Nine-forty at the freezer door
You already brushed your teeth. That's the detail you'd never say out loud, because it makes it sound worse than it is, or exactly as bad as it is. You brushed your teeth, which was supposed to be the ending, and forty minutes later you are standing in the cold light of the open freezer with the spoon already in your hand.
Dinner was fine. You ate a real dinner. You weren't even hungry in the way that word is supposed to mean — no stomach growl, no lightheadedness. It's more like a channel that won't turn off. You thought about the ice cream while you were loading the dishwasher. You thought about it during the show. You negotiated: a square of chocolate instead, a handful of chips instead, and then the negotiation ended the way it usually does.
There was also the bag of chips at your desk at three, grazed without ever deciding to.
And in the morning you will do the accounting, the way you've done it for years. You will decide that the difference between you and the women who don't do this is that they have something you don't. Discipline, or self-respect, or whatever you're calling it this month.
That accounting is the thing I want to interrupt, because it isn't what the research shows at all.
A third of women with PCOS have disordered eating patterns on formal screening, and the odds of an abnormal eating-disorder score are roughly three times higher than in women without it. Not because three times as many women with PCOS lack character. Because something upstream of character is different, and it has been measured.
What you were actually up against
Appetite is not a verdict on who you are. It's a signal, and like every other signal in this condition it arrives from somewhere you didn't choose.
Here is the part that reframes years of self-blame: you are not failing to hear the signal that tells you to stop. In women with PCOS, that signal is quieter — and researchers have watched it be quieter, in blood draws, after identical meals.
It isn't permanent either. The signals that got turned down can be turned back up, and two of the drugs that do it are ones you can ask about by name at an ordinary appointment.
But first the mechanism, because the mechanism is what makes any of this believable.
Two ends of the same curve
The background is the same machinery I wrote about in the post on abdominal weight gain: cells respond sluggishly to insulin, the pancreas compensates by making more of it, and chronically high insulin raises ovarian androgen production while suppressing the SHBG that would otherwise keep testosterone bound. That's the engine. Here is what it does to hunger specifically.
After a meal, your body is supposed to run a shutdown sequence. Ghrelin — the hormone that says find food — drops. CCK rises from the gut and tells the brain the meal happened. Leptin, from your fat tissue, reports that reserves are adequate.
In PCOS, all three readings are off. Women with PCOS show a smaller drop in post-meal ghrelin than weight-matched women without it — the find food signal doesn't switch off as far. Post-meal CCK is blunted, and the degree of blunting tracks with testosterone levels. And leptin is typically high but the brain responds poorly to it, so the reserves report arrives and nothing reads it.
Three separate "you can stop now" messages, all sent quietly. You were not refusing to hear them.
Now the afternoon. You ate lunch, something with fast carbohydrate in it, and an insulin-resistant pancreas answered with a large amount of insulin. Insulin stays elevated in the circulation after your stomach has already emptied, and blood glucose falls further than it should. In one line of research, a third of women with PCOS dropped into reactive hypoglycemia after simple carbohydrate — and among lean women with PCOS, one study found the rate was about half. Peak glucose ran high at one hour; the nadir hours later ran near 61 mg/dL.
That nadir is the two o'clock feeling. Weakness, shakiness, sweating, palpitations, the particular desperation of needing to put your head down, and hunger that arrives out of proportion to the meal you just ate.
Two more findings turn this from unpleasant into important. First, in that same work the hypoglycemic symptoms coincided with a surge of cortisol and adrenal androgens — so the crash isn't only miserable, it's pouring fuel on the hormonal fire. Second, after simple carbohydrate, ghrelin started climbing again about two hours in, while after protein it stayed suppressed across five hours. Not a diet instruction. A fact about the shape of the curve.
So the three o'clock chips and the nine-forty freezer aren't two failures. They're the downslope of the same curve, and by evening your body has spent the whole day being told to find food.
Willpower does not change a signal. It can override one, for a while, at a cost — and that override is precisely what you've been doing for years and calling the bare minimum.
