Sleep & CPAP Report — 2026-08-29

🚩 Flags

• 3-night streak of elevated obstructive events — trigger hit. OA counts Aug 27 → 28 → 29 were 14, 12, 15 (all ≥12). This is the streak yesterday's report explicitly flagged to watch for. 3-night averages: AHI 2.89 (+59% vs 30d 1.82), OA 13.67 (+63% vs 30d 8.37). Prior to Aug 27, only two nights in the trailing month (Aug 20, Aug 21) hit ≥12 OA — now four of the last ten do. This is a genuine pattern, not a blip.
• Tonight is the worst night of the trailing 10 by AHI. AHI 3.07 (up from 2.43 last night; +69% vs 30d 1.82). OA 15 ties Aug 21 as the highest OA count in the last 10 nights. Hypopneas jumped to 6 (vs 1 last night, 30d avg 2.8) — highest H count in the last 10 nights.
• Pressure delivery ramped back up appropriately tonight. Pressure_Avg 8.23 (vs 6.08 last night; back near 30d 7.44). Pressure_90th 10.96 (vs 8.90; above 30d 9.95). Unlike last night, the AirSense algorithm did respond to the elevated event load. So yesterday's "pressure didn't ramp" concern doesn't repeat — but the events happened anyway despite normal pressure delivery.
• Flow limitation up sharply. Time with flow-limit >0.10 was 35.6 min tonight vs 8.1 last night (~4×). Time >0.20 was 4.7 min. FL_P90 rose to 0.10 (vs 0.04). This is consistent with airway narrowing / positional/upper-airway issue driving the OA spike, not with mask fit.
• Deep sleep dipped. 11.5% tonight vs 7d avg 12.83, 30d avg 12.89 (−1.4 pp vs 30d). It's not extreme (Aug 25 was 6.6%), but combined with the events, worth logging. Awake time stayed low (6.0 min) and efficiency was 98.5%, so subjective sleep quality is probably still fine.
• Central apneas resolved. CA=0 tonight after 3, 3, 1 on the three prior nights. Good — the CA cluster looks like a short-lived thing.

Summary Table — Night of 2026-08-29

Metric	Value
Sleep start	23:28:56 CDT
Time in bed	401.9 min (6h 42m)
Total sleep	395.9 min (6h 36m)
Sleep efficiency	98.5%
Awake time	6.0 min
Core (N1+N2) sleep	250.9 min / 63.4%
Deep (N3) sleep	45.7 min / 11.5%
REM	99.4 min / 25.1%
CPAP usage hours	6.85
Session count	1
AHI	3.07
RDI	3.36
Obstructive apneas (OA)	15
Hypopneas (H)	6
Central apneas (CA)	0
Unclassified apneas (UA)	0
RERA	2
Pressure avg	8.23 cmH₂O
Pressure min	4.00 cmH₂O
Pressure 90th	10.96 cmH₂O
Pressure max	12.96 cmH₂O
Leak avg	0.93 L/min
Leak 90th	3.6 L/min
Leak max	19.2 L/min
Flow-limit p95	0.13
Flow-limit min >0.10	35.6
Flow-limit min >0.20	4.7

Trailing Averages Comparison

Metric	Tonight	7-day avg	Δ vs 7d	30-day avg	Δ vs 30d
AHI	3.07	1.81	+70%	1.82	+69%
OA count	15	8.57	+75%	8.37	+79%
H count	6	2.14	+180%	2.80	+114%
CA count	0	1.29	−100%	1.07	−100%
Leak_Avg (L/min)	0.93	0.66	+41%	0.77	+21%
Leak_90th (L/min)	3.6	1.37	+163%	2.28	+58%
Pressure_Avg	8.23	7.14	+15%	7.44	+11%
Pressure_90th	10.96	9.62	+14%	9.95	+10%
Usage hours	6.85	6.63	+3%	6.73	+2%
Deep sleep %	11.5%	12.8%	−1.3 pp	12.9%	−1.4 pp
REM %	25.1%	25.8%	−0.7 pp	23.7%	+1.4 pp
Awake min	6.0	9.2	−34%	16.3	−63%
Total sleep (min)	395.9	383.1	+3%	385.8	+3%

Week-over-Week Trend

Metric	Recent 7d (Aug 23–29)	Prior 7d (Aug 16–22)	Direction
AHI	2.03	1.85	↑ +9.3%
OA count	9.57	8.86	↑ +8.1%
CA count	1.29	0.43	↑ +200%
Leak_Avg	0.69	0.97	↓ −29.3%
Leak_90th	1.89	2.23	↓ −15.4%
Pressure_Avg	7.17	7.81	↓ −8.2%
Pressure_90th	9.65	10.56	↓ −8.6%
Deep sleep %	12.50	12.85	↓ −2.7%
REM %	26.88	21.69	↑ +23.9%
Awake min	7.45	9.68	↓ −23.0%

The week-over-week picture is mixed: leak, awake time, and REM all trending favorable; AHI and OA counts drifting up; pressure delivery running below the prior week. The event load is now firmly above the prior week's baseline.

