2

More CPAP for COVID

 3 years ago
source link: https://blog.plan99.net/more-cpap-for-covid-b6911f806c89
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.

More CPAP for COVID

Since Sunday’s analysis quite a lot has been happening. I’ve been asked to post a quick update.

It feels like a lot of people are simultaneously hitting on the same ideas, looking for information and ending up on my blog. Some of them have medical backgrounds, others are engineers or suppliers who want to do something.

Of those, some are then co-ordinating with each other on the mailing list:

You can join simply by sending an email to [email protected] — no new accounts or passwords are needed. As can be seen in the archives, it’s already quite active.

Three key groups of people have started to take this idea seriously (not because of anything I’m writing!):

  • Governments
  • Manufacturers
  • Doctors

Governments

The US FDA has announced it’ll allow manufacturers to press sleep apnea machines into service and do other unconventional hardware/software changes to increase the supply of ventilators. That’s a big step from where things were last week.

Input from manufacturers

ResMed is a leading manufacturer of sleep apnea devices. They posted information to their website alongside a letter from their CEO, Mick Farrell. They also emailed many customers with information on two of the most important questions we’re asking: can you use sleep apnea machines to provide assistance for patients, and what about aerosolisation risks?

Their answers are both positive and negative.

On CPAP: “Our AirCurve 10 bilevel devices, which look like our CPAP devices, can provide support to some COVID-19 patients

They also say in an email: “CPAP devices are designed to provide only PAP (positive airway pressure) and would require significant mechanical and software rework to function as a ventilator.

Bi-level devices are conceptually very similar to the cheaper CPAP variants: they’re both used to treat sleep apnea, the machines look nearly identical and they both have similar feature sets. Bi-level machines vary in two key ways:

  • They can produce slightly higher pressures out of the box.
  • They can rapidly vary between two pre-set pressure levels for breathing in and out. The cheaper [Auto] CPAP machines can also do this but with only a small amount of ‘flex’ around the target pressure the machine selected. Bi-level machines can ‘flex’ a much larger amount.

Unfortunately bi-level machines much rarer than the cheap CPAP machines used by most sleep apnea patients. Remember: this idea is interesting to explore because there are so many CPAP machines out there in the field. Being restricted to a rarer class of machines would make the tactic less useful.

ResMed have looked at conversion of in-field hardware too:

Converting in-market CPAPs to support more advanced modes of ventilation requires careful assessment to ensure risk is acceptable, and must meet labeling requirements to be set forth by the FDA. We are currently exploring options for achieving this; at the same time, we are converting manufacturing resources typically used for PAP devices to support the manufacturing of bilevel and non-invasive ventilation devices that are more suited for the immediate care of COVID-19 patients. We are also ramping up production of invasive-capable and non-invasive ventilators to double our normal levels, while working closely with the U.S. Government and other government agencies to assess their needs

End of the line? Maybe not.

Input from doctors

We’re starting to see public discussion from doctors on the front line about their experiments with simple CPAP. This tweet comes from Head of E.R. at Papa Giovanni XXIII Hospital who has seen over a thousand COVID-19 patients:

The diagram here comes from a report being written by assistant professor of pulmonology Josh Farkas. It shows how cases need to be seen as a ramp, with deterioration pushing patients towards full intensive care and invasive ventilation (i.e. a tube pushed down your throat) with less extreme forms of assisted breathing having a role to play both before and after.

In these tweets we can see debate on a key issue: Continuous positive airway pressure (PAP) vs Bi-level PAP. They’re arguing that the simpler CPAP is actually better than bi-level PAP. Mr Farkas says:

  • Atelectasis leading to hypoxemia seems to be a major problem among these patients. (Image of progressive alveolar collapse)
  • CPAP could have major advantages here:
  1. CPAP can provide the greatest amount of mean airway pressure, and thus most effective recruitment. (Image comparing mean airway pressure due to CPAP vs. BiPAP.)
  2. CPAP doesn’t augment tidal volumes, so this could facilitate more lung-protective ventilation.

