Resmed's Half Baked Algorithm Or A Miracle Of Medical Science?

My 30 day AHI averages 4.5 on Respironics where as on resmed my AHI averages only 0.6. Is Resmed providing a 12 fold better therapy? It is highly doubtful.

I have severe Central Apnea disorder (>52) with less than a 10% Obstructive Apnea and I am very interested in studying my condition. For this purpose I analyze my sleep data using various software. OSCAR cpap analyzer is my choice however.

Since my therapy is with Adaptive Servo Ventilation (ASV) type of machines, I have bought both, a Respironics Dreamstation ASV and a Resmed Aircurve 10 ASV machine. Since ASV therapy monitors my respiration on a breath by breath basis, when I monitor the data I have a very detailed set to look at, like flow rate, mask pressure, tidal volume,minute ventilation and so on, on a breath by breath basis. So when I compare graphs between resmed data and Respironics, I note that at high resolution (zoomed up to 20 seconds windows), Resmed simply ignores to flag events such as Hypopnea, clear airway,Obstructive Apnea etc, on average 80% less times than Respironics algorithm does. In fact Respironics algo never misses a single event when I examine data at high res.

This explains why Resmed machines generally report a much lower AHI than Respironics. Its too bad that I can't post images here because I can post several comparison examples of sleep data to support what I am saying here.

I'm not sure how Philips Respironics APAPS work because I am on a ASV machine. ASV machines are more akin to Ventilators than they are to PAP machines because they provide a close to spontaneous breath when you miss one while they also they provide auto EPAP pressures to minimize patency in OSA events. CSA events are essentially Clear Airway events that last more than 10 seconds, in which case a ASV algo should deliver successive and elevating IPAP pressures for the preset length of time, based on your Inspiration time. (Resemd has no alternate manual settings for any of this while Dreamstation algo allows you a fiully auto setup or manual setup).

I have no issues with my dreamstation. It is set to the max range of IPAP and EPAP and Pressure support. So when I am awake I can barely even feel it (which is what it should be like) and when I fall asleep it just automatically adjusts the pressures for each breath and I never ever feel it or get bothered by it.

In contrast my Resmed does some crazy things, like just adjusting the pressure up and up for no apparent reason and when I asked Resemed support how to fix it, they suggested that I "keep blowing back hard into the tube", which I thought was pretty stupid.

Resmed algo just looks at the last 90 seconds of my breathing pattern and sets up its backup based on that 90 seconds. It does not even have a manual backup rate setting. What this means is that if someone is seriously sick and their breathing pattern is all over the place, Resmed algorithm doesn't care and takes that pattern as its baseline:-)...So good luck if you average like 6 breaths per minute because Resemd ain't gonna give you more than that, if you stop breathing:-)

From everything I have seen of the DreamStation and ResMed APAP machines the ResMed is the superior one. I don't have much experience with the ASV machines. From the few ASV SleepyHead charts the ResMed ASV is like magic. It just makes the apnea events disappear. It is like the machine is programmed to not allow apnea (as detected by the machine) to occur, and that may actually be the simple explanation of it.

It seems that way. I think it is primarily because online forums have been littered with DME's and other parties who have an interest in Resmed products and since Philips Respironics are currently on recall, these operatives use the opportunity to flog as many Resmed machines as they can with dubious and favourable claims. Fact is however that if you do not use a Ozone cleaner to sanitize your machine, you have very little to no worries about the "foam" related recall. In fact I have used this opportunity to purchase brand new Dreamstations online and massively discounted prices, for spare and just in case they decide to discontinue this very awesome product line.

So this is an example of how Resmed fails to register Apnea events. The graph is a slice of about 12 seconds as seen on the time stamp.

"Flow Rate" graph clearly shows no movement in the flow (no air in or out of the upper airway, even though the upper airway is wide open) for over 10 seconds. According to protocol, no flow change over 10 seconds should be recognized as an apnea event, a cessation of breathing or "Clear Airway"...or "Central Apnea event", yet Resmed registers no event whatsoever. In fact that night I had at least 80 (eighty) or so of such events however Resmed only registered ONE:-)

Resmed algorithm appears to have been designed to ignore all events that last less than 15 seconds but international apnea protocol is clear: "A cessation of breathing, lasting up to TEN seconds, is an Apnea event"

Of course I must post an example of how my Philips Respironics Dreamstation handles an apnea event and how it catches absolutely all suspect events. Below is a 22 seconds slice of a 5 hour sleep session which should a cessation in flow rate (breathing stop). Graph shows the machine issuing a breath however flow does not change significantly. The machine issues another breath and registers a Hypopnea event shortly after. My Dreamstation does this all night. In fact I cannot find a single event that has been missed by the missine, throughout.

