Accuracy of CPAP event classifications

[ColoradoMom]

Following a sleep study my 16 year old son was diagnosed with mild to moderate obstructive sleep apnea & hypoxemia. Half of the events during the diagnostic sleep study were central apneas and the other half were hypopneas. My son started using CPAP in November. Sleepyhead's data shows that since starting CPAP my son's events are predominately hypopneas. Most nights he'll have a couple of clear airway events, but the clear airway index is almost always below 1. My son had a titration study about a month ago. His oxygen levels showed improvement (average O2 sat was 91%) and his sleep apnea fell into the treated category with an AHI of 4.9.
I was surprised to see that once again the sleep study showed a predominance of central events, which does not fit with his machine's data which has consistanatly shown hypopneas and OA making up the vast majority of events. The Tech's notes from the titration study stated "He slept in the supine position for most of the night with no evidence of obstructive events, however unexplained arousals interrupted sleep continually." 

Diagnostic sleep study:
AHI: 13.9
CAI: 6.9
OAI: .1
HI: 6.9

Titration study:
AHI:4.9
CAI: 3.7
HI: 1.2

2 recent nights of Dreamstation numbers:
AHI- 6.8              AHI-  4.8
CI- .7                  CI- .4
OAI- 3                 OAI-  .8
HI- 3                   HI- 3.7


In a previous thread it was suggested to me that BiPAP might be a more appropriate treatment for my son. Since my son's numbers fall into the treated range I'm guessing a BiPAP would not be approved at this time, however I plan to keep a watchful eye on my son's data. I had been relieved that few centrals were showing up on Sleepyhead, but now I wonder if perhaps he may be having centrals that the dreamstation is not classifying correctly. How accurate are the machines classifications of clear airway & obstructive events?

[Ron AKA]

2 recent nights of Dreamstation numbers:
AHI- 6.8              AHI-  4.8
CI- .7                  CI- .4
OAI- 3                 OAI-  .8
HI- 3                   HI- 3.7


In a previous thread it was suggested to me that BiPAP might be a more appropriate treatment for my son. Since my son's numbers fall into the treated range I'm guessing a BiPAP would not be approved at this time, however I plan to keep a watchful eye on my son's data. I had been relieved that few centrals were showing up on Sleepyhead, but now I wonder if perhaps he may be having centrals that the dreamstation is not classifying correctly. How accurate are the machines classifications of clear airway & obstructive events?

First I don't think your son would be an obvious case for BiPAP therapy when basically only 6 cm of pressure is being used for treatment. On classification of events, I don't have any experience with Dreamstation other than what I see in posted results here. I think the devil is in the detail, and you really have to see the expanded views of the two types of events, and most often posters do not provide that. I have looked at this kind of detail in my ResMed A10 results and I do question the accuracy of classification at times. ResMed watches the event for I believe 6 seconds and at that point inject a 4 Hz pressure flow oscillation. If the flow produces a corresponding pressure oscillation then they class it as an OA. The flow produces pressure against a closed door. And if it does not produce a pressure oscillation they class it as a CA event. That all sounds great, but when I look at the events in detail is sometimes see the events alternate between OA and CA and they look pretty much the same. I think at times they may be splitting hairs. I believe that Dreamstation uses a similar technique but they call it a pulse rather than an oscillation. Perhaps just semantics important only in patent protection, but essentially the same. 

My thoughts would be to increase pressure slowly and see what happens. If AHI comes down, it is probably a good thing. If AHI goes up or remains unchanged, then not such a good thing, and pressure should be reduced again.

[Sleeprider]

First, Ron is assuming that "BiPAP" mean a standard bilevel machine, and what I think needs to be considered is a bilevel with backup rate, or ASV (adaptive servo ventilation). The ASV is used for individuals that exhibit CPAP onset central apnea and hypopnea. The main criteria for approval is that he is diagnosed with central apnea, as happened in the titration study; that it would be medically necessary, again true due to hypoxemia; and that ASV is shown to have efficacy in the treatment of apnea and hypopnea. That remains to be determined, and would require a bilevel/ASV titration ordered by his doctor. I will add that an average SpO2 is very low, and his oxygen nadir and its duration is very important. i assume with an average SpO2 his minimum drop below 88% which is a threshold for prescribing supplemental oxygen, or considering other treatment alternatives.

