Six months into a deficit, the scale stops moving, and the advice that follows is almost always the same: tighten up, cut a little more, be stricter. That advice assumes the plateau is a discipline problem. In the trial literature, it is the expected shape of the curve, arriving at roughly the point the evidence says it should. The clinically useful question is not how to break it but whether anything about this particular plateau justifies re-measuring rather than restricting. This article sets out when plateaus normally arrive, why the field genuinely disagrees about what causes them, which findings shift the answer, and why investigating a plateau on weak grounds is its own kind of harm.

• Diet trials consistently show maximum weight loss at 20 to 24 weeks, followed by a plateau. A six-month stall is what the evidence predicts, not a sign of failure.
• Whether a plateau is caused by metabolic adaptation, gradual intake drift, or both is genuinely unresolved in the literature. Both effects are real.
• A plateau and a regain are different events with different clinical implications. Duration alone does not determine which it is.
• The signal to investigate is a non-weight finding (fatigue, cycle changes, snoring, a new prescription), not the number of weeks the scale has been still.
• Over-investigating a normal plateau is itself a documented harm: false positives, health anxiety, and unnecessary cost.

The timing is consistent enough across studies to be treated as a baseline expectation. A systematic review of weight-loss interventions found mean losses of roughly 5 to 8.5 kg, or 5 to 9% of starting weight, in the first six months on a reduced-energy diet, followed by a plateau and partial regain, with 3 to 6 kg still held at 24 to 48 months. Behavioural weight-loss trials typically reach their maximum inside 20 to 24 weeks of intensive treatment. A meta-analysis of intermittent energy-restriction trials running past six months found the same pattern, with sustained losses of 3 to 10 kg at 12 months depending on protocol.

Outside trial conditions, the numbers are smaller and the shape is identical. A retrospective cohort of a real-world behavioural programme recorded maximum mean loss of about 3.9% at six months, falling to 3.2% at 12 months and 2.3% at 24, with about a third of participants reaching a 5% loss at any point.

A recent secondary analysis of a 12-month crossover trial in 42 adults tested whether the plateau was a property of the diet. Participants lost rapidly over the first three months, slowed between three and six, and plateaued at about six months regardless of whether they had been allocated low-fat or low-carbohydrate. Switching diets at the six-month mark produced only modest further loss to nine months and did not move the plateau.

Nor is plateau something a medical route avoids. Modelling of long-term obesity pharmacotherapy trials found that suppression of energy intake decays exponentially, producing a plateau within about a year despite continued treatment, typically between six and 12 months rather than at six. Plateau is a feature of the system, not evidence that a method has failed.

One caveat matters for an Indian reader. These plateau curves come from US and European trials. India's own prevention studies report different endpoints: the Indian Diabetes Prevention Programme enrolled Asian Indian adults with impaired glucose tolerance and reported diabetes incidence falling from roughly 28.5% to 10.6%, and the Diabetes Community Lifestyle Improvement Program in about 485 adults with prediabetes recorded a 186 kcal/day reduction in intake at six months and roughly halved the one-year hazard of diabetes.

Two bodies of authoritative work explain the same six-month plateau differently, and the contest between them is unresolved. It is worth understanding because the two explanations imply different responses.

The metabolic adaptation case

Adaptive thermogenesis is the fall in energy expenditure beyond what body-composition change predicts.

In metabolic ward experiments, 17 inpatients with obesity were studied after 10% and 20% weight loss. Both resting and non-resting expenditure fell significantly below predicted values. The key findings:

• Weight-clamping experiments put total daily expenditure roughly 150 to 250 kcal/day below prediction.
• A broader review places the effect at 50 to 140 kcal/day in moderate weight-loss contexts.
• In people maintaining a 10% or greater reduction for more than one year, the suppression of expenditure persisted.
• A controlled restriction study found that a 100 kcal/day reduction in expenditure detectable after one week predicted 2.0 kg less weight loss at six weeks.

