12 weeks on semaglutide: A1c down 1.6, dawn spikes worse?

Congrats on the A1c drop. My dawn rise got sharper around week 10 too. Lower average glucose makes a 30-point bump look huge. Did overnight lows change?

Hearing the opposite from a lot of folks. That 8-12 week stretch seems to flatten the 3am rise — but the evening walk sliding later lines up with the same window. Could be timing, not the med at all.

Not necessarily dawn phenomenon. A bigger or later dinner can cause slow digestion and a 3am bump. What does your 1am–2am look like?

I've worn a CGM for two years and saw a similar pattern when my A1c fell fast. My theory is that lower average glucose makes normal overnight cortisol and glucagon bumps more visible. My 3am rise went from 130 to 155 for about six weeks, then settled to 120–135 as my baseline adjusted. Not medical advice, just my n=1. What are your bedtime and 2am readings?

The part people forget is that a 5am number can be a rebound, not a baseline. Pattern: dip to 68 around 2am, then a liver dump to 150 by 5am. A CGM is what makes that obvious — a single morning fingerstick just shows the 150 and looks fine.

That reactive pattern is what I wondered about. Did you confirm with fingersticks? My CGM reads low overnight when I sleep on the sensor.

I buy the cortisol theory, but I wouldn't write off the dawn rise. My fasting number drives morning mood and hunger. What's your time in range? Does it dip overnight?

My CGM did the same weird thing around weeks 9–11: 3am around 98–104, then 132–140 by 6am, even while my 7-day average was dropping and I was down about 16 lbs. The surprise for me was that it tracked sleep debt way more than the scale—solid 7-hour nights gave me maybe a 15-point rise, but 5-hour stressy nights were

Weeks 8–12 for me were less about the overnight curve and more about a nine-day scale stall where my waist still dropped 1.5 inches and my ring got loose — but my morning finger stick ran 8–12 points higher on nights I woke up around 2 to pee or slept under 6.5 hours. I was convinced the med was doing something weird to my mornings, then I looked back at my notes and the pattern was way more tied to late salty dinners and only sipping water after 7:30; on the nights I actually slept 7+ hours and cut the evening fluids, my fasting number was back near my usual 118–125. Not saying that explains your labs, but if you’re tracking anyway, the sleep/hydration log was the one thing that surprised me more than the overnight graph.

I’m in the two-kid chaos too, so my CGM is basically a sleep log. My 3–6am climb got steeper once my toddler started waking at 2 and my older one at 5. On the rare nights I actually slept through, the line was way flatter—same 7pm dinner and all. I started noting wake-ups right in the CGM app instead of chasing the number. Worth checking if your rise lines up with more night wakings lately, not just food. Did anything shift at home or sleep around the time it got worse?

One weird thing that cleaned up my overnight graph: timing of my last 100 calories, not dinner itself. I stopped the 8pm handful of nuts/cheese and moved my evening walk earlier; my 4am number dropped about 12 points while dinner stayed the same. Also, do you log alcohol? Even one drink around 8pm gave me a stair-step 3am bump. Curious whether your climb is smooth or stair-step — mine was stair-step on nights I got a second wind before bed.

Congratulations on the A1c and weight trend—those are meaningful wins. One new thing I track alongside weekly averages is daily steps, especially after dinner. My 3–6am slope was noticeably flatter on 8k+ step days, even when my evening routine was otherwise consistent; fasting labs took another week or two to reflect it. Have you compared high-step and low-step nights? For me, it separated activity effects from other variables. Wishing you continued progress.

Congrats on the A1c win—CGM overnight drama is such a troll. I stopped chasing every dawn bump and just looked at weekly trends.

Curious if you’ve overlaid the 3–6am rise with sleep quality/wakeups rather than food. I did that for a month and found my “dawn spikes” were mostly nights I was surfacing at 2–3am—dry mouth, weird dreams, then the graph ramped. On nights I slept straight through, my climb was way flatter. Sleep schedule was identical; sleep continuity wasn’t. Might be worth tagging wakeups in your CGM notes and seeing if the pattern splits. What does your graph look like on a rare solid-sleep night?

Curious if you’ve overlaid the overnight rise with lifting days vs rest days, not just sleep. I lift 4x/week and my 3–6am curve is way uglier after heavy lower-body sessions—higher by ~15–20 and starts earlier—even with the same 7pm last meal. I started tagging session RPE and volume in my CGM notes, and the correlation is annoyingly tight. If your A1c/weight are improving but dawn spikes are worse, has your training load crept up over those 12 weeks? That’s the variable I’d check next.

Mildly contrarian: I’d stop staring at 105→145 and plot the rise above your 2–3am nadir. When my A1c dropped, my nadir fell faster than my dawn peak, so the overnight “spike” looked worse while my absolute 6am number was basically flat. What’s your lowest 10-min median between 1–3am? The only thing that ever bent my 3–6am slope was evening alcohol—even one drink added ~15 points by 5am. Not saying that’s your story, but it’s a cheap column to check before the curve eats another spreadsheet tab.

Huge on the A1c drop and 14 lbs! One thing I’d nerd out on: not just when you finished eating, but what dinner looked like. I saw my 3–6am slope get steeper on low-protein/higher-carb dinners even with the same 7pm cutoff, and calmer on nights with more protein/fiber. Have you logged dinner composition against that rise? A quick note like “pasta” vs “chicken + veg” might show a pattern without changing anything else.

Random CGM nerd thing: is your sensor on the arm you sleep on? I lift 4x and track sleep, and I had an almost identical 3–6am climb. Turned out to be sensor compression—sleeping on that side gave me a fake dip, then a rebound. Moved it to the other arm and the “dawn spike” mostly disappeared. The only reason I caught it was fingersticking a couple of those mornings. Does yours change if you sleep on your back or opposite side? That was my tell.