Protein, Iron and B12: Nutrient Status During Rapid Weight Loss on a GLP-1

The short version, and it is not comfortable

Nearly every protein target, supplement list and blood-test panel you will read about GLP-1s is extrapolated from bariatric surgery rather than measured on people taking these drugs. That is not a rhetorical flourish. A systematic review published in 2026 went through 43 articles from 41 randomised trials of liraglutide, semaglutide or tirzepatide, covering 50,690 participants, and asked a simple question: did any of them assess what people were eating? Two did. Two out of 41 — about 5% — reported or evaluated dietary intake at all, and the quality of the assessment methods used in those two was rated "poor" and "acceptable"¹.

So the field ran trials that cut body weight by roughly a seventh and, with two exceptions, never recorded what the participants ate. The mechanisms are real and the concern is legitimate — but the numbers quoted at you are borrowed.

Why the concern is reasonable anyway

Start with the magnitude. In STEP 1, 1,961 adults with overweight or obesity were randomised 2:1 to semaglutide 2.4 mg or placebo for 68 weeks. Mean weight change was −14.9% on semaglutide versus −2.4% on placebo. The much-quoted responder rates sit on a denominator of 1,212 of the 1,306 randomised to semaglutide: 1,047 (86.4%) lost at least 5%, 838 (69.1%) at least 10% and 612 (50.5%) at least 15% — against 31.5%, 12.0% and 4.9% on placebo⁸. Losing a seventh of your body weight over a year requires eating substantially less for a year, and nutrient intake scales with food volume.

Then add the specific mechanisms. A 2026 clinical review of micronutrient risk during incretin therapy describes the vulnerability as arising from the interaction of reduced food intake, lower dietary diversity, gastrointestinal intolerance, delayed gastric emptying, rapid weight loss and baseline nutritional risk. The domains it flags as carrying the most relevant signals are haematologic, fat-soluble, bone-related, trace element and electrolyte — naming iron, vitamin B12, vitamin D, calcium, magnesium and zinc, with thiamine, folate, vitamin A and potassium as context-dependent concerns. Its key sentence for anyone deciding whether to panic: most abnormalities reported to date are subclinical or indirect, though clinically meaningful consequences may occur in susceptible people — including those with prior bariatric surgery, gastrointestinal disorders, poor baseline diet quality, older age, or prolonged nausea and vomiting³.

A narrative review in Advances in Nutrition frames the same problem from the diet-quality side: these drugs suppress energy intake effectively, but they may also have unintended effects on macronutrient distribution, ultraprocessed food consumption, and the risk of vitamin and mineral deficiencies, alongside loss of lean body mass⁴. The lean-mass half of that sentence is covered properly on our pages about muscle loss on a GLP-1 and bone density and lean mass; this page is about what is in the food.

The one real-world dietary study, and its limits

There is one published cross-sectional comparison of what GLP-1 users actually eat, and I want to be honest about how far it travels. In Turin, 103 adults with type 2 diabetes attending a diabetes clinic between February and June 2025 completed a validated food-frequency questionnaire; 52 were on a GLP-1 receptor agonist and 51 on other oral agents. There were no significant differences in energy intake, macronutrient distribution or Mediterranean diet score between the groups, and no differences by treatment duration either. What the study did find was that everyone ate poorly: fibre intake around 11 g per 1,000 kcal, fat at 39–40% of energy, and not one GLP-1 user meeting fibre recommendations².

The participants had a mean age of 66, were 65% male, and all had type 2 diabetes. That is close to the opposite of this site's reader. Take from it the one thing it supports well — poor diet quality is a baseline problem, not something the drug creates — and nothing else.