Recent 10-Night Detail

Date	AHI	OA	H	CA	Leak_Avg	Leak_90th	P_Avg	Hours
2026-08-29	3.07	15	6	0	0.93	3.6	8.23	6.85
2026-08-28	2.43	12	1	3	0.28	0.0	6.08	6.58
2026-08-27	3.17	14	—	3	0.83	3.6	7.59	6.63
2026-08-26	1.33	7	—	1	0.72	0.0	7.09	6.77
2026-08-25	1.61	7	—	2	0.73	2.4	7.05	6.85
2026-08-24	1.81	9	—	0	0.37	0.0	6.63	6.63
2026-08-23	0.77	3	—	0	0.20	0.0	6.72	6.87
2026-08-22	1.54	8	—	0	0.63	0.6	6.94	6.86
2026-08-21	2.82	15	—	1	1.03	3.6	6.98	6.85
2026-08-20	2.55	14	—	0	1.03	3.6	7.15	6.71

Four of the last ten nights (Aug 20, 21, 27, 28, 29) have hit ≥12 OA — that's a higher frequency than the trailing 30d rate would predict.

Correlation Analysis (57 matched nights)

Pair	the user r
AHI ↔ Deep sleep %	+0.15 (weak positive)
AHI ↔ Awake minutes	+0.02 (none)
Leak_Avg ↔ Deep sleep %	−0.16 (weak negative)
Leak_Avg ↔ Awake minutes	−0.06 (none)
Leak_90th ↔ Deep sleep %	−0.19 (weak negative)
Leak_90th ↔ Awake minutes	−0.04 (none)

High- vs low-leak bucket comparison (split at Leak_90th median = 2.4 L/min): - High-leak nights (n=29): mean deep sleep 14.2%, awake 15.4 min - Low-leak nights (n=28): mean deep sleep 15.3%, awake 16.6 min

Same directional pattern as prior reports: higher leak weakly associates with lower deep sleep, but the effect is small. Tonight combined moderate leak (Leak_90th 3.6 — above the median) with elevated events and below-average deep sleep — consistent with the weak but persistent relationship.

Interpretation

The pattern the last two reports have been tracking is now real. Three consecutive nights (Aug 27–29) with OA ≥12, AHI ≥2.4, and elevated hypopneas or centrals mixed in. Tonight's the highest AHI of the streak, driven by both more OAs (15) and a hypopnea spike (6 vs baseline 2.8). Flow limitation was also markedly higher than yesterday — 35.6 min above the 0.10 threshold, ~4× last night.

Two things narrow the interpretation:

1. It's not mask/leak. Leak was moderate tonight (Leak_90th 3.6 vs 30d 2.28) but the week-average leak has trended down, and last night's near-perfect leak also had elevated OA. The event pattern doesn't track mask fit.
2. The machine did respond tonight (Pressure_90th 10.96, above baseline). So the answer isn't "pressure never ramped." Even with the algorithm engaged, events still ran high — suggesting the driver is upstream of pressure delivery (positional, congestion, alcohol, weight, or an airway change).

What to check: Any of the following changed in the last week — sleep position (more supine?), evening alcohol, nasal congestion / allergies, recent weight change, or CPAP settings (mask type, pressure ceiling, EPR). The three-night pattern with normal-to-good mask fit and normal pressure response points toward something behavioral or physiological rather than device-side. Given the streak is now established, worth mentioning at the next provider check-in if it continues past this week.

Subjectively: you likely still feel rested — awake time is only 6 min, efficiency 98.5%, REM 25%. The events aren't fragmenting the night; they're clustering. That's why the pattern is easy to miss without the data.

Caveats

• SpO2/pulse columns remain empty (device not recording those channels).
• Sleep-stage percentages are Apple Watch estimates (accelerometer + HR), not lab EEG.
• Hypopnea counts for Aug 23–27 are not present in the per-day CSVs used (dashes in the recent-nights table), so H trend before Aug 28 is based on 30-day rolling only.
• Single CPAP session — no mid-night mask-off artifacts to filter.
• The "3-night streak" trigger is empirical (OA ≥12 for 3 consecutive nights vs personal baseline of ~8) — not a clinical threshold. AHI 3.07 remains well within the "well-controlled" clinical range (<5); the flag is about deviation from personal baseline, not absolute severity.