal line for lengthy periods of continuous assertion (e.g. , for more than 10 s). As well, such safety software could also check for instances in which the interval between the end of the lockout period and the next analgesic demand is unreasonably short (such as less than 100 ms). Such solutions require only a software upgrade in existing PCA machines. Another partial solution applicable to new designs is to also use a double-pole, single-throw push button and two signal lines, with software to check that the lines always carry an identical signal. Finally, we emphasize that by disseminating these safety concepts widely through this communication, we are making the idea unpatentable and thus freely available for all to use at no cost.

Jennings N: Patient-controlled analgesia: A serious incident. Anaesthesia 1992; 47: 1093–4
Notcutt W: Overdose of opioid from patient controlled analgesia pumps (letter). Br J Anaesth 1992; 68: 450
White E: A problem with the Graseby 3300 PCA pump. Anaesthesia 1993; 48: 1013–4
Kwan A: Overdose of morphine during PCA (letter). Anaesthesia 1995; 50: 919
ECRI : Abbott PCA Plus II patient-controlled analgesia pumps prone to misprogramming resulting in narcotic overinfusions. Health Devices 1997; 26: 389–91
Christie L, Cranfield KA: A dangerous fault with a PCA pump (letter). Anaesthesia 1998; 53: 827
Ma M, Chen PP, Chan S, Chung D: A potential PCA hazard (letter). Anaesthesia 1998; 53: 314
Chan S, Chen PP, Chui PT, Ma M: Unintentional bolus with Graseby 9300 pump (letter). Anaesth Intensive Care 1998; 26: 117
Brown SL, Bogner MS, Parmentier CM, Taylor JB: Human error and patient-controlled analgesia pumps. J Intraven Nurs 1997; 20: 311–6
Lin L, Isla R, Doniz K, Harkness H, Vicente KJ, Doyle DJ: Applying human factors to the design of medical equipment: patient-controlled analgesia. J Clin Monit Comput 1998; 14: 253–63