What exactly is a "flow limitation" ?

[lars_the_bear]

Hi folks

The graphs I see from OSCAR/ResMed have a "flow limitation" trace, and I'm wondering exactly what that means, and whether it reflects a genuine increase in airway resistance, or could just be an artefact.

For example, if one simply breathes less deeply, or less frequently, for reasons completely unrelated to airway resistance, could that be reported as a "flow limitation"?

I don't think CPAP machines actually have the capability to measure airway resistance -- to do this would require (e.g., ) modulating the delivered pressure with a measurement waveform. Are CPAP machines sophisticated enough to do this kind of thing?

So what, exactly, is a flow limitation?

Best wishes
Lars.

[Julie]

Have you looked at the lightbulb above - Cpap Wiki - to help answer such questions? Lots of good info there!

[lars_the_bear]

Have you looked at the lightbulb above - Cpap Wiki - to help answer such questions? Lots of good info there!

Thank you. I did not find anything there. But I did find an answer in one of the ResMed manuals:

https://media.resmed.com/sites/3/202106 ... ER-ENG.pdf

It seems that the machine guesses there is a flow limitation from the shape of the flow curve.

Best wishes
Lars.

[Respirator99]

Have you looked at the lightbulb above - Cpap Wiki - to help answer such questions? Lots of good info there!

OK, I looked at the wiki as you suggested and for flow limitation it draws a blank.

There's a reasonably clear paper here: https://www.resmedjournal.com/article/S ... 4/fulltext

and the Apnea Board wiki has a decent explanation as well: http://www.apneaboard.com/wiki/index.ph ... limitation

[lars_the_bear]

Thank you. What remains unclear to me is how much time elapses between the machine detecting what it thinks is an abnormal flow waveform, and a flow restriction appearing on the graph. It doesn't seem to be instantaneous, or even within the four-second sample window that I'm told the ResMed machines do their calculations.

Best wishes
Lars.

[rick blaine]

Lars,

I think you have mis-read what I said in the other thread. A moving four-minute window.

[lars_the_bear]

I think you have mis-read what I said in the other thread. A moving four-minute window.

Oops, that's exactly what I did. My bad.

So what happens with respiratory events that last much less than four minutes? I wonder how sensitive the machine is to respiratory restrictions that last, say, ten seconds?

Best wishes
Lars.

[rick blaine]

The definition of an apnea or a hypopnea is a flow which is reduced or absent for 10 seconds or more.

It can be much more than that. I once found a two-minute apnea in my charts.

With that definition being 'the first axiom', it follows that 'flow limitations' is used to label events shorter than 10 seconds.

[MMcG]

The definition of an apnea or a hypopnea is a flow which is reduced or absent for 10 seconds or more.

It can be much more than that. I once found a two-minute apnea in my charts.

With that definition being 'the first axiom', it follows that 'flow limitations' is used to label events shorter than 10 seconds.

I don't think that last point is correct. It's not to do with the duration but the percentage of reduction in the inhalation breath as far as I know when it comes to FL's. My Oscar print-outs show FL's of as short as 2 seconds.

[dataq1]

It seems that the machine guesses there is a flow limitation from the shape of the flow curve.

This my impression as well. The shape might suggest an actual partial obstruction (for example nasal congestion) or it may be indicative of a less-than-smooth effort by the chest musculature to inhale.

But returning to the nasal congestion-partial mucus obstruction- do you imagine that increasing the applied pressure will “clear” that ((pardon the expression) snot?

(Consider the pressure and volume needed to blow your nose to clear the nasal passages)

[lars_the_bear]

I don't think that last point is correct. It's not to do with the duration but the percentage of reduction in the inhalation breath as far as I know when it comes to FL's. My Oscar print-outs show FL's of as short as 2 seconds.

That's what I'm wondering about -- if these machines look at data in four-minute chunks, how are they able to report a flow limitation that lasts only seconds?

As for nasal congestion -- does a nasal congestion look the same, in flow/time terms as, say, a pharyngeal obstruction? I really have no idea. Does the machine increase the pressure if you've got a blocked nose? Would it help to?

My gut feeling is that auto-CPAP machines tend to work, when they do work, because the people using them already know they have OSA. The machine could over-classify events as flow limitations without doing any great harm, because raising pressure probably won't make the treatment for OSA any less effective (so long as it's tolerable). My questions are really about how well the machine detects flow limitations when it isn't clear that OSA is in play.

Best wishes
Lars.

[Pugsy]

As for nasal congestion -- does a nasal congestion look the same, in flow/time terms as, say, a pharyngeal obstruction? I really have no idea. Does the machine increase the pressure if you've got a blocked nose? Would it help to?

These machines only measure air flow...they don't have any way to know where the reduction in air flow is coming from.
The auto adjusting mode will increase the pressure no matter where the reduction is located but when the reduction is in the nasal passage it won't help. The machine just can't generate enough pressure to blow open whatever is blocking inside the nose. Usually it's swollen mucosa and also might have a bunch of drainage (snot) in there as well. No amount of pressure is going to blow open the swollen mucosa or past a glob of snot.
The machine will try though because all it knows is something is blocking the airway so it tries to prevent more blockage by increasing the pressure in an effort to prevent further air flow reductions.