The Biggest Loser cohort is the most-quoted illustration. Across 14 participants, resting metabolic rate fell about 789 kcal/day over the 30-week programme and remained about 500 kcal/day below prediction six years later. This is 14 people under extreme conditions of large deficit and very high activity. It should not be read across to ordinary Indian outpatient weight loss.

The intake drift case

Validated dynamic energy-balance models predict that under perfect adherence to a constant intake reduction, weight plateaus after one to two years. But trials plateau at six to eight months. That gap is the problem.

A computational model of human weight change reproduced the observed plateau only when adherence was allowed to decay. An 800 kcal/day deficit that progressively returned toward pre-intervention intake within the first year generated both the plateau and the subsequent regain. A separate study found that even implausibly large levels of metabolic adaptation failed to move plateau timing, while intermittent lapses in adherence produced plateaus at about 24 weeks that matched real trial data.

What the modellers conclude is this: adaptation mainly sets the plateau's level (the final weight reached), while adherence patterns set its timing (when the plateau arrives).

Both effects are real and present in every plateau. What is contested is attribution. No study has combined precise energy-expenditure measurement, detailed intake tracking, and long-term weight curves in a large outpatient cohort. Treat specific percentages that circulate for adaptive thermogenesis with more caution than the direction of the finding.

The two get discussed together and behave differently, and the distinction changes what a clinician does next.

A plateau is an equilibrium. Weight stabilises with small fluctuations and no clear upward trend, because intake has risen to meet the reduced expenditure of a weight-reduced body and the energy gap has closed from both directions. A regain is a sustained upward trend after that point, often back toward baseline.

The hormonal picture after weight loss explains why regain is the default rather than the exception. After roughly 10 to 14% loss, ghrelin rises while the satiety signals leptin, GLP-1 and PYY fall, and those changes have been shown to persist for up to a year. In regain cohorts, larger rises in ghrelin and greater reductions in GLP-1 predicted more regain over the following 12 to 24 months. The same physiology is why weight returns after stopping a medication that was suppressing intake.

Telling them apart requires serial weights over six to 12 months, not a single reading:

• Plateau at six to nine months with 5 to 10% total loss: Usually read as expected physiology plus adherence reality. Prompts a review of the programme, not more tests.
• Sustained upward trend with new symptoms: What prompts evaluation for sleep apnoea, low mood, medication change, or an endocrine cause.

None of the conditions below is diagnosed by a plateau. What makes one worth testing is a finding that has nothing to do with weight: a symptom, a new prescription, or a pattern in the body that does not follow from the diet. Prevalence is included because it is the column most often left out of these conversations.

Hypothyroidism: Hypothyroidism was found in 10.95% of 5,360 adults across eight Indian cities, including 3.47% who were previously undiagnosed. Signs that are not the weight include fatigue, cold intolerance, dry skin, hair loss, constipation, menstrual irregularity, slow pulse, and unexplained high lipids. A clinician may order TSH and free T4, with anti-TPO antibodies in selected patients.

Common in India (roughly one in ten adults), but the weight directly attributable to it is usually modest and largely fluid. Correcting it does not reliably produce weight loss. It is worth finding for its own sake.

Obstructive sleep apnoea: Pooled prevalence was around 11% across eight Indian studies, ranging from 3.7% to 21%, with about 5% having moderate to severe disease. Signs that are not the weight include loud snoring, witnessed pauses in breathing, waking choking, unrefreshing sleep, morning headache, daytime sleepiness, and blood pressure that resists treatment. A clinician may order overnight polysomnography or a home sleep apnoea test.

PCOS: Pooled Indian prevalence was around 11.3%, with a 95% confidence interval of 7.7% to 15.6% and individual studies ranging from 4.2% to 22.5%. Signs that are not the weight include long or irregular cycles, hirsutism, acne, acanthosis nigricans, and difficulty conceiving. A clinician may order testosterone and DHEAS, LH, FSH, prolactin and TSH, an OGTT and lipids, and pelvic ultrasound.

