Structural nasal damage - sharing early AutoSet data, question about airbreak firmware
Hey all. Long post but wanted to share my case properly and get input on whether pursuing the Airbreak firmware flash for VAuto mode is worth it given my data so far.
Background and suspected cause
I'm a 30 year old male based in the UK. My suspected UARS origin is a specific traumatic event - at around 14-15 I took a hockey stick directly to the face during PE, bad enough to go to A&E. I was offered surgical correction of the septum on the spot and declined because I was a scared teenager. That decision has probably cost me 15+ years of decent sleep.
Around the same time, I had four teeth removed to create space for wisdom teeth, which may have contributed to reduced posterior airway space on top of the structural nasal damage. Interestingly my palate doesn't appear narrowed despite this - my tongue can sit fully on the roof of my mouth with the tip only lightly touching the back of my upper front teeth, suggesting palatal expansion is likely not indicated.
The resulting picture is a confirmed deviated septum with bilateral nasal valve collapse (strongly positive Cottle manoeuvre bilaterally, right worse than left), mild mandibular retrognathia, and chronic mouth breathing secondary to nasal obstruction. The subjective sensation when replicating my snoring while awake is a vacuum pulling soft tissue into the airway - classic Starling resistor dynamic.
I trialled a MAD which resolved snoring completely but caused jaw discomfort and tooth mobility so discontinued. Home oximetry normal - no OSA diagnosis - consistent with UARS where RERAs resolve before oxygen drops. No access to in-lab PSG or RERA scoring so everything below is inferred from OSCAR data only.
Current self-treatment
ResMed AirSense 10 AutoSet, AirFit N30 nasal cradle, EPR Full Time Level 3, pressure range 4–15 cmH₂O. 14 AutoSet nights of OSCAR data.
OSCAR data — AutoSet nights
| Night | Date | Duration | AHI | CA | OA | Med P | FL 95th | FL 99.5th |
|---|---|---|---|---|---|---|---|---|
| 1 | 6 May | 4h 17m | 0.70 | 3 | 0 | 4.22 | 0.12 | 0.12 |
| 2 | 7 May | 8h 33m | 1.75 | 12 | 0 | 4.62 | 0.02 | 0.13 |
| 3 | 8 May | 4h 34m | 0.00 | 0 | 0 | 4.60 | 0.03 | 0.19 |
| 4 | 9 May | 5h 03m | 0.59 | 3 | 0 | 4.28 | 0.00 | 0.12 |
| 5 | 10 May | 8h 40m | 0.58 | 2 | 2 | 4.54 | 0.02 | 0.13 |
| 6 | 11 May | 7h 47m | 1.16 | 8 | 1 | 4.36 | 0.02 | 0.14 |
| 7 | 12 May | 7h 05m | 2.26 | 11 | 2 | 5.00 | 0.02 | 0.19 |
| 8 | 13 May | 7h 37m | 0.66 | 4 | 0 | 4.36 | 0.01 | 0.16 |
| 9 | 14 May | 4h 01m | 0.75 | 2 | 0 | 4.54 | 0.02 | 0.12 |
| 10 | 15 May | 5h 45m | 0.52 | 3 | 0 | 4.34 | 0.02 | 0.12 |
| 11 | 16 May | 6h 01m | 0.50 | 3 | 0 | 4.48 | 0.03 | 0.17 |
| 12 | 17 May | 7h 05m | 0.57 | 3 | 0 | 4.88 | 0.03 | 0.23 |
| 13 | 18 May | 7h 43m | 1.17 | 7 | 1 | 4.38 | 0.01 | 0.18 |
| 14 | 19 May | 8h 28m | 0.59 | 4 | 1 | 4.78 | 0.02 | 0.17 |
Key trends
- FL 95th percentile essentially eliminated - 0.00–0.03 across 13 of 14 nights from night 2 onward. The primary UARS therapy target is being consistently met.
- FL 99.5th percentile stubbornly persistent - consistently 0.12–0.23 across all nights regardless of how clean the rest of the session is. Almost certainly concentrated in late REM where the machine pushes to 7–9 cmH₂O. I'm waking consistently around 6am - one cycle short of ideal - which I suspect is this REM-phase challenge cutting the final cycle.
- AHI below 0.75 on undisturbed nights - elevated nights (1.16–2.26) all correlate with higher CA counts and median pressure, pointing to congestion as the driver.
- Median pressure 4.22–4.88 - airway needs very little support for most of the night. Over-pressure on the Elite (fixed 7.0) was likely driving much of my exhale anxiety and CA sensitivity early on.
- CA events settled at 2–4 on good nights - adaptation complete.
Airbreak question
Firmware SX567-0401, SD card flashable, no warranty on refurb unit. Rationale: the persistent 99.5th percentile FL and late-REM pressure peaks suggest a true bilevel PS gap might convert flow-limited breaths in that window without requiring IPAP to climb as high - potentially reducing the arousal cutting my final REM cycle short.
Aware of TECSA risk given CA sensitivity - planning to start at PS 2 and monitor closely. Firmware is reversible so happy to revert to AutoSet if CAs climb and don't settle.
Questions: is this worth doing given this data? What PS gap would you recommend given the CA sensitivity? And has anyone seen the 99.5th percentile FL pattern resolve with bilevel where APAP couldn't get it?
Happy to share OSCAR screenshots if helpful, and thanks in advance for any help or guidance - your time is really appreciated in helping combat this insidious problem.
As per a comment suggestion, here is a cut of my flow rate from the most recent night.
