The importance of characterising dynamic response and inertia in potential rapidly manufactured ventilator systems.

In response to the COVID‐19 pandemic, the UK Government developed a specification of a ‘minimally acceptable’ rapidly manufactured ventilator system (RMVS) [1]. This specification was highly regarded, and other countries, including Australia [2], used it as the basis for their documents. Many teams around the world began working to close the projected gap in ventilator availability and used the RMVS specification as the starting point for their efforts.

Notably, the RMVS specification was revised after many teams had commenced work to emphasise the importance of supporting spontaneous breathing. This change obviously introduced the need to trigger both inspiration and expiration, but it also introduced more challenging requirements that many teams did not notice.

We collated [3] the informally published design approaches outlined by more than 100 of these teams. Whereas noting many teams have not yet modified their approach to accommodate spontaneous breathing, we also observed that many proposed ventilators appeared to be under‐powered compared with commercial intensive care ventilators. We wish to highlight what we see as a persisting omission from the published pandemic ventilator requirements and draw the attention of working teams to that omission.

Few teams are yet to formally publish their efforts, let alone make available the various pressure and flow graphs required to ensure proper assessment. Nevertheless, two designs using rotary turbines are illustrative. One design was published before the pandemic [4], the other after [5]. Both ventilators take up to 2 s to achieve desired inspiratory pressure, and almost as long to release this pressure. This results in a saw‐tooth inspiratory pressure wave rather than the square wave observed in commercial ventilators. The long rise time is likely to mimic a trigger or flow asynchrony, and may even mimic the ‘dish‐out’ pattern that has been associated with prolonged ventilation. Delayed release of inspiratory pressure may mimic termination asynchrony as well as leading to loss of effective expiratory time, potentially resulting in auto‐positive end expiratory pressure (auto‐PEEP).

Whereas the importance of matching the ventilators’ actions with those of the patient has long been described [6], many teams have misconstrued this lack of power as providing a gentler pattern of ventilation. This misconception arises from engineers failing to appreciate the need for a responsive ventilator with rapid rise and release times in order to avoid harmful and distressing patient/ventilator asynchrony. Engineering teams have generally been unaware of the potential patient discomfort, increased work of breathing, harmful transpulmonary pressure swings, pendelluft flow, regional lung overdistension and a raft of other adverse consequences.

Yet the RMVS specification [1] is silent on the dynamic characteristics of the pressure waveform. It suggests pressure regulated volume control, or pressure controlled ventilation as the preferred modes and requires the ability to deliver 400 ml breaths at up to 30 min‐1 with an inspiratory: expiratory ratio of 1:2. The Australian specification [2] goes a little further, requiring a minimal achievable inspiratory flow of 100 l.min−1, and suggests the capacity to deliver 150 l.min−1. However, even this guidance fails to express the requirement for rapid initiation of inspiration and also the need to rapidly release inspiratory pressure back to the set expiratory pressure.

Whereas the UK Government has indicated that it is no longer seeking new RMVS designs, many teams around the world continue to invest in them. These teams need to be aware that peak pressure and flow measurements, on their own, do not adequately describe their design’s performance. In order to match the characteristics of modern commercial ventilators, time to achieve and release the target pressure are vital parameters.