Your sleep architecture changed at 45, even though your bedtime didn't.

The Fitbit says you slept seven hours. Your body says you slept four. Both are right — and the gap is what the wellness internet keeps missing.

The most common sleep complaint from men over 45 is a version of the same sentence. "I'm in bed for the same hours. The watch says I slept seven. I feel like I slept four." This is not a perception problem. It is an architecture problem.

Deep sleep collapses by 80 percent

Ohayon and colleagues, in a 2004 Sleep meta-analysis of 65 studies and 3,577 healthy subjects, gave us the cleanest age-curve we have. The headline: slow-wave sleep (N3, deep sleep) declines by about 2 percent of total sleep time per decade. From 30 to 70, that is a roughly 80 percent collapse in deep-sleep fraction.

Total sleep time barely moves. REM declines modestly. Wake after sleep onset roughly doubles between 30 and 60. The thing that collapses is the deep, restorative slow-wave fraction, the thing your watch can't reliably distinguish from light sleep.

Why deep sleep matters more than total hours

Slow-wave sleep is the physiologically distinct window during which most of the male endocrine and neurological reset happens overnight. Growth hormone pulses occur almost entirely during N3. Testosterone production rises sharply. Glymphatic clearance — the brain's overnight waste-disposal system — runs preferentially in slow-wave. Memory consolidation depends on it.

Mander and colleagues, writing in Neuron in 2017, made the case explicit. The decline in slow-wave sleep is not an inconvenience of aging. It is causally upstream of much of what we call cognitive aging: memory complaint, executive decline, mood instability, and the slow attrition of recovery.

"I slept seven hours" is often the wrong question. If your deep-sleep fraction has fallen from 22 percent at 30 to 8 percent at 50, you got the same hours and meaningfully less of the sleep that does the work.

The melatonin amplitude problem

Melatonin drops two ways with age. The mean level falls modestly. More importantly, the amplitude of the curve — the contrast between daytime trough and nighttime peak — flattens. A young man's curve has a sharp signal. An older man's has a noisy one. Bright light at 10 PM costs the older man more than it cost him at 30.

This is why morning light matters more, not less, after 45. Twenty minutes of outdoor light within an hour of waking does more for midlife male sleep than most supplements.

The prostate is now part of your sleep

Nocturia — waking to urinate one or more times per night — rises sharply through the 40s and 50s, driven by BPH and reduced bladder capacity. About a third of men over 50 wake at least twice per night. Each awakening fragments the sleep cycle and reduces the probability of returning to slow-wave on the next descent.

Two prostate-driven awakenings per night, repeated for years, will degrade the deep-sleep fraction more reliably than most factors a man can modify. Pelvic floor work, evening fluid timing, alcohol reduction, and pharmacological management of BPH when appropriate all belong in the sleep conversation.

Sleep apnea, the most under-diagnosed midlife sleep problem

The Wisconsin Sleep Cohort, in the Young et al. 2002 paper, established the prevalence numbers that still anchor the field. Roughly 24 percent of men aged 30 to 60 have an AHI of 5 or greater. Roughly 9 percent have moderate-to-severe apnea (AHI 15+) — stopping breathing at least 15 times per hour, every hour, every night. The majority are undiagnosed.

The screening criteria you can apply tonight. STOP-BANG: Snore loudly, Tired during the day, Observed stopping breathing, high blood Pressure, BMI over 35, Age over 50, Neck over 17 inches, Gender male. A score of 3+ warrants a home sleep study. 5+ warrants one urgently.

Untreated moderate OSA raises the risk of hypertension, atrial fibrillation, stroke, and all-cause mortality. It also obliterates slow-wave sleep — apneas preferentially fragment N3 — so a man with undiagnosed OSA is losing on two timescales: the slow drift of aging plus the acute nightly fragmentation.

The cortisol awakening response shifts

In a young man, cortisol rises an hour before waking and peaks roughly 30 minutes after, producing a sharp "ready to move" signal. In an older man, that spike is blunted and delayed while the evening floor is higher than it should be. The result: a man who wakes at 3 AM with the lights on in his brain, then is sluggish at 7 AM when he needs to function.

Van Cauter and colleagues, in a frequently cited JAMA paper, traced this directly. The same older-male cohorts that had lost slow-wave sleep also showed flattened cortisol and growth hormone profiles. The three signals move together because they share the same underlying architecture.

This pattern is treatable but not by supplements. The interventions are unsexy: morning light, fixed wake time, strength training, evening light hygiene, alcohol off after dinner.

Why the Fitbit lies

Consumer wearables estimate sleep stages from HRV and movement. They are reasonable at wake-vs-asleep, mediocre at distinguishing light from deep sleep, and they systematically over-report deep sleep in older adults because the HRV signature that defines N3 attenuates with age. The watch sees a steady, low-variability heart rate and codes it as deep. Increasingly with age, it is wrong.

Trust total-time and awakening data. Treat stage data skeptically.

What to actually do

The interventions that work are unfashionably simple. Fixed wake time. Morning outdoor light within an hour of waking. Last meal and last big fluid three hours before bed. Alcohol off by dinner. Bedroom under 66°F, dark, quiet. STOP-BANG, then home sleep study if positive. CPAP if warranted — the compliance gap is real, but men who stick with it report the most dramatic recovery of any intervention in the midlife stack.

Magnesium glycinate (200-400 mg) has reasonable evidence for slow-wave preservation. Low-dose trazodone short-term can help re-establish architecture. Melatonin (0.3-0.5 mg) for circadian re-anchoring, not megadoses.

The bottom line

The sleep you had at 35 is not the sleep you have at 50, even if the time on the pillow has not budged. Deep fraction collapsed. Cortisol curve flattened. Prostate joined the party. Melatonin signal gone fuzzy. And in roughly a quarter of midlife men, undiagnosed apnea on top.

The architecture is not immovable. It is responsive to light, timing, strength training, alcohol, apnea treatment. The watch number is the least important part of the picture. What matters is whether the biology of overnight repair is happening, and at 45 that question is no longer rhetorical.