Accuracy of CPAP AHI scores?

[Slinky]

Just an inconsequential "aside" remark but I think Resmed's choice of "green" for the acceptable leak "smiley" is rather "ironic"? Don't we usually use a green face to indicate "urgh! I'm sick!"?? Even given our Smiley choices here in the forum I use Mr Green "smiley" to indicate I'm laughing myself sick and use him AFTER several Laughing "smilies".

My personal opinion of Resmed's decision to use the 2 smilies to indiate Leak w/the S9s was asinine. And insulting.

God bless and long live my S8s!!!!

[scrapper]

Pugsy wrote: With avi123's pressure I would be suffocating or at least feel like it. Not enough air moving...

CPAP at 7 cm with EPR of 2. Approximately half the time the pressure is at 5 cm because of the reduction during exhale.

Everyone's pressure is different and dependent upon their throat structure, sleep position, etc...........that pressure at which you suffocate at, someone else does just fine. That factor is irrelevant in this discussion--except to you of course.

[robysue]

scrapper,

thanks for pointing out that some of us do indeed only need rather low pressures to keep our airway open.

Now that my BiPAP's pressures have been lowered to 7/4, I'm finally (some 5 months into therapy) beginning to thrive on it. And making progress with the insomnia as well.

[avi123]

Thanks Ms. Butterfly.

Here are my latest 2 weeks results.

My AHI has never been below 10

As I said before, it's my humble opinion that this does NOT (yet) represent effective treatment of your OSA. If I were you, I'd be on the phone to the doctor saying that the CPAP has only brought my apnea down to a level that would be diagnosed as "Mild--Moderate Apnea."

Statistics:

2/5/2011 - 2/18/2011

Device Settings:

Therapy Mode: CPAP Set Pressure: 7.0 cmH2O EPR: Full_Time
EPR Level: 2.0 cmH2O

Leak - L/min:
Median: 3.6 95th Percentile: 26.4 Maximum: 45.6

Even though Mr. Green Smiley keeps showing up, these leak rates ARE significant. The S9 DOES subtract the expected leak rate of the mask off the leak numbers. Hence for 5% of the night you are leaking at a rate of 26 L/min ABOVE your mask's expected leak rate. Leaks above 24 L/min can interfere with therapy, although the Resmed engineers only have Mr. Red Frowny show up if you are leaking at 24L/min for 25% of the entire night---i.e. an average of 15 minutes of large leaks every hour.

AHI & AI - Events/hr:

Apnea index: 14.8 AHI: 16.6 Obstructive: 11.5
Central: 2.8 Unknown: 0.7 Hypopnea index: 1.7

If these were my numbers night after night after night, I'd call the doctor. While we all have bad nights now and then, if this is the 14-day average (based on the rest of the data provided), that's long enough to be a disturbing trend in my opinion. And when you put this together with avi's original statement that his AHI often is above 20, that's down right alarming in my humble opinion.

p.s. I still don't know if the S9s show leaks after deducting the Venting leak or not. But I do see a Green face every day.

Read the owner's manual. It clearly tells you the leak rate is computed AFTER subtracting the expected vent rate. Otherwise there would be no way that anybody could ever have median, 95%, or maximum leak rates of 0.0.

Reply:


Hi I agree but it takes time to see the AHI indices drop. My clinician with a BS degree in respiratory therapy said, two days ago, when she downloaded my data to send to one of my Internist/Geriatric {NOT the Sleep Doc} to disregard all of the indices from the CPAP and just go by my feeling which is fine considering my other ailments such as arthritis, diabetes, neuropathy, etc.

What about responding to my excerpts from reliable sources about those CPAPers with AHI below 5 with only OSA? Could it be that they are mislead?

Question: if I go to sleep for 30 min at 1 p.m. (or 9 a.m.) with the CPAP, how does it calculate the AHI?

[billbolton]

So to repeat, a person with an AHI>20 is having 20 apneas an hour, each event lasting at least 10seconds, and more than 160 apneas in an average 8hr sleep period, right? I just wanted to see it in black/white.

AHI is a score of apneas plus hypopneas, so its not necessarily 20 apneas an hour, its 20 events which may be either apenas or hypopneas.

For example:

• 19 hypopneas + 1 apnea = AHI of 20
  
  10 hypopneas + 10 apneas = AHI of 20
  
  1 hypopneas + 19 apneas = AHI of 20

Cheers,

Bill

[rested gal]

If I were using an S9 Elite CPAP at these settings:
Pressure: 7.0 cmH2O
EPR: Full_Time
EPR Level: 2.0 cmH2O

And I saw this kind of Obstructive Apnea index for the past two weeks:
Obstructive: 11.5

I would immediately change my S9 Elite's CPAP pressure to 9 cm H2O and leave EPR set at "2" Full Time.

