Sleep & CPAP Report — 2026-10-07

🚩 Flags

• 🟢 Event metrics continued the recovery trend — AHI, hypopneas, and RERA all well below the user's baseline. AHI 1.24 (rank 5/26; 30d z = -0.96, -40.8% vs 30d mean of 2.09). OA count 7 (vs yesterday's 8, -12.5%; rank 11/26, z = -0.59); H count 1 (vs 4, -75.0%; rank 7/26, z = -0.90 — H count 1 is tied for 2nd-lowest in the recent 26-night window — only 10-01 had H=0); CA 0 (unchanged); RERA 0 (vs 1). Tonight is the second consecutive clean night after the 10-05 AHI 5.20 breakthrough — strong confirmation 10-05 was a one-off event cluster, not a trend.

• 🟢 Deep sleep returned to the upper end of the user's range — strongest reading since 2026-10-02. Deep 18.21% / 68.2 min (rank 22/26, i.e. 5th-highest in the recent 26-night window; 30d z = +1.25, +34.8% vs 30d mean of 13.51%). REM 23.43% / 87.8 min (z = -0.19, -3.0% vs 30d — in-line). Awake 6.02 min (rank 9/26; z = -0.59, -37.2% vs 30d mean of 9.58). Sleep efficiency 98.42%. Deep + REM combined is 41.6% of total sleep — one of the better architecture nights in the last two weeks, with no awake fragmentation.

• 🟡 Leak_Max 34.8 L/min ties the recent high (also 34.8 on 09-28) and is the fourth of the last six nights (10-02, 10-04, 10-06, 10-07) with a brief Leak_Max transient ≥25 L/min. Leak_Max rank 25/26 (z = +1.58, +64.8% vs 30d mean of 21.12). Leak_Avg 1.50 L/min is also elevated (z = +0.59, +41.2% vs 30d mean of 1.06), though 90th-percentile leak 3.6 (z = +0.06) stayed modest, so these are brief transients, not sustained drift. The pattern was flagged as 'worth noting' yesterday — tonight escalates it: four consecutive outlier nights is no longer coincidence. Mask fit / cushion wear / sleep position are worth checking before this starts eating into sleep stages (so far no fragmentation impact — awake stayed low).

• 🟡 Flow-limitation metrics are elevated across the board — the one channel not improving tonight. FL_P95 0.12 (rank 24/26; z = +1.36, +27.2% vs 30d mean of 0.09); Min_Above_010 23.3 min (rank 24/26; z = +1.46, +40.7% vs 30d mean of 16.56); Min_Above_020 4.6 min (rank 20/26; z = +0.38, +17.2% vs 30d). Frac_NonZero 34.94% (up from yesterday's 22.08%). FL_P95 0.12 is the highest since 2026-10-04 (0.13). None of this is clinically concerning (clinical cutoff for sustained FL is ~0.3) and these metrics are decoupled from AHI (which was clean), but it's possible the leak transients and the FL elevation are related — a worse mask seal can allow brief flow restriction that the device still catches below the hypopnea threshold.

• 🟢 Pressure stayed in the user's recent operating range — slight uptick vs yesterday but within normal variation. Pressure_Avg 7.69 cm H₂O (vs yesterday's 7.19; 30d z = +0.29, +1.7% vs 30d mean of 7.56); Pressure_90 10.72 (rank 24/26; z = +0.87, +5.8% vs 30d mean of 10.14); Pressure_Max 12.62 (vs yesterday's 12.14). The auto-titrator did not spike pressure despite the Leak_Max transient and FL elevation — because event counts were low. The system behaved correctly tonight.

• 🟡 Trailing 7d AHI avg is 2.06 (+13.9% WoW). The 10-05 outlier (AHI 5.20) is still inside the recent 7d window, inflating the mean; drop it and the 7d avg is ~1.43. The ~5-week improving event-count trend remains intact in character — tonight (10-07) is now the second consecutive clean follow-up, so the elevated 7d avg is almost entirely a single-night artifact, not a trend shift.

Summary Table — Night of 2026-10-07

Metric	Value
Sleep start	2026-10-07T22:58:53-05:00
Time in bed	380.8 min (6h 20.8m)
Total sleep	374.8 min (6h 14.8m)
Sleep efficiency	98.42%
Awake time	6.02 min
Core (N1+N2) sleep	218.8 min / 58.37%
Deep (N3) sleep	68.25 min / 18.21%
REM	87.8 min / 23.43%
CPAP usage hours	6.45
Session count	1
AHI	1.24
RDI	1.24
Obstructive apneas (OA)	7
Hypopneas (H)	1
Central apneas (CA)	0
Unclassified apneas (UA)	0
RERA	0
Pressure_Avg	7.69
Pressure_Min	4.00
Pressure_Max	12.62
Pressure_90th	10.72
Leak_Avg	1.50
Leak_Max	34.80
Leak_90th	3.60
FL_Avg	0.0217
FL_P95	0.12
FL_Max	0.55
Min_Above_005 (min)	52.4
Min_Above_010 (min)	23.3
Min_Above_020 (min)	4.6
Frac_NonZero (FL samples)	34.94%

Comparison to Trailing Averages

Metric	Tonight	7d avg	30d avg	Δ vs 7d	Δ vs 30d
AHI	1.24	2.06	2.09	-39.7%	-40.8%
OA count	7.00	8.86	9.33	-21.0%	-25.0%
H count	1.00	3.86	3.30	-74.1%	-69.7%
CA count	0.00	0.286	0.633	-100.0%	-100.0%
RERA count	0.00	1.86	0.867	-100.0%	-100.0%
Pressure_Avg	7.69	7.61	7.56	+1.1%	+1.7%
Pressure_90	10.72	10.30	10.14	+4.0%	+5.8%
Leak_Avg	1.50	1.60	1.06	-6.2%	+41.2%
Leak_Max	34.80	24.69	21.12	+41.0%	+64.8%
Leak_90th	3.60	5.49	3.40	-34.4%	+5.9%
Usage hours	6.45	6.48	6.38	-0.5%	+1.1%
FL_P95	0.120	0.094	0.094	+27.3%	+27.2%
Min_Above_010	23.30	16.84	16.56	+38.3%	+40.7%
Min_Above_020	4.60	4.57	3.92	+0.6%	+17.2%
FL_Avg	0.022	0.016	0.016	+32.0%	+39.0%
Frac_NonZero	0.349	0.263	0.251	+32.9%	+39.0%
Deep %	18.21	14.22	13.51	+28.0%	+34.8%
REM %	23.43	26.63	24.16	-12.0%	-3.0%
Awake min	6.02	7.89	9.58	-23.6%	-37.2%
TIB min	380.85	385.94	377.71	-1.3%	+0.8%
Total sleep min	374.83	378.06	368.12	-0.9%	+1.8%

Week-over-Week Deltas (recent 7d 2026-10-01–2026-10-07 vs prior 7d 2026-09-24–2026-09-30)

Metric	Recent 7d	Prior 7d	Δ	Δ %
AHI	2.057	1.806	+0.251	+13.9%
OA	8.857	7.857	+1.000	+12.7%
H	3.857	3.286	+0.571	+17.4%
CA	0.286	0.286	+0.000	+0.0%
Pressure_Avg	7.606	7.779	-0.173	-2.2%
Pressure_90	10.303	10.494	-0.191	-1.8%
Leak_Avg	1.599	1.114	+0.484	+43.5%
Leak_90	5.486	3.771	+1.714	+45.5%
Leak_Max	24.686	25.371	-0.686	-2.7%
Usage hours	6.483	6.313	+0.170	+2.7%
Deep % (Apple)	14.225	14.152	+0.073	+0.5%
REM % (Apple)	26.633	22.401	+4.232	+18.9%
Awake min (Apple)	7.885	11.039	-3.154	-28.6%
Sleep min (Apple)	378.057	364.081	+13.976	+3.8%
TIB min (Apple)	385.942	375.120	+10.822	+2.9%
FL_P95	0.094	0.097	-0.003	-2.9%
Min_Above_010	16.843	17.186	-0.343	-2.0%
Min_Above_020	4.571	3.357	+1.214	+36.2%

