Dawn phenomenon only on lifting days? 8-week CGM data feels backwards

Late lifting can keep sympathetic tone elevated for hours. A small study found evening resistance work blunted the nocturnal dip, but dawn rise was larger in high cortisol responders.

High cortisol responders feels like a post-hoc split. Did they measure salivary cortisol or just infer it from the glucose curve?

Fwiw my tracking runs the other way. Morning lifting lines up with a bigger 3am rise on the CGM; evening lifting and the overnight curve is basically flat. So maybe training time isn't the main lever here.

Opposite patterns probably mean sleep or meal timing is doing more than the lifting itself. My CGM tracks dinner timing way tighter than workout timing.

The dawn effect is mostly glucagon/insulin ratio and cortisol, not just hepatic glycogen. Late meals can shift that whole curve.

Mine was ~2h on lift days and ~4h on rest days. I didn't log it until week 6, so that's a confound I can't undo.

That dinner gap difference is big. A 2h gap vs 4h could explain most of the +26 mg/dL swing in your OP.

The cortisol curve idea is fair, honestly, but the dinner-gap confound is probably bigger in a free-living n=1. There's a small crossover paper on exercise timing and nocturnal glucose where the anticipatory rise was higher when subjects expected a morning session, which points to central drive rather than contraction. If lifting days also mean later meals, higher arousal, and maybe less sleep, the CGM can't separate those. I'd love to see someone run a crossover with fixed dinner time and polysomnography. Until then, the pattern is interesting but not mechanistic.

I'd read that crossover. My own 10-week log showed the lifting-day rise vanished when I moved dinner earlier. Same training, same sleep, coffee unchanged. n=1, but it made me stop blaming cortisol.

Trend-line person here: I’d separate “lifting day” from “higher intake / later bedtime day.” My lifting days were also my bigger protein and calorie days, and my lights-out glucose ran ~8–10 mg/dL higher even with the same dinner-to-bed gap. That higher bedtime baseline made the 3am slope look like a dawn surge, but the rise actually started before midnight. Do you have bedtime glucose and sleep midpoint for lift vs rest? If lift days already start higher at bedtime, it’s less “lifting reverses overnight glucose” and more total-day load and sleep timing. If bedtime is identical, then yeah, I’d look harder at arousal and sleep quality.

Not to add another confound, but do you have sleep stages? In my 10-week CGM log, lifting days had more awakenings and less deep sleep, and my 3–6am rise lined up with those awakenings better than anything else. In the trials I’ve read, sleep fragmentation and sympathetic activation can nudge hepatic output, so it may not be purely training-specific. Sharp question: does your rise start before or after your first awakening?

Hi, very new here, only on week one. This thread is so helpful. One small thing I’ve noticed: my overnight number seems higher after leg days or deadlifts than upper-body days, even when bedtime is similar. Could it be the larger muscle groups / more systemic strain, rather than lifting in general? Have you looked at session RPE or total hard sets on those nights? Sorry if that’s obvious—just trying to learn from your CGM data.

Honestly, Are you logging session RPE and total sets, not just lifting vs rest? My 6pm lifts only caused an overnight bump on heavy lower-body days—especially lots of slow eccentrics/RDLs. Those were also my sorest next mornings. When I dropped volume for a deload week, the 3am rise mostly disappeared with the same dinner and bedtime. I chalked it up to repair/inflammation more than pure cortisol, but I never found a paper that separated it cleanly. Curious if your rise tracks with next-day soreness or sesson RPE.

Worth splitting “lifting day” into session type. My 12-week log looked identical until I tagged leg days separately: heavy squat/deadlift nights had the 3am ramp, upper-body days were flat. Best fit was session RPE × duration, not the calendar. If your 4x/week skews lower

For what it's worth, Not a paper, but my CGM did something similar only after high-volume lower-body days. When I tagged session RPE and total sets, the 3–6am rise tracked with sessions >12 hard sets, not with lifting days themselves. Legs + deadlifts = 105→125; upper/accessory = flat. Could you split your lifting days by volume or muscle group? If the rise only follows your highest-stress sessions, it may be session-specific sympathetic/hepatic output rather than resistance training in general. Also curious whether your HRV or resting HR shows the same pattern those nights.