I just posted a graph above that you should look at. Yes, with regular CPAP therapy provided for the treatment of OSA, it doesn't really matter because all you need is some positive air pressure to maint your upper airway open, so your machine algo does not need a breath by breath analysis, even though I personally think that it should, for efficiency and also comfort of the patient so they only receive positive pressure when their airway is open.

In the case of CSA treatment, the airway is always open , but the patient simply stops breathing (at times), so providing a "Continuous Pressure" is pointless . So the algorithm must monitor breathing and ONLY provide a breath when one is missing (a cessation of breathing lasting more than 10 seconds)...This is why ASV machines are used for the treatment of complex or CSA patients, so again I think that OSA patients too will benefit from ASV therapy just as much, in that they won't have to receive a "continuous Positive" pressure when their upper airway is open.

Anyway, all this means that I am better able to test a machine algorithm for efficacy, than I would be examining a regular CPAP machine that is always ON , providing a positive pressure.

Well there is a clear scientific observation that connects patients with CHF to CSA. So they believe that CHF can lead to CSA and vice versa. Some scientists want to start studies into whether or not the development of CSA is a natural response to CHF in order for the body to moderate its PaCO2 levels...THAT is way out in the woods for me:-)

Studies also show that "treatment emergent CSA" is a thing, meaning that the long term CPAP therapy for OSA, leads to the development of CSA. (Remember that CSA itself was only isolated as a disorder back in 2008 at Mayo clinic so it is a relatively new subject and under intense study, internationally.) BUT I have also seen studies that suggest " treatment emergent CSA" will go away by itself within 6-8 weeks, after OSA CPAP therapy stops, but who would stop their therapy, unless they get an "implant surgery", like "INSPIRE" at a cost of like $30 K.??

This being said, it does not mean that people can't just develop Central Apnea without CHF (I am an example of that). There is also a clear scientific consensus that "UNTREATED CSA or OSA" will eventually lead to all kinds of Cardiovascular issues, including and not limited to CHF, however it is believed that the long term use of opiates can also be a driver of CSA.

As to O2 therapy for patients with CHF and Centrals, it makes a lot of sense because CSA folks (again like me) can watch their blood oxygen saturation level (SpO2) drop significantly overnight, if they are not using ASV therapy. I would imagine that supplemental Oxygen used in conjunction with PAP therapy for this group (CHF+CSA) would be very efficient and beneficial because CJHF itself is a cause for lowered SpO2.

We also have both, a Resmed Airsense 10 like yours and a Dreamstation-1 APAP which my wife uses for he OSA. (and no, she does not use the "FOR HER" feature of the Resmed for good reasons...See below) She finds the difference between the 2 machines, like night and day in that she now refuses to use her Resmed altogether.

1- Resmed Airsense algo does what Resmed wants to do, not what the patient wants it to do and thus its response to patient comfort is very limited, even when the EPR is turned on, where as according to my wife, she doesn't even know she is wearing a mask when she is on her DS-1. That is because Dreamstation algorithm to treat OSA is not only more responsive during inspiratory-Expiratory transition, but also much smoother in when changing pressures, unlike Resmed which according to her tends to have a much sharper (jerkier)n transition. so it is not that DS is "slower" than Resmed but that the DS algo makes the transition between IPAP and EPAP a lot smoother so patients can't feel the change with Dreamstation patented C-Flex technology..

2- Resmed's Auto PPAP algio essentially fakes mask pressure on data because they do not have separate sensors for the mask pressure built in, so they just derive this from calculation. The displayed mask pressure therefore is hypothetical, not an actual one, so I wouldn't place too much emphasis on that feature.

3-In contrast, Dreamstation-1 APA reports FLOW rate, precisely measured as you can see in my wife's chart. Her flow rate is perfect breath by breath, in comparison to your Airsense chart which reports a flow rate that is far from perfect, not necessarily because you are not breathing well naturally, but because you are not breathing well BECAUSE of Resmed's faulty pressure delivery. Worth looking into.

4-"For Her" algorithm that Resmed keeps coming up with (yes they have them on their latest model 11 too), is more of a sales gimmick than it is a real medical feature.A pap machine algo should respond to one's respiratory needs, regardless of their gender. Dreamstation machines algos are "Gender-Blind" and this even more "politically correct":-)

5-Her Dreamstation also has all the comfort settings that her Resmed has, including a cool Ramp feature that truly "RAMPS" the pressure up over the period that ramp has been set to, unlike how Resmed controls Ramp which at times jolts her awake. Ramp function should be true to the meaning of the word, a linear increase in pressure, divided precicely over TIME setting, which is what Dreamstation delivers.

At any rate, as the original inventors of PAP technology and their sizeable investment into the development of these essential medical devices, it is natural that Philips-Respironics software has been worked on a lot longer and naturally at a much deeper level than Resmed systems. Problem here is Resmed algorithm clearly does NOT recognize clear apnea events and ignores them, thus reporting a much lower AHI while not delivering full therapy. This is simply wrong .