I'd look at the "Daily" results in the machine window each morning for at least three nights at that higher pressure...hoping (and expecting) to see a considerable reduction in the Obstructive Apnea index.

OR

If I wanted to leave the CPAP pressure at 7 cm H2O, I'd turn EPR Off.
If using CPAP pressure 7, EPR Off gave me a reduction in the Obstructive Apnea Index but the Obstructive Apnea index was still not WELL BELOW "5" , I'd still keep EPR Off and raise the CPAP pressure to 8 or even 9.

If raising the pressure to 8 or 9 (with EPR Off) gave me an Obstructive Apnea index of just "1.0" or less (preferably zero Obstructive Apneas) but I wanted to go back to using EPR for the comfort of less pressure during breathing out, I'd then experiment with different levels of EPR, and would set the pressure at a level to compensate for whatever EPR reduction level I chose.

But that's just what I'd do, and I'm not a doctor.

[SleepingUgly]

My AHI values from the machine range from 4 to 8. While that doesn't sound too bad, I am one who has nearly all my apneas/hypopneas when I try to enter REM sleep. (REM AHI of 30-40 from my sleep studies).

IMHO, in the case of REM-dependent (or mostly dependent) OSA, the overall AHI is not a good index to use. One could have a very high number of respiratory events during REM, none in NREM, and end up with a "normal" AHI because all the events are averaged over the entire night's sleep. (This is an instance in which the mean is not a good measure of central tendency). If you had data, you could possibly see whether your respiratory events are clustered together during parts of the night. It may be that you need a higher pressure to ward off events in REM than you do in NREM.

[SleepingUgly]

Hi I agree but it takes time to see the AHI indices drop.

If that were true, an in-lab titration would never work! This isn't like a medication that has to build up in your system.

That said, I've never understood what effect high leaks have on the AHI. Do they raise it? Lower it? Or just render it invalid and uninterpretable?

Avi, make sure your machine is set to the type of mask you have. If memory serves, you're using a full face mask, so your machine must be set to that so that it can subtract the expected leak rate. Once that's set right, your leaks should be 0, ideally. If you have it set to nasal pillows and you're using a full face mask, it's not going to calculate this appropriately.

[DoriC]

Hi I agree but it takes time to see the AHI indices drop.

IThat said, I've never understood what effect high leaks have on the AHI. Do they raise it? Lower it? Or just render it invalid and uninterpretable?

My husband is never aware of large leaks but they do wake me up, and I have observed that he is sometimes having apneas before I can get to adjust his mask. Yet when I check his data for that time period no events are shown. So I would think LLs render AHIs invalid.Fortunately it doesn't happen too often now.

[robysue]

What about responding to my excerpts from reliable sources about those CPAPers with AHI below 5 with only OSA? Could it be that they are mislead?

Avi, CPAP is recommended for patients with an UNTREATED AHI or RDI of 5 or above. The AHIs below 5.0 that you keep seeing people referring to in their posts are TREATED AHIs---not their untreated AHIs. The goal of CPAP therapy is to take a person with a high UNTREATED AHI and reduce their AHI through CPAP therapy to normal levels---i.e. to get their TREATED AHI below 5.

You have diabetes. Do you check your blood sugar levels on a regular basis? The TREATED AHI number is rather like your blood sugar level ON DIABETES MEDICATION. The UNTREATED AHI is like the blood sugar reading on the blood sugar test you took BEFORE you started taking diabetes medication. Would you become concerned if your daily blood sugar readings while ON DIABETES MEDICATION looked just like the blood sugar readings that are used to DIAGNOSE diabetes in UNTREATED patients? In other words, would you be concerned if your daily blood sugar readings were somewhat below your untreated blood sugar level, but were also consistently way above the target range?

As for why my diagnostic AHI = 3.9, but I still have MODERATE sleep apnea: Some labs ---such as mine--- score hypopneas under both of the two different criteria (technically 4.A and 4.B) of the AASM scoring manual. For reasons I do not understand, hypopneas under criterion 4.B are not added to my lab's AHI and only show up in the RDI. In other labs, the hypopneas scored under 4.B are added to the AHI. So that's why my diagnosis is based on the total RDI in my signature: Most of my untreated OSA events are hypopneas under AASM scoring criteria 4.B: They involve a significant reduction in air flow into my lungs with continuing effort to breath and they cause arousals (and hence distrupted sleep), but they fail to cause an oxygen desaturation. Please understand that the lab that I was diagnosed at scores these hypopneas as part of the RDI but not AHI, and taking that into account, I have MODERATE sleep apnea since my RDI = 23.4---i.e. I have a traditionally scored AHI of between 15 and 30 on a full night PSG.

