Understanding the data in your sleep test report

I stopped breathing HOW many times each hour?
Understanding the data in your sleep test report

By RobySue

Introduction

I remember the only consultation I had with my first sleep doctor. It happened about a week after my diagnostic sleep test: He opened the conversation by waving my sleep report in front of my nose without really letting me see it and said, "You stop breathing about 23 times an hour. You have moderate obstructive sleep apnea. You need CPAP. You need it now. When can you do the titration study?"

Several weeks later I finally got my hands on the report itself and I learned the truth wasn't quite so simple: Technically, I actually stopped breathing only about 3 or 4 times an hour, but roughly 20 times an hour my airway was partially collapsing and the amount of air getting into my lungs was dropping significantly, and I was arousing myself in order to start breathing normally again before any clinically significant O2 desats occurred. And certain times of night---such as when I was first trying to get into deep sleep, were worse than others. And unlike most people with OSA, all my apneas occurred in NREM sleep and I more hypopneas in NREM than I had in an equal amount of REM sleep.

Sleep disordered breathing, which includes obstructive sleep apnea (OSA) and upper airway resistance syndrome (UARS) as well as central sleep apnea (CSA) and complex sleep apnea (CompSA), is diagnosed from data gathered during a nocturnal polysomnogram (NPSG or PSG) conducted in a sleep lab---i.e. a sleep study.

During a diagnostic sleep study the tech hooks you up with a lot of wires and also fastens a couple of belts around your chest and abdomen. All these leads are used to measure various things going on when you are asleep. The belts measure the effort you are making to breathe and the EEG determines whether you are actually asleep or awake. And if you are asleep, what sleep stage you are in. The sensors attached just under your nose are used to measure the airflow into and out of your lungs---i.e. whether you are breathing and whether you are breathing effectively. There are also leads to measure your heart rate, your blood oxygen saturation, and limb movements since these things are also important in determining how disturbed your sleep actually is.

And when you first get your hands on the dictated interpretation and the full summary data and graphs of your sleep studies, the numbers and acronyms can make it look like you're trying to read something in a foreign language. And all too often the sleep doc is like mine and over simplifies things and takes no time to explain what any of the numbers mean.

This essay is my attempt at a patient-friendly explanation of all those numbers. Please bear in mind that I'm only a patient myself, but as a mathematician, numbers and their meaning and computation are part of my life. So once I finally figured out what the heck a "hypopnea with arousal" was, making sense of the rest of my report wasn't that hard for me.


The types of respiratory events commonly scored in diagnostic sleep studies

NOTE: For any respiratory event to "count" towards the numbers used to diagnose SDB, the event has to start when you are asleep--as measured by the EEG. The reason why is simple: If we are awake, then we are in control of our breathing. And our wakeful breathing is much less regular than our sleep breathing is supposed to be: We sigh, we briefly hold our breath while deeply concentrating on something (perhaps as simple as rolling over in bed at night), we choose to breathe deeply for a minute or so as a form of conscious relaxation and then our breathing becomes much more shallow when we return to our normal breathing patterns. But the point remains: Even if we don't think we're consciously controlling our wakeful breathing, we are indeed in control of it.

An apnea is the complete cessation of breathing (the complete absence of air moving in and out of your lungs) for at least 10 seconds. Apneas only "count" in diagnosing sleep disordered breathing (SDB) if they occur when you are asleep---as measured by the EEG during a nocturnal polysomnogram (NPSG or PSG) conducted in a sleep lab---i.e. a sleep study.

An obstructive apnea is an apnea where there is still evidence that you are trying to breath---i.e. your chest and abdomen continue to move up and down in an effort to breath, but no air is getting into your lungs. The belts the tech puts around your chest and stomach are used to measure the effort to breathe. The assumption is that your upper airway has collapsed at some point and the collapsed upper airway is what is preventing air from getting into your lungs.

A central apnea is an apnea where there is no evidence (from the belts) that you are making an effort to breathe. In essence, the brain has forgotten to send the signal to breath to your diaphram and lungs. In a NPSG no effort is made to determine whether the airway is actually open or not during a central apnea because the "not breathing" is being triggered by a problem in the brain not the airway.

