Most adults who decide to lose weight start on their own, and for most of them that is a reasonable thing to do. A smaller group should not, and the reasons are specific rather than vague: a recent weight change nobody has explained, a history with food that restriction reliably reopens, a prescription that becomes too strong the moment intake falls.

This piece covers the situations in which running an energy deficit without clinical oversight carries a defined and documented risk, what the guidelines actually say about where that line sits, and the honest answer about how much of it has been written down at all.

• No guideline body has published a universal supervision threshold. The case rests on specific clinical situations, not on a calorie count or BMI number.
• Unintentional weight loss of 5% or more over 6 to 12 months is a clinical signal to investigate, not a running start on a diet.
• Type 2 diabetes treated with insulin or sulfonylureas carries a real hypoglycaemia risk when carbohydrate intake falls. Dose adjustment is a prescriber's decision.
• An eating-disorder history makes dietary restraint itself a relapse trigger. Changes to intake belong inside structured, supervised care.
• Several common medicines, including levothyroxine, warfarin, and lithium, have doses tied to body weight. Significant weight loss changes drug exposure even when the prescription on paper has not changed.
• Supervised very-low-calorie programmes include safeguards (thiamine, electrolyte monitoring, cardiac monitoring) that a self-run crash diet does not.

Start with the absence, because it shapes everything after it. Read across the Endocrine Society of India, RSSDI, ICMR, NICE, the US Preventive Services Task Force, WHO and the Canadian Adult Obesity clinical practice guideline, and there is no sentence anywhere that names the point at which weight management stops being something a person may reasonably do alone.

Supervision is present throughout those documents, but it arrives implicitly: through drug pathways, through surgical pathways, through references to structured programmes and specialist services. No universal calorie figure and no BMI number is published that declares dieting unsafe without a clinician.

What exists instead is a set of calorie floors. In practice they do the work a threshold would have done:

• NICE (UK): Any diet under 600 kcal per day should be used only under clinical supervision. Very-low-calorie diets at or below 800 kcal per day are classified as short-term tools that must sit inside multicomponent, supervised programmes.
• Canadian adult obesity guidance: Flags intakes below 1,200 kcal per day (and around 900 kcal in particular) for physician monitoring, on the grounds that rapid loss at those levels is linked to gallstones and cardiovascular change.
• India's dietary guidelines: Weight-reduction diets should not fall below approximately 1,000 kcal per day. Weight loss should be gradual, and extreme approaches are named as a health hazard.

Put those together and a pattern emerges that is consistent even though nobody has codified it. Moderate deficits, roughly 500 to 600 kcal per day aiming at 0.5 to 1 kg per week, are treated everywhere as routine behavioural care. Low-energy regimens in the 800 to 1,200 kcal range, and very-low-calorie regimens below 800 to 900 kcal, are consistently tied to specialist, protocol-driven oversight. Between those two positions sits an unmarked zone that every reader has to navigate without a published rule.

This matters more than it looks. Because no threshold exists, the case for supervision does not rest on a calorie count. It rests on who you are and what else is going on in your body, which is what the rest of this piece is about.

Unintentional loss of more than about 5% of usual body weight over six to twelve months is treated as clinically significant in adult assessment pathways, and it is a standard trigger for investigation rather than encouragement. It commonly opens a work-up for malignancy, serious infection, endocrine disease and HIV.
In an Indian context the differential is worth naming plainly, because it is not the same as a Western one. Tuberculosis, malignancy, uncontrolled diabetes, thyrotoxicosis and HIV are all either prevalent or clinically important here, and all of them can present as weight coming off without effort.
The practical consequence is uncomfortable but simple. Weight that has already fallen without being asked to is not a running start on a diet. Adding deliberate restriction on top of it makes the underlying process harder to see and can worsen the nutritional position of someone who is already depleted. If weight has dropped in the last six to twelve months and nobody has explained why, the next step is an assessment by a registered medical practitioner, not a plan.

Relapse rates after treatment for anorexia and bulimia sit at roughly 30 to 50% over the first one to two years, and dietary restraint and reduced calorie intake are repeatedly identified in that literature as major relapse drivers. That is the specific problem: the thing a weight-loss plan asks a person to do is the same thing the relapse literature identifies as a trigger.

