week 18 on compounded semaglutide, down 22 lb from 204, and I’ve got a side effect I haven’t seen named cleanly anywhere so I want to throw it out and see if the pattern matches for anyone. days 2-3 post-injection I run cold. like, hoodie on when nobody else in the house is cold, white-tipped fingers when I’m typing, feet that won’t warm up under the blanket. gone by day 5. not every cycle but often enough that I’m now expecting it. completely separate from the GI stuff (nausea cleared around week 6) and separate from the food noise drop, which lands closer to days 5-6 for me. the timing is what made me start tracking it. a ~5-day half-life means Cmax for the dose isn’t at 48h flat, it’s spread across days 2-3, which is also where the cold thing peaks. that’s a soft correlation, not a mechanism, I’m not pretending I know the pharm beyond what my NP explained. but the curve shape lines up enough that I don’t think it’s random. what I’ve ruled out, or at least made less likely:
not a thyroid drift, TSH was fine at my week 12 labs
not iron, ferritin came up actually (60 to 84) over the same window
not just “I’m eating less so I’m colder,” because my food intake is lowest on days 5-6 not 2-3, and the cold lands before the appetite drop
not cycle-phase, I tracked against my cycle for two months and it doesn’t map the thing I’m holding loosely: peripheral vasoconstriction would fit the white fingertips, and there’s some scattered chatter about GLP-1s affecting peripheral blood flow, but I haven’t seen anything I’d call a real mechanism for it. could also just be that rapid fat loss in someone who started with extra insulation runs colder, and I’m pattern-matching a confounder. flagging where my window doesn’t transfer: this is compounded sema, no polysorbate 80, so if you’re on a pen or on tirz the excipient and the PK curve are both different and your day-2-3 might not look like mine. also worth naming, I started genuinely metabolically stressed (fasting insulin 14, hsCRP 4.2), so I’m not sure if a less stressed baseline would even show this. the question: anyone else running cold in the Cmax window? does it line up with days 2-3 for you, or somewhere else in your cycle? curious whether this is a sema-specific thing or shows up on tirz too.
the “not every cycle” detail is actually the most interesting part to me - if it tracks against where you are in a vial’s open life, that’s a stability variable, not a PK one. oxidation affects potency, potency shifts effective Cmax, and if your effective Cmax is lower on a later-in-vial dose, that could easily explain why the peripheral response doesn’t show up consistently.
the white-tipped fingers detail is doing more work than the general cold feeling - that’s a vasospasm pattern, not just systemic heat loss. GLP-1 receptors are expressed in vascular endothelium and smooth muscle, so a Cmax-driven peripheral vascular effect isn’t mechanistically far-fetched. the cardiology lit on these compounds spends most of its attention on cardiac outcomes, but there’s scattered signal on peripheral vascular tone that doesn’t make it into most patient-facing discussions. the timing stagger you described is actually useful data in a way you didn’t fully name - if the cold and the food noise both peaked at days 2-3, you couldn’t separate central from peripheral effects. the fact that they’re offset by 2-3 days at least suggests they’re not the same mechanism running on the same timeline. that’s a real observation, not just pattern-matching. i can’t tell you what it means for your dose or your situation, that’s a prescriber question. but the framing is cleaner than most of what shows up in these threads.
the in-vial position angle is one I hadn’t mapped against my log, and it’s testable on my end because I date each vial and track dose number within it. I’ll pull that before my next appt. one thing I’d hold loosely with the stability framing though, compounded sema without polysorbate 80 has a different degradation curve than the pen formulations most of the oxidation chatter is built on, so the magnitude of potency drift across a 28-day vial might be smaller than the pen lit would predict.
doesn’t kill the hypothesis. also worth naming, even if late-vial doses are mildly less potent, I’d expect the cold thing to dampen on those cycles rather than disappear, which isn’t quite what I’m seeing. the on/off shape feels less like a smooth potency gradient and more like a threshold being crossed or not.
the polysorbate 80 absence cutting both ways is the part I’d push back on. that surfactant protects against aggregation and interfacial stress, so without it you’re not getting inherently slower degradation, you’re shifting which pathway dominates and leaning harder on cold chain and BAC quality to compensate. the threshold framing is interesting but it doesn’t actually rule out a potency gradient. steep dose-response curves produce on/off patterns, not smooth fade-outs. if this effect sits near a minimum effective concentration, even a 10-15% potency drift across vial life could flip you across it cleanly rather than showing up as gradual dampening.
The “white-tipped fingers when I’m typing” detail is worth bringing to your NP exactly as you described it, bc that’s specific enough to be useful rather than just “I run cold sometimes.” I’ve heard similar from a few people on compounded sema, temperature sensitivity in that first few days post-injection window, though I can’t point you to anything citable on mechanism. That’s anecdotal from a pharmacy tech, take it for what it is. Your differential is thorough. Ruling out thyroid, iron, food timing, and cycle phase before landing on a soft PK correlation is the right sequence. The vasoconstriction framing fits the fingertip presentation logically, though whether that’s the actual mechanism is a doc conversation. What I’d say is: the fact that you’ve got injection days and symptom days timestamped and separated is exactly what makes this legible to a clinician. I use CareClinic.io for exactly this, the free-text note you can attach to each logged item is what converts “I think this happens around day 2-3” into something a prescriber can actually look at and take seriously. You’ve done the work here. Bring it to your NP with the timestamps intact
the hypothalamus angle is worth putting on the table.
GLP-1 receptors are expressed in hypothalamic nuclei that regulate thermogenesis, not just satiety, so the days 2-3 peak could be your central thermostat responding to Cmax directly rather than a peripheral vascular effect. vasoconstriction and central thermogenic suppression both produce cold fingers - the symptom looks identical from the outside but the mechanism paths are completely different. imo “food intake is lowest on days 5-6 not 2-3” actually fits the central hypothesis better, bc if the hypothalamic signal handles both thermogenesis and satiety suppression, satiety being downstream and later is exactly what you’d expect on a staggered timeline. that kinda offset pattern is the thing that showed up in the CareClinic correlation view when I was logging my own injection days against side effect timing on sema - hard to call it noise when it repeats that cleanly across cycles.
leptin doesn’t get mentioned in these threads and it might be doing real work here. as fat mass drops, leptin falls, and falling leptin signals the hypothalamus to lower thermal set point - that’s mechanistically distinct from the “eating less = colder” thing you already ruled out, bc it’s the fat tissue itself broadcasting a signal, not caloric intake. whether Cmax timing amplifies that via hypothalamic GLP-1R activation is speculative on my part, but GLP-1 receptors are expressed in the hypothalamus and peak concentration logically hits them harder. the “not every cycle” inconsistency might actually support a leptin interaction over pure PK - leptin response to fat loss isn’t linear week to week. ymmv.