Hacker News Reader

ECMO is high risk for the patient, especially venoarterial (VA) ECMO which is probably not what is being described in this article. Venovenous (VV) ECMO doesn't dump blood into the arterial tree, so your risk of stroke or other arterial thrombotic complications is much lower. You still risk infection and venous clotting (and therefore the risks of iatrogenic anticoagulation), but I think most of us would accept those risks.

From a systems standpoint, the use of mechanical support usually makes sense in the context of being a bridge to somewhere that is not in-hospital, even in the event of non-recovery. At least where I practice, we want to be able to offer a durable device, or transplantation, if you don't recover. If you live in a place where the health system would not offer transplantation (e.g., because of some risk factor like advanced age), then offering ECMO makes less sense because it is not a bridge to anywhere, particularly for VA ECMO.

For VV ECMO, I think there has long been a recognition that people can do OK for an extended period of time (in contrast to VA ECMO, where the risks are higher and there are also destination therapies like ventricular assist devices). And if you're waiting for a lung transplant vs recovery, you may be waiting for quite awhile. To this end, there are special catheters for VV ECMO that facilitate mobility so you can retain some degree of strength and mobility while on ECMO (e.g., https://www.getinge.com/int/product-catalog/avalon-elite-bi-... ).

I'm closer to VA ECMO than to VV ECMO (but do neither); still, your comments ring more true for me about VV ECMO (which again I think is the subject of this article), whereas I think that patient risk + superior bridging/destination strategies really do dominate the VA ECMO discussion.

Thank you for this reply. Highly informative.

> whereas I think that patient risk + superior bridging/destination strategies really do dominate the VA ECMO discussion.

Is the idea here that when doctors are considering VA ECMO for a patient there is usually some non-trivial chance that the patient survives without it? In that case the choice, I guess, is between higher survival with more complications (VV ECMO) and lower survival with fewer complications (no ECMO)?

Sort of a facile response but we'll at least consider VA ECMO if we think the benefits outweigh the risks. E.g., a person in worsening cardiogenic shock because of a usually reversible insult (e.g., severe stress cardiomyopathy, or a refractory ventricular arrhythmia) has a very good chance of surviving the acute hospitalization but if I'm worried they will die tonight (even if they'd be fine 2 days from now after some recovery time) then will discuss VA ECMO with the people who actually do it for a living. VA ECMO's complications are a consideration in that calculus. That's the kind of decision where I think an observer could argue (especially if things go well) that there was a non-trivial chance that the patient would have survived without ECMO.

One important distinction is to understand that VA ECMO is more often used as mechanical circulatory support in cases of severe heart failure leading to cardiogenic shock. Used alone or in conjunction with other mechanical devices (balloon pumps or impella pumps) it can augment cardiac output to provide sufficient perfusion and oxygenation of your organs and distal extremities.

VV ECMO, on the other hand, is used purely for gas exchange (O2 and CO2) due to respiratory failure. Much of the debate in the critical care community is centered around which circumstances and patients derive the most benefit from initiation of VV ECMO. The best studied use case, is in the setting of acute respiratory distress syndrome, which is defined by very specific criteria (bilateral noncardiogenic pulmonary edema with ratio of arterial oxygenation partial pressure to fraction of inspired oxygen less than 300 mmHg). The EOLIA trial published in NEJM in 2018 looked at early initiation of VV ECMO in patients with severe ARDS [1]. It demonstrated no mortality benefit of ECMO, however many say that the study was not appropriately powered as the assumptions used to design the study were from 2008 when mortality from ARDS was much higher. Re-analysis of the data from the EOLIA trial using bayesian methods suggests that there might actually some benefit to early initiation of ECMO [2]

1. https://www.wikijournalclub.org/wiki/EOLIA

2. https://jamanetwork.com/journals/jama/fullarticle/2709620

Outside of the other comments here, I would also suggest considering the expected outcomes rather than just whether or not someone will live and die. For example, if someone spends a month on ECMO and the end result is that they're neurologically devastated, require a feeding tube, tracheostomy, and get discharged to an LTAC, was it worth it? If someone can never have a conversation again, or hold they're children, or read a book, was it worth it? If someone is likely to die, is it worth it to break their ribs or cover them in their own blood and fluid when they code? Given how sick someone is when they're put on ECMO, these are all realistic and likely outcomes.