The screening instrument used most widely is the SCOFF questionnaire, five items, with a pooled sensitivity of about 84% and specificity of about 80% for adult eating disorders across 10 studies covering roughly 4,300 people, including validation in multi-ethnic samples. It is worth knowing that a short, well-validated screen exists. It is equally worth knowing that it is a clinician's instrument for opening a conversation, not a self-administered clearance test, and a score does not confirm or exclude anything on its own.

Where there is a past diagnosis, prior specialist treatment, or current disordered eating, changes to intake belong inside structured care that explicitly manages relapse risk, and not inside a commercial programme or a self-designed plan. The evidence has a real limit here that is worth stating: studies in this field use heterogeneous definitions of recovery and relapse, which is precisely why no clean numerical rule about who may and may not diet has ever been produced from them.

This is the population where the mechanism is easiest to state and the risk is most immediate. When carbohydrate intake falls, blood glucose falls with it. Insulin and sulfonylureas do not adjust to that: their action is glucose-independent, so a dose that was correct at the old intake becomes over-strong at the new one. Hypoglycaemia, not slow progress, is the acute hazard of self-directed dieting in type 2 diabetes treated with these agents.

Indian RSSDI guidance reflects this directly, highlighting strict monitoring and dose adjustment in people on insulin, sulfonylureas or glinides, and noting that hypoglycaemia risk is highest with exactly these classes.

The trial literature makes the point even more starkly, because of what the protocols had to build in. In low-energy formula-diet studies such as DiRECT and related total-diet-replacement randomised trials, glucose-lowering drugs are stopped or aggressively down-titrated at the moment the diet starts, with structured monitoring specifically designed to prevent hypoglycaemia.

Those trials produced their results with that scaffolding in place. It cannot be replicated at home, and the reader should not attempt any part of it independently: no dose is changed, stopped or skipped except by the prescriber who wrote it.

Three separate situations, governed by the same principle: nutritional requirements are already defined by something other than weight, and an energy deficit interacts with that definition rather than sitting beside it.

ICMR's dietary guidelines for Indians state that weight-reduction diets should not fall below about 1,000 kcal per day and that extreme approaches should be avoided, with specific gestational weight-gain targets set in pregnancy. Deliberate restriction during pregnancy or lactation is therefore an obstetric decision, made against those targets, not a personal one.

After bariatric surgery the constraint is different but no less firm. Bariatric nutrition guidance written for Indian patients requires compulsory pre-operative counselling and lifelong nutritional follow-up, specifically to detect complications including hypoglycaemia, dumping and micronutrient deficiencies. A person in that follow-up who then adds an unsupervised commercial plan is changing the inputs to a monitoring programme without telling the people running it.

In CKD, energy and protein are not lifestyle variables. Indian CKD guidance prescribes them by ideal body weight and disease stage: for non-dialysed CKD, roughly 30 to 35 kcal/kg/day with 0.6 to 0.8 g/kg/day of protein, with higher protein intakes for patients on dialysis.

Those figures come with an assumption attached, which is that nitrogen balance, albumin and electrolytes are being monitored while they are applied. Strip the monitoring away and unsupervised calorie or protein restriction in CKD carries a real risk of malnutrition and metabolic complication.

There is a second-order problem worth naming. A large share of popular weight-loss advice is high-protein advice, and high-protein is not a neutral default in kidney disease: it runs in the opposite direction to the prescribed intake. Any weight-loss attempt in CKD needs to be dietitian-planned inside nephrology care rather than layered on top of it.

Several chronic medicines have doses tied closely to body weight or lean mass, which means substantial weight loss changes drug exposure even though nothing about the prescription has changed on paper.

Levothyroxine

Levothyroxine is used as replacement treatment in primary hypothyroidism, and its dose is based on body weight. As body weight changes significantly, the amount of levothyroxine needed may also change. Endocrinology reviews note that a weight change of about 10% should prompt dose re-evaluation and TSH testing to check whether the existing dose is still appropriate.

