Sleep Surgery Outcomes
Dr. Stanley Liu publishes his results. This page gathers the outcomes reported in his peer-reviewed studies, procedure by procedure, with a link to every paper.
How to read the numbers: the apnea-hypopnea index (AHI) counts pauses and shallow breaths per hour of sleep. Under 5 is normal and 30 or more is severe. Each figure below is a group average before and after surgery. The AHI is also an incomplete measure, and the next section explains why it is not the whole story.
Why the AHI is not the whole story
The AHI is the number insurers and most studies use, so it appears throughout this page. It is also an incomplete measure, and Dr. Liu has published on its limits. The AHI counts how often breathing stops or turns shallow. It does not show how hard a person works to breathe between those events, how long the events last, how deep the oxygen drops go, or how broken the night is. Two people with the same AHI can feel completely different.
It underestimates women
- 13.7 vs 22.1Median AHI of women and men seeking surgery
- 33% lowerAHI in women, adjusted for age, BMI and symptoms
- 2.2 timesAs likely to be taking anxiety or depression medication
Thresholds developed around middle-aged men leave many women below the AHI of 15 that treatments such as Inspire require.
It misses the effort to breathe
Upper airway resistance syndrome, described by Dr. Liu's mentor Dr. Christian Guilleminault, breaks up sleep and drains energy while the AHI stays normal or close to it. The airway narrows and the body works harder to keep it open, but breathing never stops long enough to be counted. It is often seen in younger, slimmer adults, and many of them are women.
Measuring what patients feel
- 15 questionsSLEAP, a quality of life measure built for sleep apnea
- 176 and 22Patients and healthy controls in its validation
Dr. Liu co-authored SLEAP so that results can be judged the way patients live them, alongside the sleep study.
So a strong AHI result is not the same as a patient who feels well. Some patients reach a low AHI and still sleep poorly or wake tired, while others feel much better after a smaller change in the number. The results below are reported in the AHI because that is how the field compares treatments. At the consultation, Dr. Liu also asks how you sleep, how you breathe through your nose and how you feel during the day. The International Consensus Statement on Obstructive Sleep Apnea, which he co-authored, names better ways of measuring the disease as one of the field's open needs.
The approach behind the results
These results come from the Stanford approach that Dr. Liu developed into the Updated Stanford (Riley-Powell) Sleep Surgery Algorithm, building on the protocol created by his mentors Dr. Robert Riley and Dr. Nelson Powell. Surgery is matched to where and how the airway collapses, seen on drug-induced sleep endoscopy, rather than following a fixed sequence. The framework is set out in his chapter with both mentors, Algorithm for Multilevel Treatment: The Riley, Powell, and Liu Stanford Experience (Elsevier, 2020).
Maxillomandibular advancement (MMA)
Both jaws are moved forward and rotated together so the airway holds open during sleep. Across Dr. Liu's studies, the AHI fell from the severe range to the mild range, with gains in sleepiness, sleep quality and nasal breathing.
What changes in the airway
- 59.8 to 9.3AHI, events per hour
- 45.0 to 5.7Oxygen desaturation index
Sleep endoscopy showed the largest change at the side walls of the throat, and their stability marked surgical success.
Airflow after surgery
- 53.6 to 9.5AHI, events per hour
- 38.7 to 8.1Oxygen desaturation index
Airflow modeling linked the improvement to slower airflow behind the soft palate.
Works for both kinds of bite
- 43.4 to 9.6AHI, class 2 bite (14 patients)
- 37.2 to 11.8AHI, class 3 bite (14 patients)
- 8.9 to 3.9Sleepiness score, class 2
- 10.2 to 4.2Sleepiness score, class 3
Deeper, steadier sleep
- 64 to 22 minTime awake after falling asleep
- 178 to 97 minTime to first dream (REM) sleep
After surgery, sleep stages resembled a population control group.
Protecting the nose
- 6.5%Needed later corrective nasal surgery, compared with 18.7% previously reported
- 34.7Average AHI reduction, events per hour
What patients reported
- 44.4 to 19.0Nasal obstruction score
- 14.0 to 5.3Nasal appearance concern score
- 9.4 to 3.3Sleepiness score
Distraction osteogenesis maxillary expansion (DOME)
DOME widens a narrow upper jaw and the floor of the nose in adults. Dr. Liu co-pioneered DOME at Stanford, working with Dr. Christian Guilleminault and Dr. Audrey Yoon, and reported the first 20 patients in 2017.
The largest DOME series
- 17.7 to 8.2AHI, events per hour
- 10.9 to 3.3Nasal obstruction score
- 10.5 to 6.7Sleepiness score
- 14.4% to 22.7%Share of sleep spent in REM
No significant adverse effects were identified.
Where the expansion matters
- 17.1 to 7.0AHI, events per hour
- 11.4 to 3.6Nasal obstruction score
- 12.0 to 7.1Sleepiness score
Widening at the front of the nasal floor predicted the drop in AHI.
How DOME eases breathing
- 17.8 to 7.8AHI, events per hour
- 88.2% to 90.9%Lowest oxygen level during sleep
- 15.6 to 7.4 m/sAirflow speed through the nose
Slower, calmer airflow through the nose lowered the suction on the throat below it.
