runaway pressure on Auto?

Currently have the range set on my machine to 10-15... it has been that way since last Saturday.

Getting good results with this. AI has been 0 for 4 of those nights, the other 2 nights it was 0.1 & 0.2. AHI varies between a range of 2.8 - 5.1. The 5.1 happened only one evening.

I have noticed on occasion that the pressure will spike all the way up to the maximum of 15... stay there for maybe 5 minutes then drop back down. Not showing any events not dealt with during this time. At the outset of that spike leak was reported at .06 L/s... then dropped down... so the leak rate is acceptable and the pressure spike cannot be due to a leak.

Is this pressure spike what I have been reading about the Auto machines here... that sometimes they can "runaway" with the pressure to the max for whatever reason?

Or is it possible that pressure was needed to deal with an event... and it was successful?

The bottom line is that I'm wondering what causes this to happen. I was thinking of narrowing the range down to 10-14 and calling it a day in terms of pressure adjustments.

Paul56 wrote:Currently have the range set on my machine to 10-15... it has been that way since last Saturday.

Getting good results with this. AI has been 0 for 4 of those nights, the other 2 nights it was 0.1 & 0.2. AHI varies between a range of 2.8 - 5.1. The 5.1 happened only one evening.

I have noticed on occasion that the pressure will spike all the way up to the maximum of 15... stay there for maybe 5 minutes then drop back down. Not showing any events not dealt with during this time. At the outset of that spike leak was reported at .06 L/s... then dropped down... so the leak rate is acceptable and the pressure spike cannot be due to a leak.

Is this pressure spike what I have been reading about the Auto machines here... that sometimes they can "runaway" with the pressure to the max for whatever reason?

Or is it possible that pressure was needed to deal with an event... and it was successful?

The bottom line is that I'm wondering what causes this to happen. I was thinking of narrowing the range down to 10-14 and calling it a day in terms of pressure adjustments.

Paul,

The A10 algorithm of the ResMed Autos is not supposed to respond to Apneas above 10 cm. (hence, the name of the algorithm)
So, the only "events" left that it may have responded to are Flow Limitations (which are essentially "partial Hypopneas"), Leaks and Snores.
Since your AI number is "good" and it seems to be taking care of your Apneas, your minimum pressure seems to be doing it's job......you'll need to determine whether the other pressure responses above 10 are worth the pressure changes during the night (if they're disturbing your sleep). You could either try straight CPAP at 10 or APAP mode with a maximum pressure of 11 and then start moving it up over a period of time......if you want to "Lab Rat" your settings. That might also tell you whether it's the pressure changes that are actually causing the Flow Limitations, etc.

Den

Wulfman wrote: The A10 algorithm of the ResMed Autos is not supposed to respond to Apneas above 10 cm. (hence, the name of the algorithm)
So, the only "events" left that it may have responded to are Flow Limitations (which are essentially "partial Hypopneas"), Leaks and Snores.

Okay, I need help here with definitions. According to http://www.sleepnet.com/tech5/messages/483.html, it would appear that an apnea is a "severe hypopnea", and you're indicating that a hypopnea is a "flow limitation", right? I'm trying to understand exactly why a machine would respond to a partial obstruction/cessation of breathing but not a full one.

ractar28 wrote:

Wulfman wrote: The A10 algorithm of the ResMed Autos is not supposed to respond to Apneas above 10 cm. (hence, the name of the algorithm)
So, the only "events" left that it may have responded to are Flow Limitations (which are essentially "partial Hypopneas"), Leaks and Snores.

Okay, I need help here with definitions. According to http://www.sleepnet.com/tech5/messages/483.html, it would appear that an apnea is a "severe hypopnea", and you're indicating that a hypopnea is a "flow limitation", right? I'm trying to understand exactly why a machine would respond to a partial obstruction/cessation of breathing but not a full one.

Try reading these links (and some of the rest of the threads listed) and see if it answers your question(s).

viewtopic/t35873/viewtopic.php?f=1&t=33 ... ns#p285716

viewtopic/t35873/viewtopic.php?f=1&t=26 ... ns#p226750

viewtopic/t35873/viewtopic.php?f=1&t=19 ... ns#p168093


Den

The standard apap has no interface to detect blood oxygen levels, so airflow is the only way for the machine to determine an event is occurring.

I'm trying to understand why at 10cm and above, one machine will respond to 69% restriction in airflow (hypopnea), but not a 71% restriction in airflow (apnea).

I do understand it to a degree, such as when I'm wearing a nasal mask and talking. I seal my nose off from allowing air in, and speak through my mouth. Obviously, the machine should see that as an apnea since the machine has no knowledge that I'm intentionally restricting airflow. However, I'd expect my machine to try to blow its way into my respiratory system one way or another.

I'm sure there's a reason, but I can't come up with one.

ractar28 wrote:The standard apap has no interface to detect blood oxygen levels, so airflow is the only way for the machine to determine an event is occurring.