Warfarin

Warfarin dosing is influenced by body size, with weekly dose requirements generally tracking BMI. Cohort data found that the weekly dose was roughly 0.69 mg higher for each one-point increase in BMI. This means that a meaningful reduction in BMI during weight loss can leave a previously appropriate dose too high, making dose review important as body weight changes.

Lithium, antiepileptics, and several psychotropics

Lithium, antiepileptics, and several psychotropic medicines have exposure that is sensitive to body weight and volume changes. However, the relationship between weight and dose is not well standardised for these medicines. Guidance therefore relies mainly on pharmacokinetic principles and expert consensus rather than large outcome trials.

For lithium and many antiepileptics, nobody has established the exact amount of weight change that should mandate a review, which is why the advice is conservative by default.

The practical rule that follows is unglamorous and effective. If you take a medicine in any of these categories, tell the prescriber before you deliberately lose more than about 5 to 10% of your body weight or start anything aggressive, not afterwards when a result comes back odd.

Thyroid replacement is the case most people encounter, and it has its own set of questions that go beyond dose arithmetic, covered in thyroid and weight. Medicines that cause weight gain in the first place are a separate problem with a separate answer, dealt with in medicines and weight gain.

Reviews of very-low-calorie diets conducted under medical supervision report few serious adverse events across 8 to 16 weeks, which is a genuine finding and should be reported as such. The same reviews note higher gallstone risk when dietary fat is insufficient during rapid loss.

The more serious hazard sits at the other end, when normal eating resumes. Refeeding guidance identifies people with very low BMI, recent unintentional loss of more than 10 to 15% over three to six months, or minimal intake for more than five to ten days as high risk for refeeding syndrome, which can produce hypophosphataemia, hypokalaemia, hypomagnesaemia and life-threatening arrhythmias.

Clinical very-low-calorie protocols therefore start at modest calorie levels, add prophylactic thiamine and electrolytes, and monitor biochemistry and cardiac rhythm closely in the first days. Every one of those safeguards is absent from a self-run crash diet, and their absence is the actual difference between the supervised evidence base and what happens at home.

One caveat belongs here and is often left out. Robust incidence figures for gallstones, arrhythmias and full refeeding syndrome specifically among unsupervised very-low-calorie dieters are sparse. Most data come from supervised programmes and case reports, so the true community risk is not known. That is an argument for caution rather than a reassurance.

There is no published clearance test, so what follows is not one. It is a description of what a registered medical practitioner looks for when deciding whether a self-managed deficit is reasonable for a particular person, drawn from the sources above.

• Whether there has been unintentional weight loss of 5% or more in the last six to twelve months, and whether it has been explained.
• Whether there is any history of an eating disorder, or current signs of disordered eating.
• Whether diabetes is being treated with insulin, sulfonylureas or glinides.
• Whether the person is pregnant, breastfeeding, post-bariatric, or living with chronic kidney disease.
• Whether any weight-sensitive, narrow-therapeutic-index medicine is in use without an agreed plan for reviewing the dose as weight changes.
• Whether the plan being considered stays clear of very-low-calorie territory below 800 to 900 kcal per day.

Where any of those is present, or where the plan under consideration is an 800 to 1,200 kcal total diet replacement or a sub-800 kcal programme, the sources point consistently towards clinical oversight rather than a self-managed attempt.

None of that reads on a single number, which is why the decision to start belongs at an assessment rather than at a checkout page. It is also a different question from what to do once a programme is already running and a measurement changes, which is covered separately in muscle loss during weight loss.

If you are weighing up a supervised route, the distinction between a programme with clinical oversight and a diet plan with a doctor's photograph on the homepage is worth understanding first, and what separates the two is the useful place to start.

Not sure which side of the line you are on? A Voy clinician can review your medications, weight history, and health picture together and tell you whether clinical oversight is recommended before you change your intake. Book a consultation.

This article is for general information and education only and is not medical advice, diagnosis, or treatment. GLP-1 and other medications referenced are prescription-only and are appropriate only for certain people under the supervision of a qualified clinician. Do not start, stop, or change any medication based on this article. Please consult a registered medical practitioner about your individual circumstances. Information reflects what was available at the time of review and may change.