Metformin vs. GLP-1 Receptor Agonists for Antipsychotic-Induced Weight Gain
Real weight gain on olanzapine has a well-established, cheaper answer and a newer, more effective one that his insurance won't cover without a fight. The tradeoff is genuinely between evidence and access.
E.G., a 27-year-old man, works as a line cook, a physically demanding job he's proud of holding steadily for the past eighteen months after years of instability made keeping any job difficult. He has been stable on olanzapine for two years following a difficult first episode that required two hospitalizations before this regimen finally controlled his symptoms. Over those two years he has gained thirty-one pounds, moving his BMI from normal into the overweight range, and he tells his psychiatrist he's frustrated and a little embarrassed by the change, though he's firm that he doesn't want to touch the medication that's kept him out of the hospital.
He mentions, when asked, that he's started avoiding photos with coworkers and has stopped playing in the weekend pickup basketball games he used to enjoy, saying he doesn't feel like himself in his body anymore even though his mind has felt clearer than it has in years.
Metformin has the longest track record specifically for antipsychotic-associated weight gain, with a substantial body of trial evidence supporting modest but real benefit, is inexpensive, and is on essentially every insurance formulary without difficulty. GLP-1 receptor agonists have shown larger weight-loss effects in this population in more recent trials, but E.G.'s commercial insurance plan requires prior authorization with documented metformin failure before it will cover one, a policy his psychiatrist has seen take weeks to satisfy even when ultimately successful. The clinical question and the coverage question point in different directions: the newer drug looks more effective, but the pathway to actually getting it starts with trying the older one first regardless of what the evidence suggests about relative efficacy.
Metformin first, or push for the GLP-1 directly
I'd start metformin now regardless of the insurance pathway question. It has the longest track record specifically for this indication — real if modest weight benefit, well established, low-risk — and starting it isn't wasted time even if we end up adding a GLP-1 agonist later.
"Real if modest" undersells the gap, though. The comparative numbers aren't close: de Silva and colleagues' 2016 meta-analysis puts metformin's own benefit around three kilograms over three to six months, while Siskind and colleagues' 2019 individual-participant-data meta-analysis of older-generation GLP-1 agonists, specifically in patients on olanzapine or clozapine, shows closer to four and a half kilograms in about half that time — and newer agents in this same population have shown considerably more than that. He's already been through two medication transitions that visibly hurt his confidence in his own body. Starting him on the option we already expect to work less well, purely because it's administratively easier, risks the same discouraging pattern again.
I don't think either of you is actually wrong on the numbers, and I don't think we have to choose between them. Three kilograms starting today beats four and a half kilograms starting in three or four weeks once authorization clears, especially given how discouraged he already sounds — but that's not an argument against the GLP-1, it's an argument against making him wait on it. I'll file the prior authorization today, in parallel rather than sequentially, so metformin functions as a genuine bridge with its own real value, not a consolation prize he's stuck defending the adequacy of for a month.
Agreed: start metformin today, submit GLP-1 prior authorization in parallel rather than sequentially, and reassess weight and metabolic labs at three months.
E.G. was told plainly that starting with metformin reflects access and required documentation as much as it reflects a clean clinical preference — a distinction the pharmacist felt mattered to state directly rather than let him assume metformin was simply judged the better option.