Analyzing the sleep study

You may be wondering why there is an article in this issue on how to analyze a sleep study. If you do not already know, a sleep study is required in order to diagnose someone's sleep breathing disorder. In my office, I do not treat any patient without first obtaining a copy of their sleep study. A study, or polysomnogram (psg), has many uses, not the least of which is the requirement by the patient's medical insurance before they ever reimburse you for treatment.

Kent Smith, DDS

You may be wondering why there is an article in this issue on how to analyze a sleep study. If you do not already know, a sleep study is required in order to diagnose someone's sleep breathing disorder. In my office, I do not treat any patient without first obtaining a copy of their sleep study. A study, or polysomnogram (psg), has many uses, not the least of which is the requirement by the patient's medical insurance before they ever reimburse you for treatment.

This report also gives us very valuable information that relates to whether an oral appliance might (or might not) be an effective treatment for your patient. Additionally, if you can explain the report to your patients, they will immediately view you as an expert, as this is rarely done in the physician's office. Finally, it serves as a baseline so that you can compare it to a future study to determine how effective your treatment has been.

Of all the questions I get from dentists getting involved in dental sleep medicine, questions about how to "read sleep studies" is among the most common. There are literally hundreds of sleep study formats, so there is no way for me to be detailed about any specific study you may be viewing. In fact, some study reports will be one page in length, while others could be as many as 15 to 20. However, should you like to have me help with analyzing a study, feel free to email me at info@sleepdallas.com and I'll be happy to assist you. I also have developed a standard analysis sheet in my own office to distill the most important data from psgs, so if you would like me to send you a copy of what I use, please let me know.

I am showing what an attended psg report might look like (below), but home sleep tests (HST), depending on complexity, will have much of this data as well. I don't know the future of the attended psg, as the HST market seems to be replacing these over time, but even after the last sleep center closes, we will continue to see copies of these psg reports come in to the office to be used as the baseline (and necessary diagnosis) for treatment.

Before looking at the data, look at the patient history that was obtained by the sleep physician and is usually contained within the report. If there is mention of any comorbidities such as insomnia, these can be used along with a mild diagnosis of obstructive sleep apnea when filing medical insurance.

When you look at sleep latency, keep in mind that this was an attended study, where your patient slept in a foreign environment with someone else's pillow while a pair of eyes was trained on him/her all night long.

There could have been strange noises as well as variations in temperature that were uncommon in the patient's bedroom. All of these can delay sleep onset, so I never put much credence in the sleep latency data in these studies.

Try not to get too concerned about a few central apnea episodes that you typically see in these reports. As patients transition between wake and sleep, most will have a few of these centrally mediated events. If the sleep report does not indicate these are a concern, you can assume these are insignificant.

One other acronym that bears explanation is the WASO.This stands for "Wake After Sleep Onset," which indicates all wake time from the moment of sleep onset until "Lights On." This is then used to calculate the sleep efficiency, which should ideally be 90%, although sleeping in a foreign environment can certainly put a damper on one's ability to have a nice, consolidated night of sleep.

The Holy Grail in dental sleep medicine would be the knowledge of whether our oral appliances will effectively treat any specific patient. Let's look at a few items that might help us in this area.

1. The PAP titration study: The final pressure prescribed can give us an idea if a MAD might be successful. Generally, the higher the pressure, the more difficult it will be to open the airway with any technique, including your appliance.

2. If the patient has supine-dependent apnea, they are more likely to have success with MAD treatment. If their AHI is twice as bad in the supine position as it is in the nonsupine position, this is considered to be positional, or supine-dependent apnea.

3. Look at the number of apneic events compared to the number of hypopneic events. If heavily weighted in hypopneas, your appliance is more likely to be successful.

4. If their oxygen does not appear to be as severe as the AHI might indicate, your appliance might be more successful.

5. An elevated BMI can lead to less success with mandibular advancement. Assuming you do not have weight scales in your office, look at the BMI in the report, and ask your patient if they have gained or lost weight since the study was done. Of course, I am not saying that we should not treat someone who is even morbidly obese. However, I am not as confident and I will be sure to temper my expectations when discussing the likelihood of success with my patient.

