To the Editor:—
The ability to supply oxygen to the patient under anesthesia is vital. We report five cases of sudden cessation of oxygen delivery during general anesthesia.
A 21-yr-old Chinese woman was scheduled to undergo emergency appendectomy. The anesthesia machine (Ohmeda Excel 210; Ohmeda, Madison, WI) was checked according to the protocol of the safety subcommittee of the Singapore Society of Anaesthesiologists. The patient was preoxygenated before rapid sequence induction and tracheal intubation. Mechanical ventilation was started with the oxygen fresh gas flow set at 1%. The isoflurane vaporizer (Isotec 5; Datex-Ohmeda) dial was then set to 1%, after which the oxygen fresh gas flow rapidly decreased to 0. Neither nitrous oxide nor air were available (zero flow when the respective flowmeters’ dials were turned). Values shown by the pipeline and cylinder pressure gauges of the three gases were normal. A self-inflating resuscitation bag was quickly brought in, and the patient underwent manual ventilation with room air. The anesthesia was maintained using intravenous propofol. Her oxygen saturation was maintained between 95 and 100%. Another Excel 210 anesthesia machine was brought in, and anesthesia and surgery proceeded uneventfully. The biomedical engineers inspected the first Excel 210 machine but could not find any fault with it.
Within the year, we had four other similar cases. All machines were checked at the start of anesthesia. All involved the Excel 210 anesthesia machine and Isotec 5 vaporizers. The patients all received uneventful intravenous anesthetic inductions followed by planned isoflurane–nitrous oxide–oxygen mixture maintenance. However, when the isoflurane vaporizer was turned on, the anesthesiologist noticed a rapid decrease in the oxygen and nitrous oxide flows in the flowmeters. The oxygen and nitrous oxide pipeline gauges recorded no change in wall supply pressure. In all cases, the event was quickly discovered, and no patient injuries resulted.
In two of the events, the total failure of fresh gas flow occurred within seconds of turning on the isoflurane vaporizer. In the remaining three cases, the loss of fresh gas flow was more gradual and occurred over 1–2 min. In one of these cases, the anesthesiologist noticed that the isoflurane output was 0% according to the agent monitor, although the vaporizer was turned on and set to 2%. In another case, the anesthesiologist attempted to increase gas flow by turning up the flowmeter dial, which resulted in a more rapid loss of gas supply. In the last of these three cases, the anesthesiologist noticed that fresh gas flow was immediately restored when the isoflurane vaporizer was turned off and replaced with a sevoflurane vaporizer.
The five cases involved five different Ohmeda Excel 210 anesthesia machines and five different Isotec 5 vaporizers. When each involved Isotec 5 vaporizer was replaced with another Isotec 5 vaporizer, normal oxygen delivery was restored. However, when the five “faulty” Isotec 5 vaporizers were mounted on other Excel 210 machines, fresh gas delivery failure was replicated.
During close examination of the “faulty” Isotec 5 vaporizers, we found that the two actuating spindles on the back of the vaporizer were not of equal length. The function of the actuating spindle is to divert fresh gas flow from the Select-a-Tec vaporizer manifold of the Excel into the vaporizer when the vaporizer is turned on, even when it is set at 0%. The actuating spindles act by depressing the ball valve system in the vaporizer manifold of the Excel machine. When the spindle in the upstream side was excessively long, the ball valve in the vaporizer manifold was completely depressed, and the fresh gas flow pathway into the vaporizer became occluded (fig. 1). The complete obstruction of fresh gas flow resulted in back pressure. Equalization of the pressure difference across the flowmeter bobbin occurred, and the output decreased to 0. Gas delivery through the flowmeter then ceased. Our biomedical engineers tested the postulated cause of the fresh gas flow failure due to the length of the upstream actuating spindle in the “faulty” vaporizer by adding O-rings around the port valve. The inserted O-rings increased the upstream level of the vaporizer on the back bar and compensated for the excessive length of the affected actuating spindle. The vaporizer then functioned normally.
When the longer actuating spindle was the downstream spindle, there was correct diversion of the fresh gas flow into the vaporizer. However, gas return from the vaporizing chamber to the vaporizer manifold was blocked. Fresh gas then leaked from the port valve into the atmosphere, causing a slower increase in back pressure. This resulted in a slower loss of fresh gas flow (fig. 2). When additional O-rings were added to increase this downstream valve port slightly, the gas leak lessened. This was associated with faster buildup of back pressure and cessation of fresh gas flow. When sufficient O-rings were added to elevate the vaporizer position to completely compensate for the excessively long downstream spindle, the back pressure effect and gas flow delivery were again normalized.
Intraoperative awareness has been reported as a result of malfunctioning of the vaporizer attached to the Siemens 900B ventilator 1(Siemens-Elema, Sweden) and also on the Select-a-Tec Vaporizing System. 2Riendl 3described a case in which a hypoxic gas mixture was delivered to the patient as a result of a faulty spring in the Ohmeda Modulus I gas machine. However, there are no reports of total fresh gas delivery failure resulting from a fault in the vaporizer system. All the vaporizers involved in our description were approximately 7 yr old and had been maintained on a regular basis by the Australian Datex-Ohmeda service center. Our suspicion that the fault was caused by excessively long actuating spindles was later confirmed by Datex-Ohmeda in four of the vaporizers. Datex-Ohmeda found that the locknuts and lock-tites holding the top of the actuating spindle in its proper place were loose or missing in the affected vaporizers. This allowed the actuating spindle to migrate downward, functioning effectively as “excessively long” spindles. The actual spindles were, in fact, of normal length. This could have occurred when the locknuts became old or when they were not replaced or tightened adequately after a routine service as reported by Datex-Ohmeda. In two of the vaporizers, the fault was in the upstream spindle. In one vaporizer, the problem involved the downstream spindle, and in the other, the exact spindle was not specified.