Question: if I go to sleep for 30 min at 1 p.m. (or 9 a.m.) with the CPAP, how does it calculate the AHI?

The S9 resets the daily data at NOON each day. So if you did NOT sleep at all from NOON to 1:00 pm, and then you sleep for 30 minutes at 1:00 pm, and then CHECK your numbers immediately after the nap ends at 1:30, the S9 will count the number of events (apneas + hypopneas) and divide that number by 0.5 since 30 minutes equals 0.5 hours. As an example: Suppose you had a total of 4 OAs, 1 CA, and 3 Hypopneas during that 30 minute nap. The AHI on the LCD at the end of the nap will be computed as follows:

(4 + 1 + 3)/(0.5) = 8/(0.5) = 16.

The AHI represents the average rate the apneas and hypopneas occur and its units are events per hour. Eight events in 1/2 hour are occurring at a rate of 16 events per hour. Hence the AHI for the nap is 16.

Now note, the S9 resets the daily data at NOON. So the events and time for that 30 minute nap will also be added into the computation of the overnight AHI for the night that follows the nap as well the events that occur during that night.

Now a 30 minute nap at 9:00 am occurs BEFORE the nightly data has been reset. So if the nap occurs at 9:00--9:30 am, the events and time just get added onto the nightly data. So for example, let's suppose that you had a total of 100 apneas and hypopneas during the eight hours you sleep overnight and that you woke up at 6:30. At that point in time, the S9's LCD will report an AHI = 100/8 = 12.5. At 9:00 you take a thirty minute nap and have another 8 events. When you look at the LCD at the end of this nap, the LCD's AHI is computed as:

(100 overnight events + 8 new nap events)/(8 hours of night sleep + 0.5 nap sleep) = 108/8.5 = 12.7

The AHI = 12.7 means that during the 8.5 TOTAL hours of sleep, your apneas and hypopneas were occurring at an average rate of 12.7 events per hour. The AHI is an average rate---that means that during part of the sleep time (such as the nap) the rate the events are occurring is GREATER than the AHI and at other parts of the sleep time, the rate the events are occurring is less than the AHI. But it all averages out to the AHI.

To better understand how the AHI is an average rate, it may help to build an analogy: The AHI is like the average speed a car is going on a long trip. If you drive 420 miles on the interstate, and drive 70 miles per hour, but also stop twice during the drive---once for an hour for lunch and once for 15 minutes to go to the bathroom and stretch the legs, the whole trip takes 7.25 hours (6 for the actual driving and 1.25 for stops). While your speed is roughly 70 miles per hour during the time you are driving, your speed is also 0 mph during the time when you are stopped. And so your average speed for the whole trip is

(420 miles driven)/(7.25 hours for the WHOLE trip) = 420/7.25 = 57.9 miles per hour

[avi123]

RG thanks, I am going to try tonight exactly as you posted:

If I were using an S9 Elite CPAP at these settings:
Pressure: 7.0 cmH2O
EPR: Full_Time
EPR Level: 2.0 cmH2O

And I saw this kind of Obstructive Apnea index for the past two weeks:
Obstructive: 11.5

I would immediately change my S9 Elite's CPAP pressure to 9 cm H2O and leave EPR set at "2" Full Time.

{sorry, first night I'll raise to 8 cm and if OK I'll raise the subsequent to 9 cm}

I'd look at the "Daily" results in the machine window each morning for at least three nights at that higher pressure...hoping (and expecting) to see a considerable reduction in the Obstructive Apnea index

Ms. Avatar, you asked: That said, I've never understood what effect high leaks have on the AHI. Do they raise it? Lower it? Or just render it invalid and uninterpretable?

IMO, the concern is valid only if the air pressure inside the mask during inspiration has fallen below treatment pressure. At that point since the AHI components are affected by the pressure the AHI would then be changed one way or another. But let's face that a large leak occurs only when we lift the mask off our face when the CPAP is ON or while having mouth leaks. All the rest of the times since we hear and feel the leaks and correct it the effect should be minimal.

As to having different AHIs during REM and NREM in different persons, you are correct.

Read this:

Half of patients with OSA have a higher NREM AHI than REM AHI

Link:

http://www.ncbi.nlm.nih.gov/pubmed/16564214

[robysue]

As to having different AHIs during REM and NREM in different persons, you are correct.