6. Finally, see if the report lists the average event time. If these events are closer to 10 seconds, they are not as severe as an average of more than 20 seconds. However, when calculating the AHI, these events are treated the same. This is one of the many reasons I feel the AHI, although presently considered the standard for diagnosis, will soon be thrown to the curb, as this formula for determining the severity of the disease has not caught up with current science.

I can't overemphasize the usefulness of understanding the polysomnogram report, and if you can get a copy prior to the patient's visit, you can effectively paint yourself as the doctor they want treating them. That's a good thing, yes?


The first thing you need to know about any sleep study is that it is a snapshot - a one-time look at the patient's sleep, which may or may not be representative of a typical night. The sleep architecture, which includes the amount of REM sleep and deep sleep, seems to have the most variability, as there are many reasons for an increase or decrease in the amount of these stages a patient might have on any given night.

As an example, look at the"lights on" time shown in the report, and ask your patients what time they usually wake when sleeping at home.

When looking at deep or slow wave sleep(N3), the amount a patient gets will decrease with age, but more so as a result of sleep disordered breathing.

Most REM sleep occurs during the latter third of sleep, so if your patient was awakened considerably earlier to conclude the study, this could have an impact on the amount of REM sleep received during the sleep study. Ideally, adults should get 20% to 25% each night, regardless of age.

Finally, an attended sleep study has variability between different polysomnographers. For example, one may interpret a hypopnea differently than another. I have seen drastic differences between different polysomnographers over the years, and there is research illustrating this as well.

VII. Polysomnographic Data:

Lights Out: 11:11:49 PM Lights On: 05:13:21 AM

Sleep architecture Epworth: 13
Total study time: 361.5 minutes Sleep efficiency 62.6%
Total sleep time: 226.5 minutes WASO 119.5 minutes
Sleep latency: 35.5 minutes REM latency 93.0 minutes
Stage % of total sleep time
SPO2intervals% of sleep time
N1 26.9% > 89% 74.1
N2 60.0% 80-89% 25.1
N3 6.1% 70-79% .5
REM 7.0% Nadir 78%

Respiratory events
Obstructive apneas
18 REM index 32.7
Mixed apneas 4 Non-REM index 16.9
Central apneas 12 Supine index 33.4
Hypopneas 40 Nonsupine index
7.8
Total events 74 Apnea/hypopnea index19.6
RDI 27.8

SPO2/heart rate statistics
Mean SpO2 95.4% SpO2 minimum 72%
Mean heart rate 65.9 HR range 54.0 to 94.0

PLMS events
# of PLMS 24 PLMS Index 6.4

Arousal statistics
PLMS arousals 11 PLMS arousal index 2.9
Respiratory arousals 31 Respiratory arousal index 8.2
Snore arousals 14 Snore arousal index 3.7
Spontaneous arousals 83 Spontaneous arousal index 22.0
Total arousals 139
Arousal index36.8

Without insomnia, hypertension, excessive daytime sleepiness, etc., a mild diagnosis will not allow your treatment to be covered by their insurance.

Looking at someone's health history can shed some light on what you might see in a sleep study. As an example, various medications can alter sleep latency, AHI,SpO2, and sleep architecture. For this reason, I strongly recommend having this information at your fingertips.

When looking at the AHI(apnea/hypopnea index)and or RDI(respiratory index), different sleep centers report this data differently. Some will only show an AHI; some will only show an RDI; and some will show both. Insurance companies are looking for the AHI. However, if only an RDI is shown, this typically represents the AHI.

The arousal indexshould hold your interest, as this can create excessive fatigue, even after their sleep apnea is treated. Alerting your patient to this can help soften their expectations so they don't get frustrated with residual sleepiness after your appliance has reached maximum clinical effectiveness.


Dr. Kent Smith is a diplomate of the American Board of Dental Sleep Medicine. He is the creator of the Best Sleep Hygiene app, has hosted a weekly radio show on sleep disorders since 2012, and teaches sleep seminars in the Dallas area. He may be contacted at info@SleepDallas.com.

More in Science & Tech