The first report
The technique, and the safety and efficacy of the first 20 Stanford patients: expanders secured with mini-implants along the palate, with minimal bone cuts, designed for adults with a narrow nasal floor and a high-arched palate.
Making hypoglossal nerve stimulation (Inspire) possible
Some patients are first told they are not candidates for Inspire, because their apnea is too severe or their soft palate collapses in a full circle. Dr. Liu's work shows surgery can change that.
Converting the collapse pattern
- 12 of 12No longer had circular collapse at the soft palate
- 54.0 to 33.1AHI after palate surgery alone
A tissue-preserving palatopharyngoplasty removed the barrier to stimulation.
Surgery first, then stimulation
- 37.6Average AHI reduction: initially ineligible, multilevel surgery then stimulation (18 patients)
- 31.5Average AHI reduction: stimulation alone (17 patients)
Adjusted for age, BMI and sex, the surgery-first group improved by about 18 more events per hour.
Putting these results in context
Most of these studies look back at patients treated at an academic referral center, and each group was selected for the procedure that fit its anatomy. The figures are averages. A strong average does not mean every patient was satisfied: some improve less than the group, and some need more than one step of treatment. That is why success is judged by sleepiness, fatigue, nasal breathing and quality of life as well as by the AHI.
The studies come from Stanford Sleep Surgery, where Dr. Liu was associate professor and directed the sleep surgery fellowship. He continues the same approach at the NSU Health Sleep and Breathe Wellness Center in Davie, Florida, and operates at Memorial Regional Hospital, Joe DiMaggio Children's Hospital and Memorial Hospital West.
Outcomes | Frequently asked questions
Understanding the numbers
What counts as a good result after sleep apnea surgery?
A lower AHI is one measure, and an incomplete one. Dr. Liu also looks at oxygen levels, daytime sleepiness, fatigue, nasal breathing, quality of life and the quality of sleep itself, because patients feel the difference in those first. In his studies these usually improved along with the AHI, though not for every patient.
Can I feel much better even if my AHI does not reach normal?
Yes. In several of the studies on this page, the average AHI after surgery was still above 5, the cutoff for normal, yet patients felt clearly better. After jaw advancement in one study, the average AHI fell from 43.4 to 9.6 and the sleepiness score fell by more than half, from 8.9 to 3.9. After DOME, nasal breathing scores dropped from 10.9 to 3.3 while the average AHI went from 17.7 to 8.2. How you breathe, sleep and function during the day matters as much as where the number lands.
Why might I still feel tired when my AHI has improved?
The AHI does not count the extra effort to breathe, short awakenings or a blocked nose, so some of what disrupts sleep can remain after the number improves. Tiredness can also come from causes outside the airway, such as insomnia, other sleep disorders, medications or a short sleep schedule. Dr. Liu looks for what is left, and some patients need another step of treatment or care alongside their sleep medicine physician.
Is the AHI different for women?
Often, yes. In a study of 408 patients seeking surgery, women had a 33% lower AHI than men with similar age, BMI and symptoms. The same cutoff can therefore make a woman's sleep apnea look milder than it is and leave her below the thresholds some treatments require. Read more about the AHI blind spot.
What if I have symptoms but my AHI is normal or mild?
Symptoms still count. Upper airway resistance syndrome can fragment sleep while the AHI stays normal or close to it. The evaluation looks at the details of the sleep study, the nose, the jaws and the airway, so treatment can be considered when the anatomy explains the symptoms. The sleep apnea page explains how the airway is assessed.
Besides the AHI, how is success measured?
With questionnaires filled in before and after surgery, such as the Epworth Sleepiness Scale for daytime sleepiness, the NOSE and SCHNOS scales for nasal breathing, and SLEAP for quality of life with sleep apnea. Oxygen levels, time awake during the night and dream (REM) sleep come from the sleep study. The patient questionnaires page has the Epworth and NOSE forms.
Do insurers use the AHI?
Yes. Insurers and clinicians rely on the AHI to grade severity and decide which procedures are covered, and some treatments have a minimum, such as an AHI of 15 for hypoglossal nerve stimulation. This is one reason Dr. Liu has published on how a single cutoff can underestimate disease in women.
Your own result
Do these results apply to me?
They show what is possible for patients with a similar anatomy. Your own plan starts with your sleep study, an examination of your face and airway and, when it helps, drug-induced sleep endoscopy. Dr. Liu reviews what result is realistic for you before any surgery.
Will I need more than one surgery?
Some patients do, and the Updated Stanford (Riley-Powell) Sleep Surgery Algorithm is built in steps for that reason. In one study, palate surgery made all 12 patients eligible for hypoglossal nerve stimulation after the way their palate collapsed had first ruled it out. In another, 6.5% of 122 patients needed corrective nasal surgery after jaw advancement. Dr. Liu discusses the likely sequence before the first step.
How will I know how well my surgery worked?
Once healing is complete, a repeat sleep study measures the AHI and oxygen levels again, and the same questionnaires are repeated so the change in sleepiness, nasal breathing and quality of life can be compared with your starting point. Both are reviewed with you at follow up.
Where can I read the full studies?
Each card above links to the paper on PubMed. The complete list of Dr. Liu's more than 120 peer-reviewed papers and chapters is on the Publications page.