A mixed apnea is a combination of the two. It starts out looking like a central apnea---i.e. there is no effort breathe for a while, but once effort to breath starts, the airflow doesn't. So it ends looking like an obstructive apnea. The usual assumption about a mixed apnea is that it starts when the brain has forgotten to breathe, but the airway has also collapsed or perhaps collapses early in the apnea. So when the brain does start sending the signal to breathe, you start making an effort to breathe, but no air gets into the lungs because of the collapsed airway.

A hypopnea is when you are attempting to breathe (there is some effort to breathe as measured by the belts) but the air flow is significantly lower than it should be for at least 10 seconds. The usual assumption is that the diminished air flow in a hypopnea is caused by a partial collapse of your upper airway. It's often described as like trying to breathe through a very small straw.

Scoring of hypopneas on NPSGs is, unfortunately not uniform. The sleep docs diagnosing us with SDB all believe hyponeas should be counted---or at least some hypopneas should be counted, but which ones should count still seems to be a bone of contention.  Which is why we've got the confusing mess of the American Academy of Sleep Medicine (AASM) having two official definitions for scoring hypopneas: The AASM Recommended Standard and the AASM Alternative Standard for hypopneas. See http://www.ncbi.nlm.nih.gov/pubmed/19238801 and http://www.journalsleep.org/ViewAbstract.aspx?pid=27368 for a scholarly discussion of how and when these two standards affect a patient's diagnosis of OSA as well as the formal definitions of the two standards:

• AASM Recommended Standard: A hypopnea requires at least a 30% reduction in airflow for at least 10 seconds AND a corresponding O2 desaturation of at least 4%. Such a hypopnea does NOT require an EEG arousal

• AASM Alternative Standard: A hypopnea requires at least a 50% reduction in airflow for at least 10 seconds AND one or both of the following conditions: A EEG arousal OR a corresponding O2 desaturation of at least 3%.

Of course, under both standards, the hypopnea has to occur during an epoch when the patient is ASLEEP according to the EEG evidence.

It's my understanding that Medicare requires hypopneas to be scored under the AASM Recommended standard, but that the folks working in sleep medicine are increasingly leaning towards believing that the arousals triggered by large numbers of "Alternative Standard hypopneas" can be just as damaging to the body as the O2 desats triggered by the hypopneas scored under the Recommended Standard. And it's also my understanding that some sleep doctors (such as mine) are deeply concerned that Medicare insists upon a 4% desat to score an hypopnea.

The upshot of all this technical stuff about hypopneas is that when you're trying to interpret all the numbers and graphs on your sleep study, you need to know which standard the lab uses. Or, if the lab is like the one that did my sleep study: My lab scores hypopneas that meet the Recommended Standard as "hypopneas with desaturation" and count them in the computation of both the AHI and the RDI; it scores hypopneas that meet the Alternative Standard, but NOT the Recommended standard as "hypopneas with arousal" and it counts these "hypopneas with arousal" when computing the RDI but not the AHI. And hence, the formal diagnosis of OSA is based on the RDI.
[Much more about all these acryonyms in the next section.]

A respiratory effort related arousal (RERA) is sort of like a "hypopnea with arousal" scored under the AASM Alternative Standard for hypopneas. Only it may not last as long as 10 seconds. Or the reduced airflow may not be as great as 50%. But the critical idea is that there is evidence of increasing respiratory effort that indicates the airway is starting to collapse, but you arouse/awake yourself before the situation deteriorates to the point of being a full fledged hypopnea or apnea. Not all labs attempt to score RERAs because to score them correctly there needs to be a way to measure the pressure in your airway. And this requires additional equipment that is even less comfortable and more invasive feeling than all the normal wires attached to you during a NPSG.

RERAs are used in computing the RDI, but not the AHI for diagnositc purposes.

For folks who get diagnosed with UARS, the problem is way too many RERAs---often with no or very few hypopneas and apneas. Since some insurance companies balk at providing UARS patients with CPAP equipment unless there are severe daytime symptoms, some labs may err on the side of scoring at least some RERA type events as hypopneas with arousal.



AHI, AI, HI, RDI, and all the other numbers

The AHI, AI, OAI, CAI, HI, RDI and other numbers with an "I" at the end of them are all indices and they represent the average number of events per hour of time actually spent asleep---as documented by the EEG. In other words, all the indices are computed in the same basic way:

index = (number of events while asleep)/(total sleep time measured in hours) 

It is usual to round the index to one decimal place.