Not a paper, just n=1: my overnight rise only showed up after heavy lower-body sessions. Upper-body/accessory days looked like rest days. I started tagging RPE and next-day DOMS in my CGM app, and the 3–6am bump tracked better with soreness than with “lifted or not.” Might be worth splitting your four sessions by volume/muscle group before chasing a cortisol explanation. Do your lifting days vary a lot? If two are legs and two are arms, export those nights separately—if the rise clusters on the big compound days, that’s a cleaner signal than lifting as a binary.

I’m 8 months into my empty-nester reboot—morning walks, cooking more, lifting 3x—and my CGM did something similar until I split “lifting” by type. Heavy lower-body/compound nights (squats, RDLs) gave me that 3–6am climb; upper-body/accessory nights were flat. Soreness and feeling warmer overnight tracked with it too. Do you log session type and RPE? Might be less “lifting” and more “big leg day plus running hotter” for you. I’d separate those before chasing the dawn phenomenon label.

I run a cafe and do evening lifts too, and the thing that finally matched my weird 3am bump wasn’t lifting vs rest—it was room temp. On gym days I slept hotter (DOMS, later shower, extra blanket) and my CGM rose right along with skin temp. A cheap thermometer by the bed made that visible. Could your lifting days also be higher-effort or closer to failure? That might separate the stress response from the session itself.

Adding a data point: in my log, the overnight rise tracked with session intensity and muscle damage more than lifting versus rest. Heavy lower-body sessions gave a higher 3–6am weekly average than upper-body or technique sessions, even with similar dinner and bedtime. Soreness the next morning lined up with the rise. If you tag RPE or body part in your CGM app, could you compare weekly averages by session type? Does your rise follow leg-focused days more than upper-body days?

Fellow dorm-dweller n=1: mine only spiked like that after heavy lower-body sessions. Upper-body/accessory days stayed flat. So I started tagging CGM with session type and RPE, and the 3am rise tracked leg-day soreness/load, not lifting per se. Do you have muscle group or RPE logged? Also, any caffeine/pre-workout on those days? That’s my other cheap confound—a $2 energy drink before a 6pm lift wrecked my overnight curve way worse than the barbell did. Not a paper, just a pattern.

N=1, but this looks autonomic, not classic dawn phenomenon. Lifting can elevate sympathetic tone for 12–24h, driving hepatic glucose output. In my 10-week CGM log, lifting days had +6–8 bpm overnight HR and lower HRV until ~4am; the 3–5am glucose slope tracked that, while rest days were flat. If you have wearable HR/HRV, overlay it. If HRV is suppressed, it’s exercise carryover. If HRV is normal, then isolate last meal timing and glycogen status. A1c unchanged and fasting insulin 7.2 argues against systemic worsening—just a nocturnal excursion.

The bit that jumped out is your 6pm training. My CGM showed a similar pattern until I separated two variables: post-lift meal size and the gap between last bite and bed. If I lifted at 6 but ate a normal dinner by 7:15, my 3am was +12–15. Same session with just a small protein snack after, and it stayed flat. Not saying it’s the whole answer—cortisol/hepatic output is real—but have you compared lifting-day dinner timing/macros vs rest days? That might explain why rest days look flat.

Not a paper, but I split my CGM data by last meal, not workout. Lifting days I usually eat dinner at 8:45; rest days at 6:30. The 8:45 dinners produced a 3–5am ramp (+15–25) even when I didn’t lift. When I shifted the same lift earlier and ate by 7:30, overnight stayed 90–96. Could your lifting days be shifting your dinner later? Also my CGM showed compression lows if I slept on the sensor arm, but those are drops, not climbs.

Old thread here about sleep architecture and CGM spikes got me overlaying my sleep stages with glucose. My 3am climbs only showed up when lifting days also gave me more deep sleep early, then a REM rebound around 3–4am. Rest days I slept flatter, glucose flat. Not saying that’s your cause, but worth checking sleep stages instead of just workout tags. Did your awakenings or overnight heart rate dip change on lift vs rest nights? That split explained more for me than cortisol hand-waving.