And, look, this is a value choice that families have to make, but also one that most have not considered before getting sick and don't really want to consider when they're thrown into that situation. Physicians know about these likely outcomes, so it affects their opinion about its use.

As a brief aside, if anyone does end up in this situation, it's stressful and I'm sorry that you're going through this. Really. If I could offer one unsolicited suggestion, it would be to discuss the possible and likely outcomes with your physician, beyond life and death, so that you can make an informed decision that's best for your family.

Not to mention the costs.

Medical resources are limited, and in a country where we are chronically under treating people, it makes no sense to torture someone with an ECMO for a month, only for them to 'recover' to the state mentioned above.

That money/medical effort could have been used to save/significantly improve the lives of dozens of people, instead of prolonging the death of one.

It all comes out of an insurance pool, it's not like people are bearing ECMO costs out-of-pocket. It all comes out of the same labour pool, when you've got your entire medical staff put on hopeless cases, there are people who could be helped who are being denied treatment.

We are both chronically under treating some patients and over treating others. Over treatment is a leading cause of iatrogenic harm. Some providers default to aggressive treatment whether due to a sincere belief that it's best for the patient, or to satisfy patient requests, or to make more money. But evidence based medicine guidelines indicate that it's often best to do nothing beyond active monitoring of the patient's condition.

https://dx.doi.org/10.1371%2Fjournal.pone.0181970

Ideally we should be allocating limited medical resources based on formulas that maximize overall benefit as quantified by quality-adjusted life years (QALY). But it's politically difficult to set rational policies around rationing care without triggering toxic disputes about "death panels" and "killing grandma".

To give some color. I haven't heard that marginal costs are ~$5000/day, I have heard they are closer to $20k/day. Granted other procedures were involved, but one ECMO stay was over $12M for 4 months. That is closer to $100k/day.

How many other people could you save (even within the same healthcare system) for that kind of money? These long tails have huge impacts on insurance costs.

Yep that's insane. That's the same as giving 1,000 families $1000 a month for the first year that their child is born. Surely that would have a significantly larger positive impact on society

> I haven't heard that marginal costs are ~$5000/day, I have heard they are closer to $20k/day.

To be clear, I wrote >$5k, not ~$5k.

No supporting data other than my decades working in healthcare and having conversations with coworkers versus my non healthcare working family/friends.

I’d say those in the medical community that have been more exposed to what that altered reality looks like, tend to not want it for themselves. That is to say, they’d rather expire. They’d probably also use that info to inform choices they made for loved ones in a similar situation.

The general population is very mixed but there seem to be more folks on relative basis that would keep someone alive at all costs or try anything to save them regardless of the risk. I also think sometimes it just comes down to medical professionals being able to come to grips with the reality of the situation and make a hard decision when others default choice is keep them alive /try anything.

Aside: I worked in a hospital from 18-25 and have basically seen it all as my job had me in every department caring for every patient and I notice when something crazy happens I still have the ability to not be consumed with shock but assist. I recently saw a pedestrian/auto accident and was on foot myself. About 30 folks around saw it happen. Everyone froze or turned away in disgust. I ran and basically took control of the situation until EMTs arrived. This has happened a few times in my life and I can’t help but to correlate it to my exposure to the hospital environment.

> Outside of the other comments here, I would also suggest considering the expected outcomes rather than just whether or not someone will live and die.

I take this to be obvious.

Cost isn't the only consideration; staffing is also important. ECMO is extremely high-touch (1:1 nursing recommended, and some patients may need 2:1), even compared to other ICU procedures; given a primary concern during the pandemic was running out of ICU capacity, a bunch of patients on ECMO can cause you to get to that point much more rapidly.