The total sleep time is usually abbreviated as TST on the sleep study report, but it is typically reported in minutes instead of hours on the report. But the TST must be converted to hours before computing the indices.

So we have the following specific definitions

• AHI = (total number of apneas and hypopneas)/(TST in hours)
• AI  = (number of apneas, regardless of type)/(TST in hours)
• OAI = (number of obstructive apneas)/(TST in hours)
• CAI = (number of central apneas)/(TST in hours)
• HI  = (number of hypopneas)/(TST in hours)

You may also see:

• REI (or RERA I) = (number of RERAs)/(TST in hours)
• RDI = (total number of apneas, hypopneas, and RERAs)/(TST in hours)

The lab that did my diagnostic test used the "hypopneas with arousal" scored under the Alternative standard to compute the RDI, but not the AHI. So you need to read your test results carefully to figure out exactly what kinds of events went into the AHI and how the hypopneas were scored to really make sense of your sleep test.



An example of sample sleep study data

To make all these numbers easier to understand, we need a working example. So let's look at data from a typical sleep study report.  We'll use the following convention:  The data written directly on the sleep study report will be shown in Bold Courier type that is easy to identfy.

Near the top of the sleep report's data, you'll probably see some figures like these:

TIB Minutes: 429.5
TST Minutes: 393
Sleep Efficiency: 91%

This means you spent a total of 429.5 minutes physically "in bed" (TIB = Time in Bed) and you were asleep (as measured by the EEG) a total of 393 minutes asleep (TST = Total Sleep Time). The Sleep Efficiency is just the percentage of time you were asleep while in bed:

Sleep efficiency = (TST/TIB) * 100%

The typical report will also include some information on how much time you spent in each sleep stage.  It may be presented as percentages (of either the TIB or the TST) or it may be listed in terms of minutes.  Or both.

The SDB events will typically be listed in a table broken down into REM/NREM and Supine/NonSupine categories.  But somewhere there will be some totals:

Apneas 158                   AI 24.1

Central Apneas 17           CAI  2.6
Mixed Apneas 18        Mixed AI  2.7
Obstructive Apneas 123      OAI 18.8
Hypopneas 366                HI 55.9 

AHI 80
RDI 80

In the 393 minutes you were asleep, you had 158 apneas---you quit breathing for at least 10 seconds 158 times.  The vast majority (123) of the apneas were obstructive, but 17 of them were central apneaa and 18 of them were mixed apneas.

You also had 366 hypopneas during the night. Recall that a hypopnea is where you "almost" stop breathing: The flow of air into your lungs drops significantly for at least 10 seconds (how much depends on the particular lab's scoring rules---remember there are two common sets of rules that are used) and there is either an associated drop in O2 saturation or an arousal or both.

The AI index (24.1) is the number of apneas divided by the time asleep in hours: 158/6.55 = 24.1 Note: 393 minutes is equal to 6.55 hours since 393/60 = 6.55. The AI is the average number of apneas per hour you had during the night. If this night is typical of your SDB, then in a typical 60 minutes of actual sleep, you have around 24 apneas.  Two or three of thes apneas are likely to be central and two or three of them are likely to be mixed since the CAI = 2.6 and Mixed AI = 2.7 are between 2 and 3.

The HI index (55.9) is the number of hypopneas divided by the time asleep in hours: 366/6.55 = 55.9 The HI is the average number of hypopneas per hour you had during the night.  Again, if this night is typical of your apnea, then in a typical 60 minutes of actual sleep, you have around 56 hypopneas.

The AHI Index is the sum of the AI and the HI: 24.1+55.9 = 80. It represents the average number of apneas and hypopneas per hour you had during the night. Your AHI = 80 means that ON AVERAGE, you stopped breathing (an apnea) or came close to stopping breathing (a hypopnea) about 80 times each hour. That works out to having an apnea or a hypopnea on average about every 45 seconds. [And remember---each apnea and each hypopnea lasts at least 10 seconds---so that means on average there's mightly little time left over for normal breathing going on during the night of the NPSG.]