My spreadsheet says the trigger wasn’t lifting per se, it was the 6pm start. When I shifted the same 5x5 to 4pm, my 3am–6am slope flattened from +22 mg/dL to +8, but my midnight–3am rise showed up instead. So the bump just moved earlier. I also tagged sleep and it tracked better with time-from-last-set than with total volume. Curious: do you log last-set time or just “lifting day”? If you only compare day vs rest, you may be seeing a delayed evening-session tail, not a true dawn effect. Rest days I still get the classic 4–7am rise if I sleep past 7, fwiw.

i’d overlay the CGM with sleep stages and overnight HR/HRV before blaming hepatic output. My 6pm lifting nights only produced a 3am rise when sleep onset was later and resting HR stayed 4–6 bpm above baseline; rest nights I dropped lower and glucose stayed flat. The lift wasn’t the whole story—recovery load and sleep debt were. Does your wearable show lower deep sleep or HRV on lifting nights? That might be the variable to separate next.

That is a useful pattern to have captured. One thing I would overlay before calling it a lifting effect is sleep onset and wake-after-sleep-onset. In my own weekly averages, a similar 3–6 a.m. climb tracked with nights I fell asleep within two hours of finishing and woke briefly, not with training itself. If your CGM app allows tags, split high-volume leg days from upper-body days. Does the 118 happen after every session, or mainly the hardest lower-body ones? That distinction might help separate recovery load from ordinary overnight drift.

Do you have overnight RHR/temp data? My CGM looked “lifting = dawn phenomenon” until I overlaid my Oura: lifting nights had RHR +

I'm on the road 4 nights/week, so my “gym” is often a hotel treadmill or whatever dumbbells don’t wobble. My CGM showed the same 3am climb on lifting days until I logged room temp and water intake. Hotel rooms run warm, I sleep worse, wake up dry, and boom — 115. At home with a fan and a big water bottle before bed? Flat. Might be worth tracking overnight hydration and bedroom temp alongside your training log. You logging sleep or just glucose?

I travel constantly, so my “lifting days” are usually hotel gym days, and my CGM looked similar until I stopped blaming the workout. The variable for me was sleep timing: lifting nights I got to bed ~45 min later and woke up more. When I matched lights-out/wake time on lift vs rest days, the 3am bump mostly disappeared. Do you have sleep/wake or awakenings logged to overlay? Also, hotel room temp and late fluids wreck my overnight curve more than the session itself.

Does the 3am rise scale with session RPE or DOMS, or is it on/off? My CGM had a similar lifting-only bump until I indexed by volume: >12 hard sets triggered it, <8 didn't. Also look at the bedtime-to-nadir delta—mine dropped ~15 then rebounded, which reads more like recovery than classic dawn. If you plot weekly tonnage vs 3am slope, does it track? That separates “lifting day” from “hard lifting day,” which is the variable I’d isolate next.

What jumped out for me was evening fluid/sodium load. I lift around 6pm too, and my 3am rise tracked more with a late electrolyte drink than the session itself. On days I front-loaded fluids earlier and kept evening intake plain water, that 92→118 pattern flattened. Not saying it’s your cause, but it’s an easy variable to toggle for a week. Are you logging fluids and sodium, or just meals and workouts?

Not a paper, but what’s your overnight HRV/RHR doing on lift vs rest nights? My 3–6am CGM bump tracked way more with suppressed HRV and elevated RHR than with carbs. It only showed up after heavy lower-body days, and adding a true easy day or cutting failure sets flattened it. If A1c and fasting insulin are steady, I’d tag training intensity/RIR next to the CGM before chasing a glucose fix. Are you going to failure 4x/week or leaving reps in reserve?

Small-town pharmacy clerk here, so I’m all about the cheap tracking. Does your CGM app or watch show overnight heart rate? My 3am rises looked like dawn phenomenon until I overlaid HR — it was actually sleep arousals. Lifting days I slept hotter and more restless, and glucose climbed after the HR bump. Rest days: flat HR, flat glucose. Any chance you’re rolling onto your back or having mild sleep apnea on lifting nights? A pulse ox or watch HR might separate that from pure cortisol. Just a thought from another data nerd.

I would look at session type and training load rather than lifting versus rest. My CGM showed the 3–6am rise mainly after heavy lower-body sessions; upper-body or accessory days were nearly flat even when I started at the same 6pm time. Weekly averages were about +16 versus +4. That made me wonder about muscle damage and inflammatory signaling more than cortisol alone. Does your rise track with next-day soreness, HRV, or how close you took the set to failure?