NB: The “CPAP” these doctors are talking about is still generated from hospital ventilators. Those machines are currently still available and give real-time monitoring, control, alarms and so on. So ResMed’s view may be based on the assumption of ‘normal’ pneumonia patients, whereas COVID-19 may behave a bit differently. If that’s the case then more hardware may become available.

On the other hand other doctors aren’t convinced and take the view that most that enter an ICU end up needing intubation, even if less invasive treatment was tried first. According to a paper from March 5th ‘the application of non-invasive ventilation for patients in the ICU is controversial’.

Thus the picture emerging so far isn’t as simple as “can sleep apnea machines be turned into ventilators” — there’s several sub-questions to this topic:

  1. What kind of pressure variance is best at each stage of the disease?
  2. Given each type, what kind of machines can generate such air flows and where can they be quickly acquired?
  3. Can some clever hack be found to allow one type of machine to be up-converted into another without a significant rework? Probably not, but it’s worth pondering the question.
  4. Even if all the above answers can be resolved, is it useful at all if most patients progress to needing the most severe interventions possible?

Aerosolisation

If hospitals need to start relying on masks rather than inserted tubes, does that spray virus all over the staff?

Interestingly, ResMed say no:

“Evidence suggests that non-invasive ventilation procedures are more likely to produce large droplets (>10 μm) rather than aerosols, and that these are largely confined to within one meter due to their large mass. This suggests that the risk of droplet dispersion as a result of using non-invasive ventilation or bilevel devices may not be that different to that of any COVID-19 patient in the hospital who is coughing or sneezing. Additionally, non-invasive ventilation systems with a good interface fitting do not create widespread dispersion of exhaled air, according to an experts’ panel cited by the U.S. National Center for Biotechnology Information (NCBI).”

That would be highly convenient if correct. If not, then people are working on a variety of solutions for that too. One is a so-called ‘negative pressure hood’. The simplest is basically a plastic box attached to a vacuum cleaner:

Image for post
Image for post
Bill Kaigler via JT TanApollo 13 for COVID

The patient wears a mask inside the box that vents exhaled air into the box, that air is then sucked out through the hose somewhere else e.g. an air vent that goes outside.

Another neat idea comes from a team in Italy who are adapting snorkelling masks by 3D printing additional parts:

Image for post
Image for post

Doctor Favero shared with us an idea to fix the possible shortage of hospital C-PAP masks for sub-intensive therapy, which is emerging as a concrete problem linked to the spread of Covid-19: it’s the construction of an emergency ventilator mask, realized by adjusting a snorkeling mask already available on the market.

As 3D printers are widespread and this kind of snorkelling gear is increasingly popular too, that might be a quick source of masks. Good thinking!

Converting CPAP to bi-level machines

Let’s assume for the moment that Dr Farkas & Dr Cosentini are wrong — using CPAP instead of more conventional bi-level PAP is a dead end.

One question keeps nagging at me; what does ‘significant mechanical and software rework’ mean in ResMed’s answer, and if that’s the case, what retrofit approach are they planning to investigate? In the UK and USA companies are designing brand new ventilators from scratch, complete with new production lines. Is it really possible that’s easier than modifying existing hardware?

CPAP motors are capable of generating higher pressures than the devices are rated for and can change speeds quite quickly. A ResMed patent from 1996 provides some insight into one of the differences: whilst large pressure changes can be done by changing motor speed it creates unnecessary noise, heat and peak power loads. The motor has to be able to double or halve its speed within about 100 milliseconds, which makes it more efficient to use a separate motor to throttle the air intake to the turbine.

Are modern CPAP electronics designed cheaper on the assumption the motor doesn’t change speeds quickly? I don’t know how to answer that. Someone with electronics experience wouldl need to dismantle a machine and figure out what it’s capable of if all the rules are broken. If bi-level machines are merely more optimal for bi-level pressures, but CPAP hardware could also swing pressures quickly (e.g. by being over-driven beyond their design tolerances), then perhaps large numbers of machines could be kludged into service as bi-level machines fairly quickly.

Community experimentation

On Sunday I was contacted by JT Tan, a man who seems to have superhuman energy and speed. He’s thrown himself into the problem and talking to as many specialists as he can. He’s rapidly putting together a network of qualified medical experts. Read an update from him here. He seems to be focusing on the question of how to protect medical staff from aerosolised virus.