Read this:

Half of patients with OSA have a higher NREM AHI than REM AHI

Link:

http://www.ncbi.nlm.nih.gov/pubmed/16564214

Not news to me: I'm one of those folks with a higher NREM AHI than REM AHI.

[avi123]

Why have you changed the butterfly avatar from micro to macro?

I agree on this:

You have diabetes. Do you check your blood sugar levels on a regular basis? The TREATED AHI number is rather like your blood sugar level ON DIABETES MEDICATION. The UNTREATED AHI is like the blood sugar reading on the blood sugar test you took BEFORE you started taking diabetes medication. Would you become concerned if your daily blood sugar readings while ON DIABETES MEDICATION looked just like the blood sugar readings that are used to DIAGNOSE diabetes in UNTREATED patients? In other words, would you be concerned if your daily blood sugar readings were somewhat below your untreated blood sugar level, but were also consistently way above the target range?

In this case the correlation is simple: The prevalence of Sleep Apnea in those suffering from Diabetes (controlled or not) is almost 50%. See it here:

Prevalence of sleep apnea in a population of adults with type 2 diabetes mellitus.

http://www.ncbi.nlm.nih.gov/pubmed/17669711

So even if my Diabetes is controlled with pills (Actos 30 mg daily) you should NOT be surprised with my AHI of 15 to 20 (CPAP treated). The question is if it's worth trying to lower it by raising the CPAP pressure (as RG suggested) b/c of the side effects?

[schnarchen]

My AHI values from the machine range from 4 to 8. While that doesn't sound too bad, I am one who has nearly all my apneas/hypopneas when I try to enter REM sleep. (REM AHI of 30-40 from my sleep studies).

IMHO, in the case of REM-dependent (or mostly dependent) OSA, the overall AHI is not a good index to use. One could have a very high number of respiratory events during REM, none in NREM, and end up with a "normal" AHI because all the events are averaged over the entire night's sleep. (This is an instance in which the mean is not a good measure of central tendency). If you had data, you could possibly see whether your respiratory events are clustered together during parts of the night. It may be that you need a higher pressure to ward off events in REM than you do in NREM.

This is why it was difficult to get the insurance company to agree to cover CPAP. My AHI was only 3.5 overall, but effectively was missing out on REM sleep. On treatment, my AHI was 1.75 which is much lower than what I'm seeing by the ResMed results. I would really like to see the graph of events by time to see if the machine is picking up my REM problem or something different.

I plant to go to the DME soon to review my data and try some different masks, and will also try to get a data cable. Failing that, I suppose I could try to return the S8 (it is currently on rental) and request a S9 so I can download the data.

[avi123]

My AHI values from the machine range from 4 to 8. While that doesn't sound too bad, I am one who has nearly all my apneas/hypopneas when I try to enter REM sleep. (REM AHI of 30-40 from my sleep studies).

IMHO, in the case of REM-dependent (or mostly dependent) OSA, the overall AHI is not a good index to use. One could have a very high number of respiratory events during REM, none in NREM, and end up with a "normal" AHI because all the events are averaged over the entire night's sleep. (This is an instance in which the mean is not a good measure of central tendency). If you had data, you could possibly see whether your respiratory events are clustered together during parts of the night. It may be that you need a higher pressure to ward off events in REM than you do in NREM.

This is why it was difficult to get the insurance company to agree to cover CPAP. My AHI was only 3.5 overall, but effectively was missing out on REM sleep. On treatment, my AHI was 1.75 which is much lower than what I'm seeing by the ResMed results. I would really like to see the graph of events by time to see if the machine is picking up my REM problem or something different.

1) CPAP machines are not suitable to diagnose Sleep Architecture and show Stages of Sleep such as REM. It was done on you during your Sleep Study. Costs of those machines run into the hundreds of thousands of dollars.

2) Once you get an Rx from your doctor the insurer or the DME can't argue about its details .

3) The S8 series are using Resmed cards (elongated). To read those cards you need a special reader and a cable to connect the reader to the laptop USB. You also need to download a software. Your DME will tell you that they don't supply any of these (except a second elongated card). See the card reader here:

https://www.cpap.com/productpage/resmed ... -cpap.html

Your sleep doctor is the key.

Also, check Judy's post here: http://www.apneasupport.org/about25693.html

I plant to go to the DME soon to review my data and try some different masks, and will also try to get a data cable. Failing that, I suppose I could try to return the S8 (it is currently on rental) and request a S9 so I can download the data.