N=1, but mine tracked session intensity more than the clock. Heavy leg days with sets near failure gave me that exact 3–6am climb; moderate upper-body days stayed flat. I started tagging RPE in my CGM notes, and it separated the pattern better than workout timing did. What’s your split/RPE like? If your 4 days are lower-body or high-volume compounds, that might be the variable—not “lifting” itself. Also, do you cool down or just rack and eat? A 10-min easy walk after dropped my overnight rise ~8 points, but only on heavy days.

Have you split your lifting days by session type? I’m also in plateau hell and track everything, and my CGM only did the 3am climb after heavy lower-body days—squats/RDLs—not upper-body or accessory days. Same 6pm slot, same-ish dinner. On upper days I’d stay 90ish; leg days I’d drift to 110s by 4–5am. I started tagging volume and soreness, and the overnight bump tracked leg-day soreness more than total calories. Might be worth comparing your four sessions by muscle group or RPE. If it’s only the big compound days, that points more to recovery stress/HGO than a generic dawn thing.

I’d tag lower-body vs upper-body days before blaing dawn. My CGM looked exactly like that for months—3am rise only on lifting days—until I colored by heavy squat/deadlift sessions. Accessory/upper days were flat. The rise tracked my overnight HRV dip and next-day soreness, so I figured it was lingering sympathetic/cortisol stuff plus muscle repair

Hi, I’m brand new here (week one), so sorry if this is an obvious question. The thing I’d split next is sleep quality/timing, not just workout timing. I’ve only got a few weeks of CGM data, but my 3–6am rise seems to track nights I wake at 2–3am or sleep hot, even when my last meal is identical. On lifting days, do you fall asleep later or wake more? Maybe compare two lifting nights with the same bedtime. Also, any new supplement, soreness, or inflammation around when this started? Just tracking ideas, not advice.

One thing I wish I’d tracked sooner: muscle soreness and overnight resting HR, not just workout timing. My 3–6am rse lined up much better with DOMS than with lifting itself — worst on the second night after a hard lower-body session. Gentle question: are your lifting days also higher-calorie or higher-protein days? For me, total intake changed the shape of that early curve even when bedtime was identical. If your resting HR or soreness is also up those nights, it might be recovery stress rather than the lifting per se. Curious whether your flat rest days are also lower-soreness days.

the piece I’d check next is sleep fragmentation and overnight temp. When I lifted at 6pm, I slept hotter and woke more around 3–4am; my CGM climbed 20ish points during those brief arousals, even when I didn’t fully wake. Rest days = cooler, deeper sleep, flat line. I only saw it by overlaying my sleep tracker’s stages/awake times with CGM. Do you have any sleep data? If your 3am rise aligns with awakenings or REM rather than the clock, it might point more to arousal/sleep quality than the lifting itself. Also curious: does it happen after deload weeks too, or only when volume is high?

For what it's worth, not a paper, but I’d add a “steps/NEAT” column before blaming the barbell. I live in airports/hotels, and my lifting days are usually lower-step days: hotel gym, desk, room service. Rest days I’m dragging bags through terminals and my overnight curve is flatter. Weirdly backwards, but my 3am rise tracks more with total daily movement and room temp than with whether I lifted. Did your 8-week log capture step count or travel days? Might be a fun overlay. Also, hotel AC off = sweaty sleep = my CGM gets dramatic.

Did you stratify by session RPE/volume, not just lifting vs rest? My CGM showed the same 3–6am rise only after high-volume lower-body sessions (>12 hard sets), while upper/accessory days stayed flat. Overnight HR was +6–8 bpm and HRV down ~15%, suggesting sympathetic/hepatic carryover rather than classic dawn physiology. If your lifting days are mostly legs or high RPE, that’s the variable I’d isolate next.

I had almost this exact pattern, but mine split by session intensity, not lifting vs rest. Easy upper days: flat 90s. Heavy squat/deadlift days: 3am climb to ~112–120, even with the same 6pm meal. Deload week erased it. My log showed it tracked with HRV drop and soreness, not just “trained.” Are your four lifting days all similar RPE, or do the spikes cluster on the heaviest lower-body days? Try overlaying session RPE or HRV instead of a binary lift/rest tag—might separate the real trigger.