This is why when we see a really active flow limitation graph we try to ask if the person is experiencing much nasal congestion or the nose is clear or not. If the nose is clear and the FL graph is really ugly then the usual fix is either more baseline pressure or adding some pressure support if available or sometimes both.
Now if they report the nose is congested a lot...we know that more pressure won't help the congestion and other more traditional ways of dealing with congestion will need to be attempted.

Here below is a good example. This person has one of the worst FL graphs I have ever seen and you can clearly see the machine tried to fix things with more pressure but was unsuccessful. This person had really bad nasal congestion partly from allergies and partly from humidity settings on the machine.

[dataq1]

As for nasal congestion -- does a nasal congestion look the same, in flow/time terms as, say, a pharyngeal obstruction? I really have no idea. Does the machine increase the pressure if you've got a blocked nose? Would it help to?

These machines only measure air flow...they don't have any way to know where the reduction in air flow is coming from,,,
all it knows is something is blocking the airway so it tries to prevent more blockage by increasing the pressure in an effort to prevent further air flow reductions.

Pugsey, please correct me if this is wrong but:
The machine supposes that something is blocking the airway because of the shape of the flowrate curve. (there are other possible explanations for non-symetrical inhalation curves). In anyevent, the algorithm guesses that there is a blockage. The only tool that the machine has in it's arsenal to clear that blockage is to increase the pressure. If increasing the pressure fails after repeated and prolonged attempts, then perhaps the shape is being incorrectly diagnosed (as a pharyngeal obstruction)

[Pugsy]

As for nasal congestion -- does a nasal congestion look the same, in flow/time terms as, say, a pharyngeal obstruction? I really have no idea. Does the machine increase the pressure if you've got a blocked nose? Would it help to?

These machines only measure air flow...they don't have any way to know where the reduction in air flow is coming from,,,
all it knows is something is blocking the airway so it tries to prevent more blockage by increasing the pressure in an effort to prevent further air flow reductions.

Pugsey, please correct me if this is wrong but:
The machine supposes that something is blocking the airway because of the shape of the flowrate curve. (there are other possible explanations for non-symetrical inhalation curves). In any event, the algorithm guesses that there is a blockage. The only tool that the machine has in it's arsenal to clear that blockage is to increase the pressure. If increasing the pressure fails after repeated and prolonged attempts, then perhaps the shape is being incorrectly diagnosed (as a pharyngeal obstruction)

Remember that none of these machines will actually increase the pressure DURING the whatever apnea event...be it airway or nose or whatever. The auto adjusting algorithm knows it can't blow open anything while it is happening and won't even try. The machine sits by and twiddles its little thumbs DURING the event (OA/hyponea).
So we don't expect the machine to clear the blockage from whatever....what the increase in pressure is supposed to do is PREVENT whatever happened in the past from happening again in the future.
One the apnea event has resolved at that time the machine uses its algorithm/brain to decide what all transpired in the past and what might be going on now and its best course of action to hopefully prevent all that from happening again.

At least that is the way things were designed to work.

We don't CLEAR active blockages with more pressure and the machine won't even try.
A lot of people think that the machine blows past collapsed airway tissues to open up the airway...it doesn't...it can't because it can generate enough pressure fast enough to physically move much of anything in terms of something that has sagged into the airway and is either fully or partially blocking the airway and reducing the air flow. They simply don't work that way in auto mode.

Now if after several "attempts" at better preventing the machine perceived reductions in air flow we don't obtain any reduction in those air flow reductions then we start looking at nasal caused air flow reductions and not airway caused reductions. The machine doesn't necessarily get it "wrong" because it simply has no way of knowing where the air flow reduction is coming from. It's making an educated guess based on the algorithm built into the software given the data it has at its disposal.

People tend to expect these machines to know a lot more than they are capable of knowing and expect these machines to do things that they are incapable of doing.
And yes...pretty much educated guesses as to what is going on and how best to deal with whatever that educated guess might point to.

Remember when the auto adjusting algorithms were in their infancy they were directly compared (in lots of trials) to in lab sleep study data that is much more definitive and what the machine perceives as some sort of flow limitation based on breath shape was compared to breath shapes documented in the in lab sleep studies and they got a favorable comparison.
All the stuff that the auto adjusting algorithm does or doesn't do...all compared to in lab sleep studies.
They didn't just pull this stuff out of the air with no basis to document what and why they are doing anything.

Now is it perfect? Nope...nothing is ever perfect because there are always potential variables in medical science in general along with people in general.
Me...I decided a long time ago to trust the machine even with the known limitations because it's not like I can do anything about it anyway. I don't expend time nor energy worrying about stuff I have no control over anyway. Waste of my time.

So yes...the machine does sort of make a "guess" when it does something but it is based on an educated guess that was proven in comparisons with known data from real in lab sleep studies where the data is much more comprehensive.

[dataq1]

It may have been more accurate for me to say (with respect to auto modes)"

The only tool that the machine has in it's arsenal to prevent that blockage/limitation is to increase the pressure.

However if an increase in pressure does not prevent similar "limitations or blockages" ... the machine has no more arrows in it's quiver.

(and you are certainly correct to point out that auto machines are reactive, while continuous machines are intended to be proactive.)