Now it's also important to remember that you don't have a little stop watch in your airway that says, "45 seconds, it's time to collapse now." Rather---over the course of the whole night, sometimes there will be more time between apneas/hypopneas---as in several minutes (or more) of normal breathing between apnea episodes. And sometimes there will be substantially less than 45 seconds---as in several apenas can occur in a row with only two or three recovery breaths between them. That's why the summary graph that shows when the events occurred is also important. For many people, there will be distinctive clustering of events----perhaps when they are in REM or when they are on their back. In my diagnostic data, the worst of the clusters happened just as I was finally really falling asleep more deeply than Stage 1.

Sometimes you also see an "arousal" or an "awakening" index or both on your sleep report. An arousal is any sudden EEG change from a deeper sleep stage to a lighter sleep stage. It may be related to an apnea, a hypopnea, a leg movement, or it may just be labeled "spontaneous" meaning there was nothing that it was clearly related to. An awakening is a sudden EEG change from any sleep stage all the way to WAKE (if I recall correctly). But most people do NOT remember awakenings unless they are awake for at least 5 minutes or so. Awakenings that are shorter than that are usually not consciously remembered. So on your sleep study report, you might also see something like:

Arousals 552    Arousal Index 84.3

So this means you had 552 arousals and awakenings during the 6.55 hours you were asleep. The Arousal/Awakening Index is just the number of arousals and awakenings divided by the total sleep time:

arousal index = (552/6.55) = 84.3

And this number means that you were arousing or awakening (going from a deeper sleep state to a lighter one or all the way to WAKE) about 84 times an hour on average. That means that on average you were "arousing" or "waking up" about every 45 seconds. The arousal index is frequently broken down into a Respiratory Arousal Index, a Periodic Limb Movement Arousal Index, and a Spontaneous Arousal Index: with the obvious definitions:

• Respiratory Arousal Index (RAI) = (total number of arousals associated with SDB)/(TST in hours)
• Periodic Limb Movement Arousal Index = (number of arousals associated with limb movements)/(TST in hours)
• Spontaneous Arousal Index = (number of arousals NOT associated with SDB or PLM)/(TST in hours)

In general, if your SDB events usually trigger an associated arousal, the RAI and the RDI (or AHI) will be pretty close to each other.

Often a minimum blood oxygen saturation level is reported:

Min SpO2 77

Your minimum oxygen saturation dropped to 77%. Anything less than 90% is a concern. There may be more detailed data about your oxygen levels during the night. In addition to the minimum oxygen level, you may see a table that indicates how long your oxygen levels were below 90%. The O2 numbers are usually pretty straightforward to interpret.



REM and NREM indices; Supine indices

First, a quick clarification: REM sleep is rapid-eye-movement sleep, which is when the most vivid dreaming occurs. NREM (non-REM) sleep consists of Stage 1 (transitional), Stage 2 (light), and Stage 3/4 (deep or slow wave) sleep. Note that the number of events in REM vs. the number of events NREM does NOT indicate which kind of sleep has the higher AHI---you need to know exactly how much sleep time occurs in REM and NREM before you can compute the REM AHI and the NREM AHI.

Because most people's apnea tends to be worse in REM sleep than NREM sleep, labs also break the SDB data down into counts and indices for sleep stages. Sometimes the data are broken down as far as separate indices for Stage 1, Stage 2, Stage 3, and REM sleep, but more commonly the non-REM data are all lumped together and just reported as NREM. In the case of the indices, the denominator of the fraction is just the time spent in the appropriate kind of sleep. So we get:

• NREM AHI = (number of apneas and hypopneas occurring in NREM)/(total NREM sleep time)
• REM AHI = (number of apneas and hypopneas occurring in REM)/(total REM sleep time)

And the interpretation of these indices is still in terms of average number of events per hour of the designated sleep type. In other words, if your diagnostic REM AHI = 108, that means that on average for every 60 minutes of REM sleep, the data indicates you would have about 108 apneas and hypopneas.

NREM OAI,  NREM CAI,  NREM HI, NREM RERA I, and NREM RDI are all defined in the same way NREM AHI is defined. Likewise REM OAI, REM CAI, REM HI, REM RERA I, and REM RDI are all defined in the same way REM AHI is defined.