Staffing is a component of the cost.

It’s also an absolute availability issue. A hospital has, in the near to intermediate time horizon, the staff that it has. So dedicating staff in such ratios make them unavailable to other patients, so triage considerations come into play.

Staffing and cost are related, but there are staffing requirements that no amount of money thrown at them can fix.

Nurses are not an infinitely fungible commodity.

No, but a significant cause of current staffing shortages are stagnant wages (especially for non-travel nursing roles).

Hospitals paying decent wages to full-time/permanent staff would go a long way towards solving staffing shortages.

not trying to play doctor here, but a quick skim through the wikipedia article shows a high risk for developing neurological issues, infection, and blood related ailments

I imagine, and I hope someone with more knowledge can add on or correct me, that medicinal professionals are generally reluctant to use ECMO for extended periods due to the aforementioned issues. I personally know if my odds of waking up brain dead from a procedure were 50-50%, I wouldn't take that risk. so I wonder if doctors in this case saw a corresponding decrease in such risks?

https://en.wikipedia.org/wiki/Extracorporeal_membrane_oxygen...

I'm not a doctor either, but doesn't chemo have similarly high risks of serious side effects? When the odds of death are close to 100% otherwise, a treatment that improves those odds to 50% doesn't sound so bad.

This is a serious decision, not just between 100% death and 50%, but a significantly worse death for the other half of people than the original certain one.

Take the final period of a person’s life from them and their family for a devastating treatment that doesn’t help and then see how simple of a choice it feels.

My understanding is that hesitation to employ ECMO isn't about the risk of failure.

It's not like chemo or dialysis where you can live some kind of meaningful life while the treatment has an X% chance of buying you Y years.

It's extremely labor intensive. Another commenter mentioned 1:1 fulltime nurse care. If true, that means multiple fulltime employees, working in shifts, per ECMO patient. Given finite medical resources (doctors, nurses, beds, ECMO machines) we would like to make sure we're employing this only for patients who have a hope of recovering afterward.

In that sense it is similar to other "extreme measures" like ventilators and so on. We can sometimes keep somebody alive in a totally nonviable state for an extended amount of time after heart and/or lung failure, but if there's zero chance for recovery it just doesn't make sense from an ethical or practical standpoint.

(edit - when I say "we" I mean society. not "we" as in "we medical professionals." I am not a medical professional!)

> We can sometimes keep somebody alive in a totally nonviable state for an extended amount of time after heart and/or lung failure, but if there's zero chance for recovery it just doesn't make sense from an ethical or practical standpoint.

There are people on ECMO who have permanent heat and/or lung failure but are conscious, comfortable, and do not wish to have care withdrawn. In this case, it can absolutely make sense to continue to support them even if they have no hope of getting off the machine. The non-trivial ethical question is what to do if the cost is extremely high.

> Given finite medical resources (doctors, nurses, beds, ECMO machines) we would like to make sure we're employing this only for patients who have a hope of recovering afterward.

Outside the short-term, we can always hire more nurses and docs, and build more machines and hospital rooms. We should not pretend that the amount of these are fixed in order to justify withdrawing expensive but beneficial care (other than in emergency situations). If we are going to withdraw care because we just don't want to pay for it given the size of the benefits, we should acknowledge this explicitly.

I'd love to know more about this infinite pipeline of doctors and nurses you're referring to! Most coverage I've read lately has been about healthcare workers quitting in droves, so it's a relief to know that there are actually no limitations here.

   If we are going to withdraw care because we 
   just don't want to pay for it given the size 
   of the benefits, we should acknowledge this
   explicitly.

I feel like the discussion is super explicit about this?

For nearly all situations, I would agree that nobody should be suffering or dying because of a lack of willingness or ability to provide or pay for medical services.