It would probably be very revealing to see some #Sleepyhead data of your son's CPAP data. I strongly suspect he has what we call complex apnea, and there is a good reason to pursue ASV as a treatment goal. Again, the charts would help as they give a lot more detail on machine settings, therapy response and other factors.

Let's talk about ASV, or what your doctor or therapists may be referring to as "BiPAP". A bilevel machine simply provides two different pressures for inhale (IPAP) and exhale (EPAP). This pressure difference makes it easier for some people to inhale and exhale, so it's comfortable, and it can improve ventilation in those that need it. A bilevel ASV is a machine that also maintains the breathing rate and volume on a breath by breath basis. When a patient breaths more shallow, the ASV increases the pressure during inspiration to increase the breath volume. When he stops breathing during a central apnea, the ASV provides enough pressure to inflate the lungs with no effort on the part of the patient. So the ASV can provide CPAP treatment using positive pressure at low pressure to maintain the airway and prevent obstruction, but it also provides pressure as needed and when needed to maintain the breathing rate and volume.

Let's take a closer look at what is going on, and help you to help your son get the care he probably needs.

[Ron AKA]

Unless things have changed since you posted SleepyHead charts in March, it would seem your son does not have a central apena problem. Most times it is under 1.0. Hypopnea and obstructive apnea are more of a problem. I don't think there is any way that your machine could class a central apnea event as a hypopnea event. Both hypopnea and obstructive apnea are most directly corrected with more pressure. My suggestion from March remains the same. I would put the machine in Auto mode and give it some room to adjust pressure. Alternately you could just increase the pressure in CPAP mode. If more pressure reduces the OA and H events then it essentially verifies they are real and properly classified. This said if oxygen levels are the main concern, I'm not sure that CPAP is going to have all that much potential to improve them. Perhaps oxygen supplementation might be the more direct route to bringing O2 levels up...

[SarcasticDave94]

"...Diagnostic sleep study:
AHI: 13.9
CAI: 6.9
OAI: .1
HI: 6.9..."

The titration was following a similar pattern AHI number wise.

Certainly if I'm incorrect others will put me in line. I see the above info, if CAI refers to indicate centrals and not clear apnea, they count for about half these sleep study events. Hypopnea about the other half. That would indicate to me a qualification for an ASV sleep study, which is needed to move onto what I believe is the right tool (ASV) in this case. I say that because this situation sounds close to my own, obviously not exactly the same, but it seems to fit the pattern.

As was mentioned already, post a chart via Sleepyhead. I'm sure there will be a more certain answer as more data would be telling a better idea of what is taking place.

Sincere best wishes to you and your son on getting what is required to treat his therapy needs.

[ColoradoMom]

My son had his titration study in late March the and sleep tech filled me in on the findings the following morning. At that time I sought advice on this forum because I was considering making changes to my son's pressure myself. The tech had told me my son's oxygen levels showed improvement with a higher pressure. The tech also recommended that I utilize his machine's auto mode. The sleep lab does not score upper airway resistance, but the tech believed that my son's many unexplained arousals may have been caused by upper airway resistance. The tech told me that APAP mode would help resolve the upperairway resistance. After reading thru many helpful replies to my last thread in this forum I decided to be patient and wait for my son's doctors recommendations. 
So in answer to Ron's question, the only thing that has changed since my post in March is that this week they finally got around to reading my son's titration study. His doctor has put in an order to increase his pressure to 8. They will reevaluate after a couple of weeks and may raise the pressure to 9. We have to wait until July for my son's follow up appointment with his sleep doctor. 

My impression from talking to the nurse at my son's doctors office is that they seem to feel that the hypoxia has been resolved with the higher pressure. My primary concern with this determination is that the data set they are extrapolating this conclusion from was collected over a very short amount of time. During the titration study they had my son at pressures of 5 & 6 for the vast majority of the night. 
His oxygen saturations did greatly improve at 8 and 9, but unfortunately they had him at those pressures for a short amount of time. They only had him at 9 for a grand total of 31 minutes. At 9 his mean O2 sat was 93.7 and he had 4 centrals in the 31 minutes he was at 9. They had him at 8 for 62 minutes and at that pressure his mean O2 was 93.4 and he had 3 centrals and 2 hypopneas.  Fortunately I do have a recording oximeter, after his pressure is raised I plan to monitor his O2 levels to see if that 31 minutes of titration is truly representative of the entire night. Based on the titration results the worst pressure was 6. At 6 he had his lowest oxygen saturations of the night and his AHI was 16.3. Based on the patterns I have seen in Sleepyhead my guess is that the bad numbers at 6 are likely the result of sleep stage, not pressure.

