Secondary Logo

Journal Logo

Clinical Letter

Delivery of an hypoxic gas mixture due to a defective rubber seal of a flowmeter control tube

Hay, H.

Author Information
European Journal of Anaesthesiology: July 2000 - Volume 17 - Issue 7 - p 456-458



The delivery of an hypoxic gas mixture from an anaesthetic machine is today a less common event due to the many safety features that have been incorporated in currently used equipment. These include:

• Colour coding of flexible hoses and specificity of probes and collars.

• Specificity of connections within the anaesthetic machine.

• Arrangement of the common gas manifold so that oxygen joins downstream of the other gases.

• The pin indexing system of gas cylinders.

• Non-return valves and blanking caps for cylinder yokes [1].

• Oxygen failure warning devices, linked to a means of flow control to shut off the flow of nitrous oxide and open the system to air.

• Monitoring the inspired gases with both a fuel cell electrode and a paramagnetic analyser.

• Oxygen Ratio Controllers whereby any inadvertant attempt by the anaesthetist to deliver a hypoxic mixture to the patient is prevented.

In spite of these measures incidents where hypoxic gas mixtures are delivered to the patient continue to occur. Defective nitrous oxide/oxygen gas mixers can lead to the retrograde pollution of the oxygen supply with nitrous oxide [2]. A defective gas loaded regulator can allow communication between the oxygen and nitrous oxide pipelines and the flow of oxygen through a nitrous oxide flowmeter. This has led to a case of awareness under anaesthesia [3]. Oxygen contaminating a nitrous oxide pipeline has been reported resulting from a perforation in the diaphragm of an oxygen failure alarm [4]. In both these cases an hypoxic gas mixture could have resulted if the pressure gradient between the pipes had been reversed. The oxygen failure devices and nitrous oxide shut-off valve are only effective if oxygen pressure from the gas supply is low. Improvements have been made with oxygen ratio systems, preventing the delivery of hypoxic gas at fresh gas flows higher than 1 L min−1[5].


A rare form of oxygen failure is described which resulted from a defective rubber seal at the top (downstream) end of a Rotameter® flow tube on an Ohmeda Selectatec SM anaesthetic machine. The Rotameter® barrels (tubes) are seated in rubber sleeves, which form a push-on gas-tight seal at the top and bottom of the barrels. They are not visible to the user without removing the surrounding plate. A critical incident arose because the upper seal on the oxygen flow tube was cracked. Routine testing did not expose this defect and the flowmeter appeared to be operating normally. An hypoxic gas mixture was delivered to the patient.

The incident proceeded as follows. Our anaesthetic machines are routinely tested prior to use. Pressure occlusion tests were performed at the common gas outlet (CGO) and at both the inner and the outer coaxial tubes of a Bain-type breathing system, which was to be used. No leaks were demonstrated. A 41-year-old-lady was due to undergo a percutaneous nephrolithotomy; she smoked 20 cigarettes per day but was otherwise in good health with no respiratory problems. After the uneventful induction of anaesthesia and tracheal intubation, the patient was transferred to the operating theatre and positioned in the prone position. Her lungs were mechanically ventilated using a Bain breathing system and a Nuffield Penlon 200 ventilator. Anaesthesia was maintained with isoflurane and a gas mixture initially comprising oxygen 2 L min−1 and nitrous oxide 4 L min−1. The inspired gases were monitored using a paramagnetic oxygen analyser (Ohmeda RGM 5250, Ohmeda, Louisville, CO 80027, USA) which had been appropriately calibrated. A fuel cell electrode oxygen analyser (Ohmeda 5120) was positioned close to the CGO; the gas sampling line for the paramagnetic oxygen analyser was attached to the heat and moisture exchange filter close to the endotracheal tube.

The flowmeter control valves indicated oxygen 2 L min−1 and nitrous oxide 4 L min−1 but almost immediately an alarm sounded because the paramagnetic oxygen analyser had detected an inspired O2 concentration of 18% at the patient's end of the breathing system, and the fuel cell recorded an O2 concentration of 8% at the common gas outlet. The inspired concentration of N2O was measured as 70% by the gas analyser, which was accepted as being within reasonable limits of error. The patient looked well and pulse oximetry showed an oxygen saturation of 95%. There was no disconnection between the breathing system and the anaesthetic machine and the patient's lungs appeared to ventilate satisfactorily. It was then assumed that there must be a fault with the monitor and a newly calibrated monitor of the same type accordingly replaced it, but exactly the same readings were obtained. It was therefore assumed that the machine was delivering an hypoxic gas mixture despite evidence to the contrary from the flowmeter control valves and the satisfactory condition of the patient. In an attempt to give the patient 100% oxygen the N2O flowmeter was closed and the oxygen flow increased to 8 L min−1. Both the fuel cell and the paramagnetic analyser recorded oxygen concentrations of over 90%; pulse oximetry showed a rise in oxygen saturation to 98%.

The operation had by now reached a stage where it would have been very difficult to rapidly abandon the procedure so a new anaesthetic machine was procured and the defective machine taken out of service. There was now no discrepancy between the indicators of the flowmeters and the readings from the O2 analyser. The operation concluded uneventfully and the patient made a good recovery. Remarkably the defective machine subsequently passed a pressure test by the hospital's medical engineering department in which a pressure gauge was attached to the common gas outlet, the oxygen flowmeter was opened and the pressure rose to 34 kPa at which point oxygen was vented by the pressure relief valve. The oxygen flowmeter was closed and the pressure in the system appeared to be maintained. The anaesthetist insisted that the manufacturer's maintenance engineers were called in: when they dismantled the flowmeter tubes, the cracks in the rubber seal were revealed (Fig. 1). Had it been the lower seal that leaked, below the Rotameter® bobbin, then the flowmeter reading would have approximated to the gas flow supplied. A more dangerous situation arose because the leak was above the bobbin in the flowmeter tube; although oxygen appeared to be delivered at 2 L min−1, more than 20% was leaking out into the atmosphere before reaching the common gas manifold. The leak was also downstream from the oxygen failure device so the oxygen failure alarm did not sound.

Fig. 1
Fig. 1:
Rubber seal from flowmeter tube showing split.


This incident demonstrates that anaesthetic equipment remains fallible, however sophisticated it may be, and that the monitoring equipment should not be assumed to be wrong when unexpected readings are obtained.


1 McQuillan PJ, Jackson IJ. Potential leaks from anaesthetic machines through open Rotameter valves and empty cylinder yokes. Anaesthesia 1987; 42: 1308-1312.
2 Otteni JC, Ancellin J, Cazalla JB. Defective gas mixers, a cause of retro-pollution of medical gas distribution pipelines. Ann Fr Anaesth Reanim 1997; 16: 68-72.
3 Puri GD, George MA, Singh H, Batra YK. Awareness under anaesthesia due to a defective gas-loaded regulator. Anaesthesia 1987; 42: 539-540.
4 Lye A, Patrick R. Oxygen contamination of the nitrous oxide pipeline supply. Anaesth Intensive Care 1998; 26: 207-209.
5 Rathgeber J. Anaesthesia equipment and respirators. Fundamentals. Anaesthesist 1993; 42: 885-909.


© 2000 European Academy of Anaesthesiology