But there are practical limits. There is a reason why we don't all get weekly mammograms, prostate exams, and dental checkups even though this would inarguably be the most effective way to catch things early.

> I'd love to know more about this infinite pipeline of doctors and nurses you're referring to! Most coverage I've read lately has been about healthcare workers quitting in droves, so it's a relief to know that there are actually no limitations here.

I clearly state in my comment that I was not talking about the short term. In the long term we just train more and pay more, it’s not rocket science. But if you are curious about how we could have a for-all-practical-purposes infinite pipeline in the short term, that is also easy: just recognize international medical degrees and give foreign doctors work visas.

> I feel like the discussion is super explicit about this?

I don’t think it is, although I agree that different people will interpret conversational norms differently. IMO, there is constant equivocation between medical rationing and medical triage. The mealy-mouth-ness is so ubiquitous that we take it for granted.

Like, why are we even talking about nurse shortages in this thread? If we were discussing whether it’s worth it to pay for an extra floor on our expensive houses, people wouldn't keep bringing up how we have to consider last year’s lumber shortage and how the US has too few skilled craftsmen because it doesn’t properly support trade schools (which is ofc true). We would just say “having another floor is nice, but it’s not worth an extra $100k”.

death and serious brain damage are not the same risk

> I personally know if my odds of waking up brain dead from a procedure were 50-50%, I wouldn't take that risk.

You would accept guaranteed death to avoid a 50% chance of brain death? That’s not necessarily an inconsistent preference, but it’s a highly unusual one.

Did you consider the impact on those responsible for your long term care?

Brain dead is dead. [1] There is no long term care. They turn off the machines pumping air in your chest and that's it.

I think you are confusing brain death with vegetative state caused by severe brain damage.

1: https://www.nhs.uk/conditions/brain-death/

The poster's point would've made more sense, I think, if they'd used "significant cognitive impairment" instead of brain death.

Being left in a permanent state where I'd be conscious enough to want to die, but not be able to legally access that option, is a scary thought. I'd rather not live in a nursing home bed for fifty years.

Fortunately for your case, once people are bedridden, they very rarely survive fifty years.

Our bodies don't work very well when they can't keep moving.

I think the right thing to do for the time being is to consider prolonged use of ECMO as a research practice where not everyone is accepted but those accepted taken care off as long as they are conscious or have a decent chance of regaining consciousness and making their own informed decision. Costs of doing so are part of research budget that determines how many patients can be included. Not providing care to everyone is ethical so long as we don't know if the care is going to help anyway.

In turn, research can determine - Exactly what kind of patients are most likely to recover after prolonged use of current ECMO machines - Potential ways to bring equipment and nursing costs down while preserving most of effectiveness. - Ways to wean the patient off intensive care long term.

For example, the very first genetically modified pig heart transplant resulted in 2 months conscious survival. Even if animal transplants never become a long term solution, they could prove more cost and medically effective than ECMOs to buy time for own organ recovery or a human transplant.

What do the survivors deal with though? I recently had an injury that required surgery and occupational therapy. My entire limb was a sad, useless appendage for months mostly due to immobility and how it reacted to pain. I can’t imagine the mountain one must surmount to get back to even 25% of their pre illness self. I was down to 15% use by the time my OT started. And 6 months later it still is a daily task to keep improving (much better now though).

It must take an incredible toll and on already sick body.

A lot of my infection control colleagues are rightly terrified of keeping people on ECMO or mechanical ventilation due to infection risk, that are extremely hard to treat.

This was especially acute during the early phase of the pandemic, when the combination of the intensity of care COVID patients required + the PPE shortage meant some compromises that they'd really prefer not to make.

The negative side effects are not to be taken lightly, especially if they cause profound suffering prior to a patient’s death or severe lifelong morbidity for a survivor. Yes, there are amazing uses of the technology, but the doctors who refer to proper use are speaking from a place of having seen the improper use.

Some countries do not perform hemicorporectomy, since they think survival would not be worth it.