When did the sleep signal separate from placebo?
The time-course report gives an early date and a later date. Each comes from a different comparison.
Week 4 and Week 20
A post hoc analysis of SURMOUNT-OSA reported significant improvement in peripheral AHI relative to baseline as early as Week 4. The estimated treatment difference versus placebo became significant at Week 20 for both peripheral AHI and sleep apnea-specific hypoxic burden.
Those dates do different jobs. Week 4 compares the treated group with its own starting measurement. Week 20 concerns the separation between treatment and placebo. The later date also covers the oxygen-burden measure. Compressing all of that into “sleep improves by Week 4” would discard the most interesting feature of the analysis.
There is something satisfyingly awkward about that interval. A treatment can produce an encouraging early trace while the more demanding comparison takes longer to emerge. The paper is a post hoc examination of the trials, which makes the timing an analytical finding rather than a timetable for an individual.
Baseline and placebo are different reference points.
The oxygen burden alongside the event count
The same time-course analysis tracked sleep apnea-specific hypoxic burden alongside peripheral AHI. It reported that the peripheral measurements tracked with the polysomnographic findings—the findings from the sleep recording.
An event count already gives the night a shape. The oxygen-burden measure adds another aspect of the disturbance. Fewer events would leave my curiosity about oxygen unsatisfied; here, the investigators actually examined another measure.
That extra measure earns its place here. Counting fewer interruptions and examining their oxygen burden give the overnight test more substance than a single tally could supply.
How closely did the sleep changes follow weight?
The investigators found that the magnitude of AHI and hypoxic-burden improvement was associated with achieved weight reduction in both studies. That association makes weight central to the explanation. It leaves room for further work on exactly how the changes connect.
Mechanism is the unfinished business here. How much can each aspect of improvement explain, and would the relationship look the same across people with different starting sleep problems? The cited analysis supplies a relationship to investigate, rather than a personal prediction.
The other half of the night’s account comes from the participants’ reports: did the measured improvements arrive with changes they could describe?