The row with the most Indian data and the most nuance, since the four phenotypes differ in how strongly metabolic features feature at all.

A medicine already prescribed: The prevalence varies sharply by drug and class. A relevant sign is weight gain or a stall beginning soon after a medicine was started or a dose was raised, often with rising appetite or new sedation. There is no single test. A clinician reviews the full medication list against the timing of dose changes.

The row most often missed, because the reader is rarely looking at a prescription written for something else. The characteristic signal is a weight trend that changes when a medicine starts or a dose goes up. The classes with measured effects are set out in full under drug-induced weight gain. Nothing about that is self-managed. It goes to the prescriber.

Perimenopause: Perimenopause typically occurs between ages 45 and 55, with around 1.5 kg per year reported in Western cohorts. Signs that are not the weight include cycle irregularity, hot flushes, disturbed sleep, mood change, and vaginal dryness. Assessment is mainly based on clinical judgement, with FSH and estradiol used selectively.

Cushing's syndrome: Cushing's syndrome is rare, with 1.8 to 4.5 new cases per million people per year in international series. No large Indian population incidence dataset exists. Signs that are not the weight include facial rounding, a supraclavicular fat pad, truncal fat with thin limbs, purple striae, easy bruising, muscle weakness, and hypertension resistant to treatment. A clinician may order 24-hour urinary free cortisol, late-night salivary cortisol, or an overnight 1 mg dexamethasone suppression test.

On the list because it is the classic teaching example, not because it is a realistic explanation. At 1.8 to 4.5 new cases per million people per year, an Indian city of a million produces one to five new cases annually. Its weight pattern is also wrong for a plateau: continuous central gain despite restriction, not a stall.

The failure mode of an article like this is that every reader with a stalled scale requests a full endocrine and sleep panel. That outcome is worse than the advice it replaced.

Start with the arithmetic. Plateaus at six months are the normal outcome of lifestyle intervention, so the population being screened is almost entirely people whose plateau is expected. Cushing's runs at a few cases per million per year. Hypothyroidism, at roughly one in ten, is common enough to be worth testing on symptoms but rarely explains a plateau that followed a 5 to 8 kg loss.

Screening an unselected population for a rare condition mostly returns false positives, incidental findings and results that are ambiguous rather than reassuring. Guidance on Cushing's is explicit that a work-up is justified only where pre-test probability is non-trivial, which in general practice it usually is not.

The costs are documented in the general diagnostic-stewardship literature rather than in Indian plateau-specific studies, which do not exist. They are cascades of further tests and procedures triggered by an incidental finding, health anxiety arising from an ambiguous result, and direct and indirect costs: out-of-pocket spending in a system where most of this is self-paid, plus time away from work. A strategy that sends every plateaued adult for a full panel would be disproportionate on the evidence and would harm more people than it helped.

The discriminating question is not how long the plateau has lasted. It is whether anything in the non-weight column is present.

There are two possible responses to a plateau. One assumes the problem is effort and prescribes more restriction. The other treats the plateau as information and asks what else is in the picture.

The second is a narrower and more checkable task:

1. Look at serial weights over six to 12 months, not six weeks.
2. Decide whether this is a plateau (weight stable) or a regain (weight rising).
3. Check whether any non-weight sign is present: a symptom, a changed prescription, a new pattern.

Where none is present, the plateau is behaving exactly as the trial literature says it should, and further testing is more likely to cost than help. Where something is present, it goes to a Registered Medical Practitioner who decides what to order. Which of the two applies to you is a question about measurement rather than motivation, and it is the same handover point described in what a blood panel finds that a diet plan cannot.

Is your plateau behaving like normal physiology, or is there something else in the picture? A Voy clinician can look at your weight history, symptoms, and medication list together and decide whether any further investigation is warranted. Book a clinical 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.