The insurance catch-22 with tirzepatide is worth approaching from the trial data side, because the concern is real. SURMOUNT-4 is the clearest reference: participants who discontinued after a year saw substantial weight regain within 52 weeks, and metabolic markers including fasting glucose moved back toward baseline. Not immediately, not uniformly - but the trend was there. For pre-diabetic or newly-normalised patients, this is the specific problem. Drug works, A1C drops into normal range, insurer reclassifies you as no longer qualifying, coverage lapses. How durable is the effect, actually? My working hypothesis: appetite suppression is the scaffolding maintaining the metabolic improvement. Remove it before insulin sensitivity is genuinely stable and you’re borrowing time. I track fasting glucose weekly; the monthly trend summary in CareClinic is what I’d bring to my GP if coverage ever lapsed - a trajectory, not a snapshot. The literature on maintenance dosing for metabolic indication specifically is thin. That’s the conversation worth having here.
“appetite suppression is the scaffolding” is probably the right model, but it implies there’s a point where the scaffolding can come down once the structure is load-bearing on its own. fwiw SURMOUNT-4 doesn’t show that window exists within 12 months. the part that gets undersold in these conversations is the pre-diabetic vs established T2D distinction - for someone who normalized from 8.4 to 5.6 like me, describing that A1c as “stable” vs “managed” is doing a lot of work. my endo’s framing: tirz isn’t treating a deficiency, it’s compensating for a chronic dysregulation. discontinuing without an alternative compensation mechanism is just hoping the underlying pathology resolved on its own, which is a bet the current data don’t really support. the “maintenance dosing for metabolic indication” literature gap you’re pointing to is real, but I’d want to see the SURMOUNT-4 cohort tracked out to 3 years before drawing strong conclusions about where the rebound actually plateaus
“treating a deficiency vs compensating for chronic dysregulation” is a useful clinical frame, and your endo’s read is probably right for the established T2D case. where I’d push back slightly: pre-diabetic pathology, especially someone who came in at 6.1 with significant adiposity, may not be the same category of irreversibility as someone who normalized from 8.4. beta cell dysfunction and insulin resistance driven primarily by visceral fat are different problems mechanistically, and the latter has at least some evidence of durable improvement with sustained weight loss even independent of continued GLP-1 agonism. I’m not arguing the scaffolding definitely comes down at 12 months. I’m saying the pre-diabetic subgroup may have a different answer than the T2D-who-normalized cohort, and SURMOUNT-4 doesn’t cleanly separate them. the 3-year data ask is right, but I’d want it stratified by baseline A1c and by how much weight was maintained, not just reported as a single cohort. the heterogeneity within “people who normalized on tirz” is probably pretty large.
The pre-diabetic subgroup distinction is worth pressing on, and your point about mechanistic heterogeneity is solid. Where I’d add a caveat though: “durable improvement with sustained weight loss even independent of continued GLP-1 agonism” is doing a lot of work there. The evidence for that is mostly observational and confounded by who actually sustains the weight loss. In my experience, the people maintaining without the drug are a select group to begin with. The SURMOUNT-4 stratification ask is right, and I’d add baseline insulin secretory capacity to the list alongside A1c and weight maintained. That variable is probably carrying a lot of the heterogeneity you’re describing.
the “appetite suppression is the scaffolding” framing is probably underselling the direct receptor activity here - GIP agonism specifically has hepatic and pancreatic effects that aren’t purely downstream of eating less. so the rebound when coverage lapses may hit faster than you’d expect from equivalent caloric restriction ending, which is actually a grimmer story than the scaffolding metaphor implies. thin maintenance dosing literature for T2D indication specifically is the real gap you’ve named, and i don’t think anyone has good data on what minimum effective dose looks like long-term.
“scaffolding maintaining the metabolic improvement” is the right frame and I’d keep it, but I think it undersells how much of the SURMOUNT-4 regain curve is just energy balance reasserting itself once the satiety signal goes away. The fasting glucose drift back toward baseline tracks weight regain pretty tightly in the published curves, which is consistent with “insulin sensitivity improved because adipose mass dropped” rather than anything durable happening at the receptor level independent of weight. So the scaffolding metaphor is right, I’d just argue the scaffolding is holding up the weight loss, and the metabolic stuff is downstream of that. The piece I’d push on is the “borrowing time” framing, because it implies there’s a finish line where insulin sensitivity becomes self-sustaining off-drug. I haven’t seen anything in the trial data that suggests that line exists for most people. SURMOUNT-4’s withdrawal arm wasn’t designed to find a durability threshold, it just showed regain at 52 weeks post-discontinuation. The maintenance dosing question you flagged is the actual unknown, and the reason it’s thin in the literature is that the trials weren’t powered or designed to answer it. Lilly has zero commercial incentive to study a lower maintenance dose, fwiw. The weekly fasting glucose tracking is a reasonable proxy but worth flagging the intra-individual CV on a single fasting draw is genuinely high, like 5-8% on a good day, more if sleep or stress varies. A weekly point can swing 10+ mg/dL on noise alone. The trend summary view is useful precisely because it smooths that, and a 12-week trajectory tells you something a single value can’t. If coverage ever does lapse, that smoothed trajectory is the one document I’d want in front of a GP, agreed.