NOTE about REM indices: On many diagnostic sleep studies there is very little time spent in REM sleep. And this can lead to the AHI number being VASTLY greater than the actual number of events that occurred in REM. For folks with severe OSA, it is not uncommon that during the diagnostic sleep test there the person gets as little as 5--10 minutes of REM sleep all night long. So lets suppose that you have 8 OAs and 15 Hs during those 10 minutes of REM sleep. What are the REM indices? Because 10 minutes equals 1/6 hour, the indices are computed as follows:

REM OAI =  8/(1/6) =  8 * (6/1) = 48

REM HI  = 10/(1/6) = 10 * (6/1) = 60

REM AHI = 18/(1/6) = 18 * (6/1) = 108

And the size of these numbers astonishes many people---particularly when they also look at the REM Counts: REM OA = 8, REM H = 10. It's important to realize that REM AHI = 108 means that if you were able to get a total of 60 minutes of REM sleep during the night, then the expected number of apneas and hypopneas occurring in REM would be about 108. Or equivalently that during REM sleep, you're having a breathing problem about every 30 seconds or so since the average time between events is:

(60 minutes) / (108 events per 60 minutes) = 0.555 minutes between events

And many people's apnea is worse when they sleep on their back. Sleeping on you back position is called supine sleep. On most sleep study reports, the event counts and the indices are also broken down into Supine and NonSupine counts and indices. So on your report you may also see:

• Supine AHI = (total number of apneas and hypopneas occurring in supine sleep) / (total supine sleep time)
• NonSupine AHI = (total number of apneas and hypopneas occuring in supine sleep) / (total supine sleep time)

As with REM AHI, if you spend very little time on your back, then the Supine AHI (as a number) may be far greater than than the number of events that actually occurred on your back because you are dividing the number of events by a small fraction, which turns into a multiply the number of events by a number bigger than one.


Understanding the data from the titration study
(THIS SECTION IS INCOMPLETE)
All the same kinds of data recorded in the diagnostic sleep study are also recorded in the titration study---the one where the tech puts a mask on your nose and after you go to sleep, the tech remotely adjusts the pressure in response to what he/she is seeing in your data.

There is an algorithm that is used to determine when to increase the pressure and, under certain circumstances, when to lower the pressure. (NOTE TO SELF---find the appropriate link to a description of this algorithm.) Because the test starts off at low pressure---pressure that is NOT high enough to be therapeutic for most of us---the overall AHI, OAI, CAI, HI numbers may not be as low as that magic "AHI < 5.0" that's the target for all xPAP therapy. The more important numbers on the titration study are the AHI numbers at each pressure level that was tried during the night AND whether that pressure was enough to reduce the SUPINE REM AHI to (much) less than 5 for something like at least 15 minutes.

Of course, many of us don't get 15 minutes of supine REM sleep during the night. And in that case, the doc is looking at was the final pressure (or the highest tolerated pressure) enough to eliminate or come close to eliminating all events. And was there any REM sleep or SUPINE sleep at that pressure? And was there REM or SUPINE sleep at a lower pressure?

For example, in my first titration study, my over all RDI was 5.4 for the entire night---because of the events that occurred when I was at 4 and 5 cm of pressure. In both my REM cycles, 6 to 7cm of pressure controlled my apnea almost completely---only 1 hypopnea with arousal occurred during all 46.5 minutes of REM sleep which occurred in two 20+ minute long REM cycles. But about a half hour after the second REM cycle, I flipped onto my back for the last half hour of the sleep study. And there were just enough events on my back that the tech increased the pressure to 9cm, at which point, I had no more hypopneas or apneas for the remaining 20 minutes or so of the sleep test. So at 9cm of pressure, I had a SUPINE AHI = 0.0 for about 20 minutes. Hence, the doc prescribed a pressure of 9cm (which turned out to be a bit too much, but that's another story).

And the important thing to realize when interpreting all the index numbers on the titration study is that the AHI at 6 cm of pressure is simply computed as:

AHI at 6 cm = (total number of events occurring at 6 cm) / (total sleep time at 6 cm) 

And again, if you slept less than an hour at a given pressure setting, then the AHI number will be larger than the number of events that occurred----because in this context the AHI should be looked at as the average number of events expected in an hour of sleep at the given pressure.