I'm trying to understand why at 10cm and above, one machine will respond to 69% restriction in airflow (hypopnea), but not a 71% restriction in airflow (apnea).

I do understand it to a degree, such as when I'm wearing a nasal mask and talking. I seal my nose off from allowing air in, and speak through my mouth. Obviously, the machine should see that as an apnea since the machine has no knowledge that I'm intentionally restricting airflow. However, I'd expect my machine to try to blow its way into my respiratory system one way or another.

I'm sure there's a reason, but I can't come up with one.

ResMed designed their A10 algorithm to be cautious of treating apneas that may be "Central" in nature. 10 cm. is apparently a threshold at which they can start....."pressure induced Centrals". So, if the person does NOT have any indications of Central Apneas......and wishes to use a ResMed Auto......AND needs a pressure above 10 cm., the Auto's minimum pressure needs to be set to where it will take care of all/most of the Apneas.

Den

ractar28 wrote:The standard apap has no interface to detect blood oxygen levels, so airflow is the only way for the machine to determine an event is occurring.

I'm trying to understand why at 10cm and above, one machine will respond to 69% restriction in airflow (hypopnea), but not a 71% restriction in airflow (apnea).

I do understand it to a degree, such as when I'm wearing a nasal mask and talking. I seal my nose off from allowing air in, and speak through my mouth. Obviously, the machine should see that as an apnea since the machine has no knowledge that I'm intentionally restricting airflow. However, I'd expect my machine to try to blow its way into my respiratory system one way or another.

I'm sure there's a reason, but I can't come up with one.

While it is true these machines are flow-based they convert that flow into a sine wave with a numeric value. The percentage of flow reduction for scoring of the event can vary by mfg's for their particular algorithm, many of which are patented algorithms. So for them all do have an algorithm they each put a twist on them to get past the patent issue.

Nearly all autopaps limit any response to frank apnea at or above 10 cm pressure. Resmed calls theirs the A10 algorithm, Respironics NRAH or Non-Responsive Apnea/Hypopnea. The PB420e and Sandman Auto use another patented algorithm with Cadiac Oscillation detection to avoid responding to central apnea. All these algorithms do is better differentiate the difference between a obstructive and central apnea.

From the machine's perspective, a 40-second obstructive Apnea has NO airflow. Same can be said for a 40-second central apnea NO airflow. SO to the machine those two events look identical. However, one you respond to with pressure, the other it only makes the number of events worse.

Statistics done on the general OSA population show that once you get above 10 cm pressure that the risk of pressure induced central apnea is greatly increased called the apneic threshold. So knowing that statistic, CPAP mfg's incorporate algorithms to avoid responding to an event it cannot easily distinguish.

Vibratory Snore: Those same statistics also show that if a person snores if an apnea were to happen it will most likely follow a snore. Knowing this and that snore is easy to spot when you convert the flow signal to a sine wave, snore looks like dragging a stick along a picket fence. So if you increase pressure to eliminate the snore you most likely will also prevent the follow-on apnea. So nearly all mfg's use snore detection to preemptively treat apnea.

Hypopnea, Flow Limitations: These are residual or partial obstructions. When converted to sine wave from the flow signal, these events have specific flattening seen on the sine wave signal once converted. Numbers are assigned to specific points of the sine wave signal, with this these events will show up with flattening. The machine can better detect these events (flow limitation more so than hypopnea). So if the machine can clearly identify what the event is and not be at risk of it being central, it can more accurately respond to it.

Puritan Bennett and Covidian (same company) makers of the 420e and Sandman Auto use a different method of differentiation of the obstructive/central events. While they also employ the same method as Resmed with their A10. They use their sensitive sensor to listen in the flow signal for Cardiac Oscillations. Basically what it means is a Central event will have an "open" airway. So during that 40 second apnea, the obstructive apnea might have a tongue blocking the airway. That blockage also blocks the Cardiac Oscillations heard from your heart. Just like a doctor uses a stethoscope to listen to your heart, that heart beat comes to their ears via a tube. These machines do the same thing but the signal comes from your airway is heard via the CPAP hose. If they don't hear the Cardiac Oscillation, the algorithm says that the apnea is obstructive like your tongue is blocking the airway, so it applies pressure. Now if that same apnea produced a Cardiac Oscillation then the machine knows your airway is "open" from a Central Apnea and the machine knows not to respond.

Command on Apnea: Next, statistics also show that only 60% of Central Apnea is associated with an "open" airway, the others could be closed. So listening for cardiac oscillations is only about 60 percent accurate or 6 out of 10 apnea. What about the other 4? These machines then deploy a similar method of avoidance as the Resmed machines, referred to as Command on Apnea. Now if the machine cannot positively differentiate the apnea by hearing the Cardiac Oscillation, it will use the Command on Apnea Maximum pressure value before responding to that suspect apnea. If the pressure is at 10 cm or above, it will not respond to stand-alone frank apnea. It will still respond to events it can positively identify such as FL and Snore. On these machines the Command on Apnea or A10 parameter is changeable, because we are not all the same. Some people have a lower apneic threshold, yours might be 9 cm instead of 10 cm. In that case you can lower the A10 parameter from 10 to 9 and limit how the machine responds to those particular events. Similar feature is found for Flow Limitations called Command on Flow Limitation. On the Resmed and others this values are hard-coded in the algorithm and you cannot adjust them.