I have attached some sleepyhead shots. Which brings me back to my initial question, how accurate are the machines classifications of different types of events? As you will note very few clear airway events typically show up on my son's sleepyhead reports. I had been operating under the assumption that the centrals were in large part resolved, but because centrals were the predominate event in my son's sleep studies it makes me question whether or not the machines classifications of events are accurate.

Thank you to you all for the wealth of information and support!

[DeepBreathing]

Which brings me back to my initial question, how accurate are the machines classifications of different types of events?

Typically the machines are thought to over-diagnose central apnea (ie false positives) due to pauses in breathing which may not be apneas at all. Examples can be rolling over in bed while holding your breath. Note that this is anecdotal and I haven't personally seen any studies on this. If you zoom in closely on the flow chart the type of apnea can usually be determined by the shape of the wave-form.

There's also the possibility that your son is experiencing brief pauses in breathing which don't quite reach the 10 second threshold to be scored as apneas. SleepyHead has a facility to detect and record this type of event - go to File | Preferences | CPAP |Custom CPAP User Event Flagging. You can set two different thresholds for flow reduction and duration, and SleepyHead will then show them on the charts.

I agree with you that titration studies are at best a snapshot in time, and I think you're doing exactly the right thing with your ongoing monitoring of his situation. It's the trends over time which are important, not a single night in isolation.

[ColoradoMom]

Thank you DeepBreathing, I will play around with resetting the thresholds for flow reduction and duration. What values would you suggest I enter for the flow and duration?

I have read thru the Sleepyhead tutorial and have gained a very basic understanding of the flow charts, but more often than not when I zoom in on events I don't really know how to interpret what I'm seeing. I have attached a couple of closer views that are all from the same night. The first shot is of a couple of events early in the evening that look similar to the waveform examples of apneas that I have seen. The next shot is a closeup of his breathing later in the evening before he has a number of events. The next 3 shots are closeups of a cluster of events. Do those of you who are skilled in reading flow charts see anything notable in any of the closeups?

[Ron AKA]

I have attached some sleepyhead shots. Which brings me back to my initial question, how accurate are the machines classifications of different types of events? As you will note very few clear airway events typically show up on my son's sleepyhead reports. I had been operating under the assumption that the centrals were in large part resolved, but because centrals were the predominate event in my son's sleep studies it makes me question whether or not the machines classifications of events are accurate.

I think about the only way you can second guess what the machine classifies as an event is to look at them in detail, one by one. I took a screen shot snip from my Friday night sleep that included three types of events; CA, OA, and H. You can see that with the H events there is no stoppage in flow, just a reduction. I don't believe a machine would mistake a H event for CA or OA. However telling the difference between a CA and OA is more difficult. One sign is that an OA event usually ends with a gasp or large air flow cycle when the obstruction clears, whereas a CA can just start up again slowly, without a gasp as the airway is clear. The way the machine does it is oscillate the air flow and then measure how much of an oscillation in pressure you get in response. If the airway is closed you will get a larger oscillation than if it is open. In the image I have included you can see the difference in the height or amplitude of the oscillation when the machine is testing to see what it is. 

On the treatment side, I agree it may be better to go with a fixed pressure increase rather than APAP. The reason I say this is that it seems your son may need more pressure just what is needed to stop the obstructive apnea, and the Dreamstation machine is not so quick to respond with pressure increases and seems overly quick to reduce pressure again. For that reason a fixed pressure may be more effective. I however would try it in steps, of 1 cm. It is likely you will have to try different levels to see what works best. In my opinion OA and CA events are equally as bad, and hypopnea events less bad. The machine will not reduce real CA events, only OA. More pressure than what is needed to reduce OA may be required for reduce the flow reduction hypopnea though. I would think a week at each pressure level should give you a good indication as to what works best, and more indicative than an overnight titration test.

[Ron AKA]

I forgot to include the snip, and website glitches prevent me from adding it. Here it is below: