One strong recommendation, 41 patients at 3,650 metres, and 481,857 scored nights
Europe's new CPAP guideline gives its strongest grade to daytime sleepiness alone; a 41-patient trial cut the apnea index by a third, and 481,857 scored nights show what that index leaves out.
A joint European guideline published on October 11 graded continuous positive airway pressure by the outcome it changes, and gave its strongest grade to one symptom: excessive daytime sleepiness. Everywhere else the recommendation is conditional, including a suggestion that CPAP or no treatment is reasonable for people without daytime sleepiness. 1
Two studies published in the same seven days test what the apnea-hypopnea index can carry. Among 41 people living at 3,650 metres, two weeks of a cheap drug cut the index by about a third. 2 Across 481,857 scored nights, a second movement index fell as breathing events rose, a direction its authors read as undercounting. 3
The count and the symptom are different layers of a night, scored by different rules and moving for different reasons. The week's evidence says which layer a treatment decision can rest on.
The strongest recommendation CPAP has is for daytime sleepiness
The guideline comes from a joint task force of the European Respiratory Society and the European Sleep Research Society, with pulmonologists, cardiologists, patient representatives and methodologists on it. The panel framed its questions in the PICO format, ran a systematic review of CPAP against no treatment or sham treatment, and graded the certainty of each outcome with GRADE, the standard system for separating strong recommendations from weak ones. 1
The two grades mean different things. A strong recommendation says the benefits clearly outweigh the harms for most patients, so most clinicians should follow it. A conditional one says the balance depends on the individual's circumstances and values, which turns the decision into a conversation.
CPAP carries the strong grade for adults with obstructive sleep apnoea and excessive daytime sleepiness. The panel suggests CPAP conditionally for its effects on sleepiness, other patient-reported outcomes and blood pressure. Further conditional recommendations cover moderate-to-severe disease, older adults, and people with arterial or treatment-resistant hypertension. 1
Three groups get a suggestion that CPAP or no treatment are both reasonable: patients without daytime sleepiness, patients with atrial fibrillation where the aim is to reduce recurrence, and patients with impaired glucose metabolism where the aim is glycaemic control. 1
So the machine at the centre of this week's guideline has its certainty concentrated on how awake a person feels in the afternoon. The cardiac and metabolic reasons people seek treatment appear as conditional recommendations, and for two of them the panel puts treatment and no treatment on the same footing.
A drug lowered the apnea index by a third in 41 people at 3,650 metres
Acetazolamide is a carbonic anhydrase inhibitor that improves sleep apnoea in lowlanders and in newcomers ascending to high altitude. Whether it helps people who have always lived high was the open question. 2
The design was a randomised, double-blind, placebo-controlled crossover. Forty-one adults with untreated moderate-to-severe apnoea in Lhasa, at 3,650 metres, took two weeks of acetazolamide at 500 mg a day and two weeks of placebo in random order, with a one-week washout between them. Each treatment period ended with an overnight laboratory polysomnogram. 2
Against placebo, the total apnea-hypopnea index fell by 15.5 events an hour (95% CI 7.1 to 23.9) and the obstructive index by 16.6 (9.9 to 23.4), roughly a one-third reduction. Mean overnight oxygen saturation rose from 83% to 86%, a gain of 3 points (2 to 4). Self-reported sleep quality improved 7% (2 to 13), and stage 2 sleep and sleep efficiency both rose. 2
Nothing in the trial connects that one-third to a life. There was no cardiovascular endpoint, no test of daytime function, and no follow-up beyond the two treatment periods. The abstract reports no adverse-event rates. Its authors propose acetazolamide as an adjunct where CPAP is not feasible, which is a claim about the index and about oxygen, earned in a population living at an altitude most readers never sleep at. The 86% saturation means oxygen remained well below a sea-level normal even on the drug. 2
481,857 scored nights, and a movement index that shrinks where breathing events cluster
Periodic limb movements are rhythmic leg jerks during sleep, counted per hour as a PLM index. They are usually described as something that travels with sleep apnoea. A retrospective analysis of 481,857 adults who completed overnight polysomnography found the opposite relationship. 3
The cohort spanned 240 sleep centres in 30 American states, with technician-scored studies and a sleep and medical history questionnaire. Mean age was 54.4, mean apnea-hypopnea index 14.7 events an hour, mean PLM index 10.3. The PLM index was strongly negatively associated with every measure of sleep-disordered breathing, and the association held when the analysis was stratified by CPAP use. 3
It was also higher in White patients than in Hispanic, Asian and Black patients, and associated with diagnosed depression, insomnia score, elevated blood sugar and anaemia. Objectively it tracked a higher arousal index and lower sleep efficiency. 3
The scoring rules supply the likely explanation. A limb movement is counted only when a respiratory event does not account for the arousal that ends it, so where apnoeas are dense the respiratory event claims the arousal and the movement that follows disappears from the limb count. The authors reach that reading themselves and call for the criteria to be reviewed. 3
What the design supports is narrower than what the numbers invite. The study is retrospective and cross-sectional, comorbidities came from a self-report questionnaire, the abstract gives no effect sizes, and the authors describe the racial difference as possibly genetic and sociocultural at once. The finding is about an index that thins out where another index thickens. Any claim that leg movements cause depression or high blood sugar sits well beyond it.
Three ways of judging a night now sit side by side. The guideline rests on how a person feels during the day. The trial in Lhasa moved the event count with a pill and left the person unmeasured. The 481,857-night cohort shows a second event class fading as the first thickens, so each count it reports is the product of a scoring decision. A night holds several event classes, each scored against its own threshold, and the class that dominates a report suppresses the others. The symptom has one advantage over every count: it is measured against the person rather than against a rule.
Insomnia's liability runs toward inflammation
The complaint itself has a measured biological footprint, and the direction of it was the week's question. In the UK Biobank, C-reactive protein was elevated in 130,611 people with insomnia compared with 327,375 controls. 4
A genetic analysis then mapped what the two traits share: a significant positive genetic correlation, 125 loci associated with both, and colocalisation supporting shared causal variants at six of them, in or near APOE, HLA-B, NLGN1, AK5/DNAJB4, FAM114A2/GRIA1 and SATB1. Functional annotation pointed at chromatin organisation and nucleosome assembly; pathway analysis at immune signalling, cellular senescence and programmed cell death. 4
Bidirectional Mendelian randomisation gave stronger support to insomnia liability raising CRP than to the reverse. That direction is the result to keep, with its conditions attached: the method assumes the genetic variants act on the outcome only through the exposure, the sample is UK Biobank, the abstract reports no effect size, and CRP is one inflammatory marker among many. 4
For a reader deciding how hard to push on a sleep complaint, this is the layer beneath the symptom. The guideline graded daytime sleepiness strongly because that is what the trials measured. The genetics add a route from the complaint into systemic inflammation, which raises the stakes on treating it as something more than discomfort.
What the ring calls sleep efficiency is still a family of definitions
Most consumer trackers report something called sleep efficiency, and the values get compared across brands in forums, in review articles and in research. A specification posted online this week in Sleep Medicine says the underlying measures are defined, computed and represented differently across the industry. 5
A working group of academic and industry experts set out a minimum viable standard: a common semantic framework for what it calls fundamental sleep measures from consumer health trackers. The group standardised the data layer rather than the product interface, on the reasoning that manufacturers need room to compete on algorithms and apps. The document specifies metadata, provenance and quality indicators, and names the circumstances in which a wearable-derived sleep measure should be read with more caution. 5
None of that changes tonight's score. It changes how much weight two numbers from two devices can carry when they are set beside each other, and it identifies the missing layer under a comparison any reader makes the week they switch rings. A specification is also the first step in a long process: manufacturers would have to adopt it before two brands' numbers could be compared.
At 71 degrees north, the roster set sleep and the season barely did
The Arctic is the natural test of the idea that light runs the body clock. At 71 degrees north the sun stays up for weeks in summer and stays down for weeks in winter, a swing that should leave a large seasonal mark on sleep. 6
Industrial process operators at a Norwegian plant were followed through both photoperiods: 111 workers, 81 of them across both seasons, with daily sleep diaries and a smart ring. Their roster ran seven consecutive morning shifts, four free days, then seven consecutive night shifts, and linear mixed-effects models compared shift types, seasons, chronotypes and consecutive workdays. 6
Shift type moved sleep timing and duration markedly. Season produced no overall effect, beyond a delayed sleep onset and mid-sleep after the first morning shift during the dark months. The probability of sleeping under six hours was higher on both morning and night shifts than on free days. Late chronotypes slept later on free days and longer after night shifts. 6
For anyone on a rotating schedule, that is a fair account of the leverage available: the roster decides, and where a chronotype sits against the roster decides the rest. The study is observational, drawn from one plant with one ring model, and its average worker may not resemble the reader.
Five more numbers
- Children did not compound. Seventy-nine children aged 6 to 12 wore actigraphy at home for 16 days; 40 of them slept under nine hours on weekdays and at least nine at weekends, and 39 were well rested throughout. No measure showed a group-by-day interaction (p>0.61), and the short-sleep group showed no weekday-to-weekend difference or worsening across a second short-sleep cycle. That group did report more sleepiness (Karolinska scale 4.51 against 3.77, p=0.01) and less positive affect (12.84 against 15.15, p=0.01). One season, 79 children. 7
- Anxiety showed up as slower slow-wave dissipation. In 70 conscripts aged 18 to 24 who spent a night in a sleep laboratory, the 24% with elevated anxiety had a smaller fall in slow-wave activity between the first and third NREM period, and a larger share of N3 sleep. Anxiety without depression also came with less REM and higher alpha power during REM. Sixty-seven of the 70 were male. 8
- Sleep pressure is encoded as an oscillation, in flies. Dorsal fan-shaped body neurons in Drosophila flip between two membrane states through competing half-centres joined by inhibitory glutamate synapses. Disrupting glutamate release abolished slow-wave activity and rebound sleep; replaying the oscillation optogenetically induced sleep. The amplitude tracks sleep need continuously, while behaviour switches between two states. 9
- A questionnaire-built circadian index predicted hypertension. Among 107,270 UK Biobank participants free of hypertension at baseline, an index built from chronotype, sleep duration, neuroticism, coffee intake and vitamin D carried a hazard ratio of 1.34 (95% CI 1.25–1.44) for the highest group against the lowest, over 10,311 cases and 1,690,083 person-years; combined with high genetic risk the hazard ratio was 2.43 (2.11–2.81), with additive interaction. The index is a composite of self-report and one blood test, so it describes a profile rather than anyone's clock. 10
- Higher narcolepsy doses helped, with no control arm. In an open-label study, 48 people with narcolepsy moved from 9 g of low-sodium oxybate to a mean 11.1 g a night. Mean Epworth sleepiness fell from 11.4 to 8.3, and the narcolepsy severity score fell from 17.3 to 9.4 in type 1 and 9.7 to 4.3 in type 2. Apnea-hypopnea index and oxygen saturation held steady; nausea and vomiting each affected 12.5%. Without a placebo arm, regression and expectation sit inside the improvement. 11
The maker desk and the researcher desk
No wearable maker published new sleep research this week, which is worth recording because two of them did last week.
Oura's blog feed carried one new post, a buying guide dated October 13, and the newest post in that feed with original data is still its September 11 tennis analysis. 1213 Eight Sleep's newest entry is still the pregnancy and postpartum study of October 11, covered here last week. 14 WHOOP's Locker page rendered with no October 2026 entry among its listings, and its press centre answered with a bot check. 15
On the research side, Matthew Walker's episode 155, published October 12, takes apart nocturia, the condition of waking at night to urinate. The episode argues that the bladder is usually the messenger: age-related falls in nocturnal vasopressin, cardiovascular problems, sleep apnoea and insomnia are what drive the awakenings. 16
One two-week record worth keeping
The guideline's strongest grade rests on a symptom, so the most useful artifact a reader can produce this month is a record of that symptom. Pick a fixed hour in the afternoon, rate your sleepiness from 0 to 10, and score it every day for two weeks before changing anything. Add a standard sleepiness questionnaire once a week if a second handle on the same quantity helps.
Then change one thing: a bedtime, a morning-shift week, a machine setting agreed with your clinician. Keep scoring for another two weeks, and compare the halves on their averages and on their worst three days. A drop that survives both readings is a change worth keeping.
If you use CPAP, the same record answers a question the apnea index cannot, because the index and the symptom are partly independent layers. Bring both weeks to the appointment. The guideline's strong recommendation was earned on the layer the record measures, and the case for staying on treatment gets settled in the conversation that follows.
Nothing here is a reason to stop a prescribed treatment on your own.
References
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12Oura blog feed
ouraring.com
- 13What to Know When Buying an Oura Ring
ouraring.com
- 14Eight Sleep blog
eightsleep.com
- 15WHOOP The Locker
whoop.com
- 16#155 - Sleep & Nature's Call
themattwalkerpodcast.buzzsprout.com
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.