So without the Cardiac Oscillation detection method imposed by the 420e and Sandman, like with Resmed and others, you have to use the Minimum pressure to manually titrate the apnea just as you would with CPAP. So I guess you could say the 420e and Sandman machines have better differentiating algorithms and can respond to frank apnea throughout their pressure range. The advantage of this is you can use a lower Minimum pressure and still eliminate those apnea seen at the higher pressure. This can help if you have apnea at 14 cm pressure and also have aerophagia.

Paul56 wrote:Currently have the range set on my machine to 10-15... it has been that way since last Saturday.

Getting good results with this. AI has been 0 for 4 of those nights, the other 2 nights it was 0.1 & 0.2. AHI varies between a range of 2.8 - 5.1. The 5.1 happened only one evening.

I have noticed on occasion that the pressure will spike all the way up to the maximum of 15... stay there for maybe 5 minutes then drop back down. Not showing any events not dealt with during this time. At the outset of that spike leak was reported at .06 L/s... then dropped down... so the leak rate is acceptable and the pressure spike cannot be due to a leak.

Is this pressure spike what I have been reading about the Auto machines here... that sometimes they can "runaway" with the pressure to the max for whatever reason?

Or is it possible that pressure was needed to deal with an event... and it was successful?

The bottom line is that I'm wondering what causes this to happen. I was thinking of narrowing the range down to 10-14 and calling it a day in terms of pressure adjustments.

As Wulfman already mentioned, the only items that can drive pressure up to 15 cm is Flow Limitation and Snore. The machine can easily differentiate these so I would let it. Set your maximum pressure to 20 cm and allow it to go as high as it needs to eliminate the event.

Snore is generally followed by an apnea if one is to occur, allow the machine to eliminate the snore and it may also eliminate the follow-on apnea as a result. Snoring is also thought to contribute to microarousals and disturbed sleep, allow it the pressure to eliminate them. Same for Flow-Limitations, those are easy for the machine to differentiate, so allow the pressure to do its job and eliminate them, you will get better sleep.

The risk of runaway pressure is low, more an old wives tale with today's Autopaps, that is what the A10 algorithm does.

Thanks for the responses folks.

The results seem to be good right now with min pressure at 10. After my first week on CPAP at 8 the RT seemed to think 10 would be my pressure after looking at the data. Given results of the min. move to 10 about a week ago that seems about right.

At this point it doesn't seem major changes are needed. I am thinking of lowering the max by 1 pressure point/week and gauge the results... at some point it may become clear what the max should be... or perhaps that I can simply operate the machine as a CPAP on constant setting of 10.

I don't want to mess things up at this point as the results have been doing nothing but get better and better with the small changes.

Does this seem like a reasonable approach?

Paul56 wrote:Is this pressure spike what I have been reading about the Auto machines here... that sometimes they can "runaway" with the pressure to the max for whatever reason?

No. From what you described, it sounds more like this....

Paul56 wrote:Or is it possible that pressure was needed to deal with an event... and it was successful?

Your machine was probably moving the pressure up to normalize an airflow from you that was showing signs of getting limited. It was proactively going up, just like it's supposed to, to open the airway better. It apparently accomplished that goal since no "events" were marked during that time. It successfully prevented your throat from collapsing more. It prevented a limited air flow from getting worse.

rested gal wrote:

Paul56 wrote:Is this pressure spike what I have been reading about the Auto machines here... that sometimes they can "runaway" with the pressure to the max for whatever reason?

No. From what you described, it sounds more like this....

Paul56 wrote:Or is it possible that pressure was needed to deal with an event... and it was successful?

Your machine was probably moving the pressure up to normalize an airflow from you that was showing signs of getting limited. It was proactively going up, just like it's supposed to, to open the airway better. It apparently accomplished that goal since no "events" were marked during that time. It successfully prevented your throat from collapsing more. It prevented a limited air flow from getting worse.

Yep, I agree.

Since Saturday October 25th I have noted the following:

Oct 26 - AI 0.1 - 1 event at 5am
Oct 29 - AI 0.2 - 2 events, one at 5am the other at 7am
Oct 31 - AI 0.1 - 1 event at 6:30am

I am fairly certain for the event recorded in the early morning for the data of Oct 31 that I was awake. For the one recorded at 7am on Oct 29 I know I was awake. If I am awake... those events do not count.

During this period from last Saturday I had 4 nights recorded at AI 0.0. The AHI for the week looks like this: 2.80, 2.70, 5.10, 2.70, 4.80, 3.70, 4.40.

In any case, I hesitate to mess with the pressure range any more at this point at it seems to be pretty much dialed-in.

I have had EPR set off but will give it a go now to see if I notice any difference or the data changes.