What the borrowed guidance actually says

Two 2026 reviews converge on a similar framework, and both are explicit that it is a proposal, not a validated protocol. The more specific one recommends daily protein of at least 1.2 g/kg, up to 1.6 g/kg in adults without chronic kidney disease, distributed as roughly 0.3–0.4 g/kg per meal with about 2.5–3 g of leucine; a laboratory panel of vitamin D, B12, iron studies, folate and zinc, adding thiamine in high-risk patients; energy floors high enough to keep micronutrient intake plausible; and progressive resistance training alongside⁵.

Where does that panel come from? Largely from the surgical literature. The American Society for Metabolic and Bariatric Surgery's micronutrient guidelines — 92 graded recommendations covering vitamins B1 and B12, folate, iron, vitamins A, E and K, calcium, vitamin D, copper and zinc — note that it is common for patients presenting for weight-loss surgery to already have at least one vitamin or mineral deficiency before any operation, that preoperative screening has not been the norm in most practices, and that the prevalence of deficiencies continues to rise while follow-up monitoring declines⁶. That is the document underneath most GLP-1 supplement advice. It was written about surgically rerouted anatomy, which a weekly injection does not create.

The iron point deserves its own paragraph

Of everything on this page, iron is where a mother's starting position matters most, and where the pre-existing evidence is strongest even though none of it is about GLP-1s. In 1,098 UK women aged 18 to 49, iron deficiency was present in 49.7% using a ferritin threshold of 30 μg/L and 19.6% using 15 μg/L; anaemia affected 9.2% and iron-deficiency anaemia 6.8%. Women with greater central adiposity had markedly worse iron status — iron-deficiency anaemia in 12.1% of those with a high waist-to-hip ratio versus 4.9% of those with a low one. The authors' conclusion is that at least one in five women of reproductive age in the UK is iron deficient⁷.

If you menstruate, have been pregnant recently, and are now eating perhaps a third less food, the arithmetic is unkind before any drug-specific mechanism is invoked. Fatigue and hair thinning on a GLP-1 are usually attributed to the weight loss itself — and our page on hair loss on a GLP-1 explains why that attribution is often right — but a ferritin and a full blood count are cheap, and worth asking for rather than guessing at. Persistent exhaustion, breathlessness on stairs you used to manage, pins and needles, or a sore tongue are reasons to be seen and tested, not reasons to start buying supplements online.

What to ask for

Ask for a baseline before you start, not only when something goes wrong: ferritin, full blood count, B12, folate and vitamin D at minimum. Ask what protein target they want you at in grams, for your body weight, and how you are meant to reach it while eating less. Ask when the bloods get repeated. If nausea or vomiting persists for weeks rather than days, say so — prolonged vomiting is the one scenario in which the reviews stop talking about subclinical signals and start talking about real deficiency³. And if you take a multivitamin because someone online told you to, tell your clinician which one; a supplement is only useful against a deficiency you actually have.

Where this sits on the evidence scale

On what happens to nutrient intake: almost nothing has been measured, and that is the finding. Two of 41 randomised trials — 5% — reported dietary intake at all, and both were rated poor or merely acceptable on method. One cross-sectional study of 103 people, mean age 66 and 65% male, is the real-world dietary evidence base. Anyone quoting you a confident figure for how much protein GLP-1 users eat is quoting a model, not a measurement.

On micronutrient deficiency: plausible mechanism, weak direct evidence, honest reviews. The 2026 clinical review that names iron, B12, vitamin D, calcium, magnesium and zinc as the priority domains also says the abnormalities reported so far are mostly subclinical or indirect, and calls for prospective studies to define incidence and clinical relevance. That is a research agenda, not a finding.

On the monitoring advice itself: borrowed, and openly so. The panel of tests and the protein targets come from bariatric surgery nutrition and general weight-loss nutrition, applied by analogy to a drug that does not reroute the gut. The analogy is reasonable — less food is less food — but it is an analogy, and the surgical guidelines were themselves written about a population where at least one deficiency was already common before surgery. Treat the panel as sensible precaution with a thin evidence base, and the iron question, for a mother of reproductive age, as the one with real numbers behind it.