Variability in oxygen delivery with bag-valve-mask devices: An observational laboratory simulation study

A bag-valve-mask (BVM) is a portable handheld medical device commonly used in airway management and manual ventilation. Outside of the operating theatre, BVM devices are often used to pre-oxygenate spontaneously breathing patients before intubation to reduce the risk of hypoxaemia. Pre-oxygenation is considered adequate when the end-tidal expiratory fraction of oxygen is greater than 0.85. There are reports that some BVM devices fail to deliver a satisfactory inspired oxygen (FiO(2)) in spontaneously breathing patients due to variability in design. The primary aim of this study was to evaluate the efficacy of oxygen delivery of a broad range of adult and paediatric BVM devices at increasing tidal volumes using a mechanical lung to simulate spontaneous ventilation. The secondary aim was to evaluate the effect of BVM design on performance. Forty BVM devices were evaluated in a laboratory setting as part of a safety assessment requested by HealthShare New South Wales. The oxygen inlet of each BVM device was primed with 100% oxygen (15 l/min) for two min. The BVM device was then attached to the mechanical lung and commenced spontaneous breathing at a fixed respiratory rate of 12 breaths/min with an inspiratory: expiratory ratio of 1:2. For each device FiO(2) was measured after two min of spontaneous breathing. This process was repeated with small (250 ml), medium (500 ml) and large (750 ml) tidal volumes simulating adult breathing in adult BVM devices, and small (150 ml), medium (300 ml) and large (450 ml) tidal volumes simulating paediatric breathing in paediatric BVM devices. The test was repeated using up to five BVM devices of the same model (where supplied) at each tidal volume as a manufacturing quality control measure. Eight of the 40 devices tested failed to deliver a FiO(2) above 0.85 for at least one tidal volume, and five models failed to achieve this at any measured tidal volume. Concerningly, three of these devices delivered a FiO(2) below 0.55. Six of the eight poorly performing devices delivered reducing concentrations of inspired oxygen with increasing tidal volumes. Devices which performed the worst were those with a duckbill non-rebreather valve and without a dedicated expiratory valve. Several BVM devices available for clinical use in Australia did not deliver sufficient oxygen for reliable pre-oxygenation in a spontaneously breathing in vitro model. Devices with a duckbill non-rebreather valve and without a dedicated expiratory valve performed the worst. It is imperative that clinicians using BVM devices to deliver oxygen to spontaneously breathing patients are aware of the characteristics and limitations of the BVM devices, and that the standards for manufacture are updated to require safe performance in all clinical circumstances.