Recent 10-Night Context

Date	AHI	OA	CA	PAvg	P90	LAvg	LMax	L90	FL_P95	MinAbv010	MinAbv020	Deep%	REM%	Awake
2026-09-28	1.23	7	0	7.54	10.36	1.04	34.8	3.6	0.11	22.2	3.5	13.79	25.60	7.53
2026-09-29	1.45	8	0	8.10	11.34	0.44	33.6	0.0	0.07	10.6	2.5	13.30	15.99	21.08
2026-09-30	1.44	7	0	8.01	10.40	1.18	16.8	3.6	0.10	16.9	2.0	16.10	18.32	24.59
2026-10-01	0.56	4	0	6.82	9.00	0.53	8.4	2.4	0.07	10.9	2.2	16.96	22.96	16.56
2026-10-02	2.04	9	1	8.50	10.32	1.68	25.2	8.4	0.09	15.9	2.8	18.86	34.36	8.53
2026-10-03	2.33	7	1	7.21	9.30	2.77	22.8	10.8	0.08	11.3	2.5	9.63	25.46	10.54
2026-10-04	1.18	6	0	7.90	10.18	1.91	28.8	4.8	0.13	25.7	7.1	7.88	27.87	1.51
2026-10-05	5.20	21	0	7.93	12.36	1.76	22.8	7.2	0.09	15.7	8.2	15.92	25.21	8.03
2026-10-06	1.85	8	0	7.19	10.24	1.04	30.0	1.2	0.08	15.1	4.6	12.12	27.14	4.01
2026-10-07	1.24	7	0	7.69	10.72	1.50	34.8	3.6	0.12	23.3	4.6	18.21	23.43	6.02

CPAP → Sleep-Stage Correlations (all 99 matched nights)

Correlations computed across full available history (2026-06-23 → 2026-10-07).

Pair	the user r	Interpretation
AHI vs deep_pct	+0.057	negligible
AHI vs awake_min	+0.034	negligible
OA vs awake_min	+0.090	negligible
CA vs deep_pct	+0.046	negligible
Leak_Avg vs deep_pct	-0.137	weak negative
Leak_90th vs deep_pct	-0.109	weak negative
Leak_Max vs awake_min	+0.029	negligible
FL_P95 vs awake_min	-0.082	negligible
Min_Above_010 vs deep_pct	+0.016	negligible

Bucket comparison (Leak_90 median split, median = 2.4): - Low-leak nights (n=49): mean deep% 15.24, awake 14.33 min - High-leak nights (n=50): mean deep% 14.03, awake 12.26 min - Δ (high – low): deep% -1.21 pp, awake -2.06 min

The direction remains consistent with earlier reports: higher 90th-percentile leak is weakly associated with less deep sleep. Effect size is small (~2 percentage points of deep sleep lost on average on high-leak nights) but has been stable across weeks of data. Tonight's data goes against that trend — Leak_90th was only mildly elevated (3.6) while deep% was the highest in the window, showing the bucket signal is a weak average, not a reliable per-night predictor.

Caveats

• Tonight (10-07) is the second clean recovery night after the 10-05 breakthrough. The AHI 5.20 spike on 10-05 is now surrounded by clean event counts on both sides (10-04 AHI 1.18, 10-06 AHI 1.85, 10-07 AHI 1.24) and can be confidently treated as a one-off.
• Leak_Max 34.8 L/min is the fourth consecutive-pattern outlier (10-02, 10-04, 10-06, 10-07). This is now a persistent weekly pattern worth investigating — mask fit, cushion wear, or sleep position. 90th-percentile leak stayed modest (3.6 L/min), so these are brief transients, but the frequency has shifted from occasional to recurrent.
• Flow-limitation metrics (FL_P95, Min_Above_010, Min_Above_020) are all elevated vs 30d baseline, in some cases substantially. These are subthreshold events the auto-titrator did not catch as hypopneas; they do not represent clinical airway closure (clinical cutoff ~0.3 for sustained FL) but are the signal to watch over the next few nights, especially if correlated with the Leak_Max pattern.
• Deep sleep 18.21% is the 5th-highest reading in the recent 26-night window (after 09-25, 10-02, 09-23, 09-17) and well above the 7d avg of 14.22%. One strong night does not reset a trend — the 30d avg of 13.51% remains the baseline reference.
• Matched-night count is 99 (one more than the previous report — 10-07 joined the matched set after both exports became available this morning).
• The 2026-10-06 Apple Watch JSON was cloud-only (iCloud/Windows) at the start of this run and had to be staged to force materialization before trailing 7d/30d sleep-stage averages could include it. Yesterday's report was unaffected (that run materialized its own 10-06 copy); this is a routine iCloud eviction, not missing data.
• Early 2026-06 nights (via multi-day roll-up) retain zero-valued pressure/leak channels and remain low-confidence for CPAP metrics prior to ~2026-06-22.
• Correlations are directional/small-effect only — single-night deviations often exceed what correlations alone would predict (tonight is a clear example — Leak_90th modest, deep% record high).
• All flagged metrics for tonight remain within clinically safe ranges (AHI < 5, leak within mask spec). Flags mark deviations from the user's own recent baseline and connections to recent trends, not clinical thresholds.