Glossary: the terms used on this site, defined plainly

Terms are grouped by topic rather than alphabetically, because that is how they are met. Each links to the page where it matters most.

Trials and statistics

Arm. One group in a randomised trial, defined by what it received (for example, tirzepatide 15 mg, or placebo).

Baseline. The measurement taken at the start, before treatment. Percent changes on this site are from baseline unless stated.

Confidence interval (95% CI). The range within which the true average effect plausibly lies given the data. STEP 4's difference of -14.8 points (95% CI -16.0 to -13.5) means the data are consistent with a true difference anywhere in that range.

Estimand. A precise statement of what a trial result measures. The treatment-policy estimand includes everyone randomised whether or not they kept taking the drug (STEP 1: -14.9 percent). The trial-product or treatment-regimen estimand estimates the effect if everyone had stayed on it (STEP 1: -16.9 percent). Both are correct; they answer different questions. See rate of loss by week.

Lead-in or run-in. An open-label period on the drug before randomisation, used in STEP 4 (20 weeks) and SURMOUNT-4 (36 weeks) so that the randomised comparison is between continuing and stopping. See maintenance after stopping.

Mean. The arithmetic average. Half of any trial arm did worse than it.

Placebo. An inactive injection given so that neither participants nor investigators know who is on drug. Placebo arms in the GLP-1 trials also received diet and activity counselling; see what the trials asked participants to do.

Randomised withdrawal trial. A design in which everyone takes the drug first, then is randomly assigned to continue or to switch to placebo. It is the strongest way to measure what happens when treatment stops.

PubMed id (PMID). The unique number for a paper in the National Library of Medicine's index. Every trial cited here carries one; paste it into pubmed.gov.

Body composition

Adaptive thermogenesis (metabolic adaptation). The fall in resting energy expenditure after weight loss beyond what the change in body size predicts. See why weight returns.

BMI. Weight in kilograms divided by height in metres squared. Blind to composition and fat distribution. See body composition measures.

Bioelectrical impedance (BIA). The method home body-fat scales use. Sensitive to hydration; treat single readings with suspicion.

DXA (dual-energy X-ray absorptiometry). The scan STEP 1 and SURMOUNT-1 used to measure fat, lean and bone mass in substudies. See lean mass evidence.

Fat-free mass / lean mass. Everything that is not fat: muscle, organs, bone, water. DXA reports lean soft tissue separately from bone. A change in lean mass includes water shifts, which is one reason lean loss in the first weeks overstates muscle loss.

Sarcopenia. Loss of muscle mass and function with age or illness. The reason older adults get a higher protein range (PROT-AGE) and the reason lean loss matters more after 60.

Visceral fat. Fat stored around the abdominal organs, more strongly linked to metabolic risk than fat under the skin. Waist circumference tracks it.

Waist circumference. Measured at the top of the hip bone or midway between lowest rib and hip bone; 102 cm (men) and 88 cm (women) are the most-used thresholds. See body composition measures.

Nutrition and energy

Energy deficit. Eating less energy than you expend. The GLP-1 trials counselled 500 kcal/day below estimated requirement.

g/kg/day. Grams of protein per kilogram of body weight per day, the unit position statements use. The 2025 GLP-1 joint advisory range is 1.2 to 1.6. See protein targets.

Mifflin-St Jeor. The 1990 equation for resting energy expenditure from weight, height, age and sex. See energy needs.

PAL (physical activity level). The multiplier applied to resting expenditure to estimate total expenditure; FAO/WHO/UNU bands run from 1.40 to 2.40.

Protein first. The joint advisory's practical rule for low appetite: eat the protein portion of a meal before the rest so it is not the part left over. See eating on low appetite.

REE / RMR / BMR. Resting energy expenditure, resting metabolic rate and basal metabolic rate: closely related measures of energy used at rest. Equations estimate REE; indirect calorimetry measures it.

TDEE. Total daily energy expenditure, REE times PAL plus the thermic effect of food.

Training

Progressive overload. Gradually increasing the weight, repetitions or sets so the muscle keeps adapting. The principle behind every resistance programme. See resistance training basics.

Repetition maximum (RM). The most weight you can lift for a given number of repetitions. Programmes are often written as a percentage of 1RM or as a repetition range.

Resistance training. Exercise against a load: weights, bands, machines or body weight. The joint advisory recommends it at least three times a week on GLP-1 therapy.

Moderate and vigorous intensity. Moderate: you can talk but not sing (brisk walking). Vigorous: a few words at a time (running). The 150 and 75 minute weekly guideline figures refer to these.

Medication

Compounded drug. A medication prepared by a pharmacy for an individual prescription rather than manufactured under an FDA-approved application. Compounded GLP-1 products are not FDA approved, their safety and effectiveness have not been evaluated by FDA, and they are not interchangeable with brand products. The trials on this site used brand products.

GLP-1 receptor agonist. A drug that activates the receptor for glucagon-like peptide-1, slowing gastric emptying and reducing appetite among other effects. Semaglutide is one. Tirzepatide activates both the GLP-1 and the GIP receptor.

Titration / escalation. The stepwise increase in dose over weeks written into the label, designed to reduce gastrointestinal effects. Dose decisions belong to the prescriber. The Semaglutide Hub and Tirzepatide Hub cover the schedules.