Different AHI results between two CPAP machines

[Arik]

Hi Guys,
 
I have two CPAP machines at home (ResMed S9 and Lowenstein PRISMA 20A). I noticed that there are considerable differences in AHI readings between the two machines as follows:
 
1.      ResMed AHI average reading per night is about 1 (mixed Apnea and Hypopnea events).
2.      PRISMA AHI average reading per night is about 5 (mainly Hypopnea events, very few Apnea events).
 
Pressure limits are set the same on both machines. Mask leakage is also about the same.
Lately, I took two home test with Itamar Medical’s Watch PAT 200 at different days. This machine is an FDA-approved portable diagnostic device that uses finger based physiology and accurate SPO2 monitoring to indirectly calculate AHI and other sleep parameters. The result of the two home tests was AHI=7. At the same nights, the ResMed results were about 1 (I used both devices concurrently).  
It looks to me that there are 3 possibilities here:
1.      The ResMed therapy is much better than the PRISMA, therefore I get lower AHI results with the ResMed.
2.      One of the machines is underestimating (or overestimating) the true AHI result.
3.      OSA drives me nuts (my wife supports this option….).
I looked deeply into the "Flow" charts of the two machines and came into the conclusion that ResMed ignores events that the PRISMA will count as Hypopneas.  In the attached file you can see two "Flow" charts, taken at different night but support my assumption. In the lower chart you can see how ResMed algorithm is "ignoring" two events which should be defined as Hypos to my opinion. Also in the upper chart you can see a more “strict” behavior of the PRISMA’s algorithm.
My assumption is also supported by the two home test results, although the AHI is calculated indirectly and there is no breakdown between AI's and HI's .
Needless to say, we need to know what the true AHI value is in order to pursue the best treatment.
Please let me know what you think. Note that the deference between the two reading is made mainly of Hypopnea events.
Thanks, Arik

[Sleeprider]

Another member tested six different machines including the Resmed and Prisma, and compared the results to a polysomnogram (PSG). http://www.apneaboard.com/forums/Thread-...rification His findings confirm a higher event rate using the Prisma vs Resmed. We have seen other CPAP machine shoot-outs that all seem to show that equal results are not something to be expected between different auto CPAP machines with the same settings, and we see the same thing in member's posted results on teh forum.

[DanGagner]

My son in law and I switch cpaps one night. He always gets 0 to 0.4 ahi. Mine is usually 0.6 to 1.5ish. When we switched mine was also lower and his higher. He has a resmed and mine is a dreamstation. I think the algorithms about what is an apnea is different in each.

[DeepBreathing]

I think the algorithms about what is an apnea is different in each.

I believe the apnea definitions are the same, but hypopnea may be different.

Hypopnea
Hypopnea is a partial blockage of the airway. During a hypopnea, breathing is shallow at 30%-50% less than your normal breathing for 10 seconds or longer.
https://www.resmed.com/au/en/consumer/su...aries.html

From Philips Respironics:

What is a Hypopnea?

An episode of overly shallow breathing or an abnormally low respiratory rate. Our auto algorithm establishes a baseline of patient flow based on a moving flow signal window. When your device detects a hypopnea as a 40% reduction in flow lasting at least 10 seconds, followed by a recovery breath, then a Hypopnea event is recorded.
https://www.mysleepmapper.com/Help/Faq#

The Philips definition is more precise than Resmed's, which make comparison difficult, but they are obviously not exactly the same.

[Arik]

Hi,
 The problem I raised with this post was a HUGE difference in AHI score between the ResMed and Lowenstien CPAP machines, which cannot be explained by different definitions. The question is which AHI value should I consider when trying to optimize my treatment. I assume that I have to compare both machines by manually counting respiratory events on the FLOW chart. Alternatively, I can use my Oximeter to figure out which treatment will result the best saturation results (=less desaturation events per night).

Arik

[DeepBreathing]

Arik, if I understand you correctly, you used the home diagnostic device at the same time as your Resmed Autoset and they gave a similar result AHI = 1. Did you do the same with the Prisma and the device?

Looking at the charts you provided, it might be useful to show graphs of the same period for both machines (on the present charts you've got 3 minutes of the Prisma but only 1:24 on the Resmed). That would give us a better feel as to how the two machines compare.

The Resmed is treating these events as potential apneas and is starting the forced oscillation technique (FOT) to determine the type of apnea. In each case there is a tiny uptick in the flow rate after the apnea has commenced and the FOT ceases, only to start again a few seconds later. I think the machine is being tricked by the small uptick. On the Prisma, the machine is flagging two events, but seems to be ignoring a third one at 00:14. Also on the Prisma, there is a fairly persistent leak recorded through the whole period, which is distorting the wave flow diagram.

Both machines are showing the flat-topped wave form consistent with flow limitations. This appears more severe on the Prisma.

In Oscar you can set "user flags" which allow the software to measure and flag events which fall outside the standard definitions of apnea and hypopnea. For example, you could set one flag to detect events with full loss of flow for 5 seconds and the other for a 40% loss for 8 seconds. Or any other combination you like. You could use this feature to help determine what's happening in more detail and easier than manually scoring the events.

[bluesboybob]

Do you feel any physical difference in quality of sleep with the two machines?

[Arik]

Hi,
Thanks for your useful comments. 
Let me clarify:  

• I slept one night with both ResMed and Watch PAT 200 (ResMed AHI=1, Watch PAT AHI=7).
  
• I slept one night with both PRISMA and Watch PAT 200 (PRISMA AHI=5, Watch PAT AHI=7).
  
• Watch PAT does not distinguish  between Apnea and Hypopnea events. It does use professional oximeter.
  

The charts I added were for the purpose of demonstrating ignored events with the ResMed. 
I do not understand your note about the tiny upticks.
The user flag at Oscar is a great feature I was not aware of. I'll take a look at it and see what can be learned.
Finally, I want to comment that the AASM definition of Hypopnea requires that "There is a ≥3% oxygen desaturation from pre-event baseline and/or the event is associated with an arousal". Most CPAP machines including ResMed and PRISMA are not able to detect either Oxygen disaturation or an arousal.
Thanks,
Arik

[Arik]

No, I don't think so.
Thanks,
Arik

[Hydrangea]

I have 2 machines: DreamStation Auto is my daily machine, and ResMed Airsense 10 AutoSet is my travel machine. I have them both tuned in to similar settings that make me feel similar symptom relief. But my ResMed always yields an AHI that's about half of the AHI on the DreamStation.

I've personally concluded that they must have their own proprietary way of counting AHIs, and thus why the numbers are so different.