It’s worth asking if all this flurry of activity is pointless; shouldn’t we all just sit it out at home and let the professionals get on with it? Arguably yes, but there are a few reasons why volunteers on this isn’t a bad thing.

Firstly, not enough professionals. Right now only a few doctors seem to have both enough experience with COVID-19 and enough time to be discussing unconventional therapies in public. Everyone else on the front line is swamped with patients and as invasive pro-grade ventilation is known to work, that’s what they’re going with. What happens if all the ventilators get used up? It’s not obvious many medics are thinking about it — they’re busy!

Meanwhile at the CPAP manufacturers it seems all hell has broken loose. ResMed have had employees working in hazmat suits in Wuhan and their manufacturing operations were interrupted when the Malaysia/Singapore border was closed (workers could no longer reach the factory). Their focus is on scaling up standard therapy as fast as possible whilst dealing with huge levels of business disruption — it’s plausible they don’t have much time to R&D weird hacks of products they sold years ago, especially not to provide therapy conventional wisdom says doesn’t work well enough. They face a tough choice of focusing on the best possible solutions for the immediate need, or worse solutions for the potential 10x wave following it. Not an easy call!

Secondly, the need for ad-hoc solutions. If sleep apnea hardware does get pressed into service there are going to be basic problems you don’t need much expertise to work on, like how to keep infected air expelled by the patients away from staff and how to monitor them. Sure, the average mechanical or software engineer may not know much about an ICU ward, but right now there’s a planet full of them with a sudden excess of time. If even 1% of them hit on the right idea by chance, that’s a ton of energy being focused on useful solutions.

Finally, in war-time availability beats quality. This is one of the machines being proposed to the UK government as a supplemental ventilator:

Image for post
Image for post

Other hospitals are experimenting with connecting several patients to a single ventilator with a simple tube splitter. All the fancy expert wisdom in modern ventilators go out the window the moment you consider things like those.

At times like these volunteers have a key advantage over the established institutions: they’ve got nothing to lose. ResMed are respectable purveyors of fine medical equipment and would never dare produce the contraption pictured above. And as their reference to FDA labelling requirements reminds us, just because there’s a crisis doesn’t mean the law evaporates. Does the exact labelling of whatever ResMed produces really matter right now? Probably not, but regulations don’t flex (libertarians, this is your cue). Established players could shove out untested recipes for hacking their own products, but they have so much to lose: 7000+ jobs in ResMed’s case. In contrast the legal blast radius of a lone hardware hacker whose ideas spread virally is one: themselves. In theory they could even go full Satoshi and never be found.

Last thoughts

Is it possible we won’t actually face a ventilator crisis? That’s not what governments are planning for but it’s worth considering:

  • Manufacturers seem to be reporting that they’re able to double or triple production capacity of standard pro-grade hardware, and there are a bunch of different manufacturers. So a whole lot of new machines are heading down the pipe to hospitals.
  • Old hardware that still works is being pulled out of storage. Some countries seem to have a lot of that.
  • Lockdowns are being predicted to bring demand down below the (temporarily increased) capacity of healthcare systems.

Today in the Daily Telegraph we see this:

Prof Ferguson said that the new social distancing measures announced by Boris Johnson earlier in the week meant the NHS would now be able to handle the incoming cases of coronavirus. “The strategy being done now in some areas ICUs will get close to capacity but it won’t be reached at a national level,” he said. “We are reasonably confident that at a national level we will be within capacity.”

How long can these extreme measures last? Entire countries are practically under house arrest, it’s more restrictive than even London during the Blitz. If hospitals are not going to overflow then it’s unclear what happens next, but it’s possible lockdown would be steadily relaxed to try and ensure case loads match available capacity. “Available” in this case is unlikely to be calibrated based on what’s possible with unusual hardware hacks but rather what doctors recommend … and that is likely to be what they’re used to and know works.

How to balance the crippling and totalitarian measures currently being imposed against quality of care will be one of the toughest challenges our political class has ever faced. I hope they’re up to it.


About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK