Ask a clinician what to measure, rather than buying a multivitamin. A structured review covered six studies and 480,825 adults on these drugs[1]. Vitamin D deficiency ran 7.5% at six months and 13.6% at twelve. More than 60% were eating below the estimated requirement for calcium and for iron. Vitamin D intake averaged about a fifth of the recommended amount. Ferritin ran 26% to 30% lower than in a comparison group on a different diabetes drug. The data are observational, and the review says causality cannot be established.
These drugs work by making you eat less. Eating less means taking in less of everything, not only calories. That is obvious once stated and almost nobody states it.
What the review found
The search covered PubMed and the Cochrane Database from January 2019 to May 2025. It turned up six adult studies with nutritional endpoints, together covering 480,825 people. Vitamin D deficiency was the most common abnormality, in 7.5% of users at six months and 13.6% at twelve. Iron depletion was frequent. Ferritin ran 26% to 30% lower than in a comparison group taking SGLT2 inhibitors. That is a different diabetes drug, not an untreated control.
Intake was worse than status. More than 60% of users were eating below estimated requirements for calcium and for iron. Vitamin D intake averaged about a fifth of the recommended amount. Thiamine and B12 deficits grew over time, and protein and calcium shortfalls contributed to loss of lean mass.
How much less is being eaten
That simpler explanation has been measured. A meta-analysis pooled standardized test-meal data from four arms across three trials, 209 participants in total [3]. Energy intake at an observed lunch fell 1,132 kJ against placebo, on an interval from −1,449 to −815. That is roughly 271 kcal at one meal.
One lunch is not a diet, and the authors say habitual intake remains underreported. It is still the number that makes the nutrient findings unsurprising. The full reading is in how much less you eat on a GLP-1.
Why it goes unnoticed
The early symptoms are the symptoms of everything. Tiredness. Feeling cold. Thinning hair. Getting out of breath on stairs. A person who has lost twenty pounds attributes all of that to the diet, and so does everybody around them.
Nausea makes it worse by narrowing what people can face eating, which tends to mean less meat, less dairy and fewer vegetables. That is a predictable consequence of the titration period and nobody schedules a blood test around it.
The rare version, and its warning sign
Thiamine is stored for a matter of weeks. A systematic review found six published cases of Wernicke’s encephalopathy after semaglutide prescribed for obesity [2]. Every case followed the same sequence: prolonged gastrointestinal symptoms first, then substantial rapid weight loss, then neurological deterioration.
Six cases cannot produce a rate. Millions of people take these drugs, and case reports have no denominator. What the sequence gives is an early warning that the person experiencing it can see. Persistent vomiting for weeks is a reason to contact a prescriber rather than push through. The outcomes in several of those cases were poor. The detail is in the six reported cases.
What lean mass has to do with it
Protein shortfalls and lean mass are linked in the review, and the body composition evidence is thinner than the argument around it. The best-controlled figure comes from 101 adults over 24 weeks. Semaglutide changed lean body mass by −2.5 kg against placebo, on an interval from −6.6 to 1.6 [4]. That interval includes no change at all.
Lean body mass is also not muscle. Bioimpedance measures fat-free tissue, which includes water and organ mass, and the same analysis recorded extracellular water down 0.9 L. Nothing there supports a claim that these drugs spare muscle or destroy it. The state of that question is in what these drugs do to muscle and what comes back afterward.
What would actually catch it
A blood test. Vitamin D, ferritin, B12 and a basic metabolic panel cover most of what the review flags, and they are cheap and ordinary. The review’s own recommendation is targeted nutritional assessment and individualized laboratory evaluation for people at higher risk of malnutrition. That is a clinician deciding what to measure, rather than a reader buying supplements on the strength of a percentage.
Who is at higher risk is not mysterious. People losing weight fast. People who were already eating poorly. People with a restricted diet, and people who have had bariatric surgery. That last group is already monitored for exactly this. The medical system has a protocol for rapid weight loss and nutrition, and it exists on the surgical side of the market.
What sellers say about it
Close to nothing. A prescription arrives, a titration schedule arrives, and no seller on this roster publishes a monitoring plan or suggests a baseline panel. That absence is not unique to nutrition. It runs across almost every clinical question this market touches, and the census counts how wide the silence runs.
There is no accusation in that. A telehealth seller is not set up to run longitudinal labs, and most of them do not pretend to be. The point is that the gap is yours to close. You are buying a year of a drug that reduces how much you eat. Ask a clinician what to measure and when. Ask before the tiredness starts and gets attributed to the diet.