Recent months have seen an extraordinary global focus on intensive care and mechanical ventilators, as governments respond to the COVID-19 emergency. Understandably, much of the media attention has concentrated on urgent efforts to create the capacity to cover worst case pandemic scenarios. But beyond the headlines, what role will innovation now play in improving ventilators and ventilation?
Building a life-saving ventilator at lightning speed.
Ventilators are complex devices, capable of meeting a wide range of respiratory support needs. They must help to optimise gas exchange in a way that avoids further injury to the lung. Ventilators help relieve patients of some or all of the work of breathing – but once the lungs have sufficiently recovered this needs to be reversed so that patients are able to breathe for themselves again.
With such complexity, managing ventilation demands a high level of operator skill and plenty of scope for ventilator settings to be less optimal, particularly as a patient’s condition changes over time.
Overcoming poor synchronisation
A key functionality of ventilators is for the device to synchronise with the patient’s own breathing efforts to support them at the right time. Ventilators use changes in flow or pressure to trigger the inhalation and exhalation phases of support.
However, poor synchronisation of the ventilator with the patient is relatively common. Studies indicate that it occurs in up to 85% of patients at some point during treatment and that nearly a quarter of all breaths show signs of patient-ventilator asynchrony (PVA). Sometimes it is obvious, with the patient “fighting” the ventilator, but often this is not the case. PVA can result in patient discomfort and the need for deeper sedation, disrupted sleep patterns, wasted effort and lung injury. These are all associated with prolonged need for ventilation – with higher care cost – and poorer patient outcomes.
PVA may be caused by factors relating to the ventilator settings, the patient or both. In many cases, an experienced intensivist can identify and rectify the problem by inspecting the pressure and flow curves on a ventilator. But this requires a level of skill and experience that few attain to a high level of accuracy. And because patient conditions change over time, there is a need to keep checking and adjusting settings.
This is just the sort of problem that classification algorithms are excellent at, either to flag an issue or to go a further step and include synchronisation as one of the end-point goals in automated ventilation. I expect to see this type of smart ventilation – currently limited to only one or two high-end ventilators on the market – becoming much more widespread.
Diaphragm monitoring
Analysis of pressure and flow curves will not be a panacea for all incidences of PVA. There is good correlation, but some types of asynchrony are difficult to detect by this means. In some patients it is very difficult to reliably pick out the inspiratory signal from the airway pressure trace.
Two options to improve reliability are to measure either the oesophageal pressure or the nerve signal to the diaphragm with a technique known as electromyography (EMG). These approaches require the placement of either a balloon catheter or an EMG electrode array into the patient’s oesophagus.
Electromyography has the additional advantage of being able to measure not just the timing but also the magnitude of the inspiratory effort. This allows the level of ventilatory support to be adjusted in proportion to diaphragm activity. But perhaps because of their invasive nature, these devices are not used widely for this purpose.
From my viewpoint, there are a number of less invasive options that could be attractive for more mainstream use. Surface EMG, using electrodes placed on the thorax, has been demonstrated to be suitably reliable for the purpose. Electrical impedance tomography, used for helping optimise lung recruitment, provides additional information for the detection of asynchrony. And an ultrasound-based wearable could also be used to measure diaphragm movement and thickening to measure timing and a correlate of inspiratory effort.
Maintaining breathing muscles
When mechanical ventilation takes on the work of breathing, there is a rapid weakening of the patient’s diaphragm, the main muscle that drives inspiration. The patient can become ventilator dependent and the therapy is difficult to withdraw, increasing the risk of complications such as a lung infection.
But there are some really interesting approaches coming into early clinical use right now that use electrical pacing to preserve or restore breathing muscle function. The Lungpacer device, for example, depends on an electrode array being inserted through a central venous access to stimulate the nerves that drive the diaphragm to inspire.
Liberate Medical, on the other hand, has taken a non-invasive approach, using externally placed electrodes to drive the abdominal muscles during expiration. Other diaphragm pacing systems developed for patients with chronic conditions such as spinal cord injury or central sleep apnoea require surgical electrode placement and are probably less well suited for use in acute settings.
The use of these types of devices in critical care settings is very much in its infancy. Although early clinical study results indicate that they can reduce the average length of mechanical ventilation, it will take some time to get a better picture of which patient groups they are most effective for. The current COVID pandemic is likely to generate some insight into this because of the long durations of mechanical ventilation we are seeing. Emergency Use Authorisation from the FDA for devices with this indication will help.
One of the requirements of these devices when used in support mode is to synchronise with the patient’s own efforts. The Lungpacer device requires manual control, with the carer initiating the stimulation when the inspiration cycle starts. The Liberate device requires insertion of an additional sensor into the ventilator circuit to synchronise. If electrical pacing of breathing muscles fulfils its early potential, it is likely that an electronic triggering signal will become a function of ventilators in the future. Ventilator manufacturers may want to make it proprietary to their own pacing therapy solution.
Challenges and opportunities
While recent focus has been on the availability of ventilators, there are some fascinating developments in the space that could help in the goal of improved patient outcomes. They all could be supportive in the automation of ventilation setting to deliver a prescribed therapy goal. While automated ventilation has been on the market for several years, we’ve yet to see its wholesale adoption. But the demographics of developed economies – together with the rising costs of healthcare – means that the pressure to adopt technologies that drive staff efficiency will continue.
There is a natural and justified caution in handing over control of a ventilator to an algorithm. Healthcare professionals worry that such devices will dumb down care and that staff will no longer trust, or even make, their own judgements. In the worst-case scenario, they’ll have no instinct of what to do in a crisis. These are key challenges. Device manufacturers must create interfaces that clearly communicate the status of the system, to allow seamless transfer back to human control. They should also provide an understandable model of why the current settings have been selected, to support an understanding rather than blind acceptance of automation.
Significant early adopters of automated ventilation could be emerging markets, where there is a huge shortfall in skilled ICU staff and a willingness to embrace new technology that leapfrogs developed markets. A great example of this is the M-Pesa telephone system pioneered in East African markets years before the launch of Apple Pay. Then there’s the fact that Rwanda is the first country in the world to build a universal primary healthcare service based on digital technologies. It’s important to note, of course, that any initiative will need to be founded on devices that are specifically designed for the needs of such markets, rather than the premium ventilators that feature advanced technology.
If you’d like to discuss any aspect of mechanical ventilation – and the potential pathways to improvement – please don’t hesitate to drop me an email.