What to ask for
Ask for measurements first. A HbA1c or, better, an oral glucose tolerance test, which the 2023 international guideline names as the first-choice test for glucose status in PCOS. Describe the post-meal crash specifically, including the timing — two hours, three hours — because the timing is the clinical clue and it's the detail most likely to go unmentioned.
Then the two drugs.
Metformin lowers the glucose your liver releases and improves how your tissues respond to insulin, which flattens the spike that the crash follows. Its effect on appetite itself is real but modest, and the mechanism is genuinely unsettled — there's good evidence it raises GDF15, a hormone acting on appetite circuits in the hindbrain, and other work arguing that pathway isn't required. It's off-label for PCOS, meaning approved for type 2 diabetes and used here on guideline recommendation rather than FDA indication, which is ordinary rather than alarming. Cost of entry: GI upset, worst early and worst on immediate-release, often better on extended-release, with B12 worth monitoring long term.
GLP-1 receptor agonists act directly on the system this post is about. They slow gastric emptying and act on GLP-1 receptors in the hypothalamus, and in trials they reduce hunger and food preoccupation and improve the sense of fullness. Researchers now study this under the name food noise — intrusive, unwanted thoughts about food — with validated questionnaires, which means the thing you couldn't describe at an appointment has a literature and a measuring instrument.
The honest limits: in PCOS specifically, pooled trials show modest weight reduction and improved insulin resistance but no demonstrated effect on testosterone, and the reproductive data are thin. Side effects are nausea, vomiting and constipation. There are insurance fights, muscle loss without adequate protein and resistance training, and weight regain for many people who stop. They are not for use in pregnancy and need a washout before trying to conceive.
Bring the sentence:
I get an intense crash about two hours after eating, and cravings at night that I can't override. That pattern makes me think insulin resistance. Can we check a HbA1c or a glucose tolerance test, and talk about whether metformin or a GLP-1 makes sense for me?
Two doors, and then it's yours
The first door is your PCP, OB-GYN, or endocrinologist. Bring the timings. Ask for the labs.
One thing to hold onto while you go. Given how often disordered eating travels with this condition, there's a real chance you'll be offered a food diary and a calorie target by someone who never checks a glucose curve. If you've been bingeing, that deserves actual care — not a stricter plan handed to someone whose satiety signaling is already impaired. A clinician who treats this as a discipline problem has told you about their training, not about you.
The second door is telehealth built for this. Allara Health is a virtual practice for PCOS and women's hormonal and metabolic health that pairs clinicians with registered dietitians and works with many major insurance plans, though state availability varies. Sesame is a direct-pay marketplace where you book a named clinician at a listed price without insurance or a referral, useful if you're uninsured or facing a long wait.
I want to be clear that I have no financial relationship with either of these. I'm not an affiliate and I don't get paid if you book. I mention them because "find a doctor who actually knows PCOS" is useless advice without somewhere to start.
PCOS is managed, not cured, and I won't promise you a number or a date. But the quiet at nine-forty that you've been trying to produce with resolve is something many women only got when the signal changed, and the signal is treatable. You have been doing this on hard mode, without being told it was hard mode, and doing it anyway. That was never a character defect. It was an untreated one.
This is my experience and my reading of the research, not medical advice. Nothing here should replace a conversation with a clinician who knows your history.
Sources
- Food Cravings and Obesity in Women with Polycystic Ovary Syndrome: Pathophysiological and Therapeutic Considerations — Stefanaki et al., Nutrients, 2024
- Reactive Hypoglycemia: A Trigger for Nutrient-Induced Endocrine and Metabolic Responses in Polycystic Ovary Syndrome — Karakas, Journal of Clinical Medicine, 2023
- Increased odds of disordered eating in polycystic ovary syndrome: a systematic review and meta-analysis — Lee et al., Eating and Weight Disorders, 2019
- Recommendations from the 2023 International Evidence-based Guideline for the Assessment and Management of PCOS — Teede et al., Journal of Clinical Endocrinology and Metabolism, 2023
- Metformin-induced increases in GDF15 are important for suppressing appetite and promoting weight loss — Coll et al., Nature Metabolism, 2020
- Food noise: definition, measurement, and future research directions — Nutrition & Diabetes, 2025