"Belly fat is always the last to go": the evidence on spot reduction

Your face has changed. So have your forearms, and someone at work has noticed. The waistband has not moved, or has moved so little that you have stopped believing the tape.

Almost everyone losing weight arrives convinced the abdomen is being held back until last, and the imaging evidence does not support that. Fat leaves unevenly, but the abdomen is usually among the more readily mobilised depots rather than the most stubborn. What makes it look otherwise is arithmetic and attention, both worth understanding before you decide to train harder at the one place that seems not to be listening.

Key Takeaways:

• The abdomen is typically a mid-stage depot in studies that track regional fat loss, not a late-stage one. The idea that it goes last is not supported by imaging evidence.
• A larger store means a given absolute loss is a smaller fraction of what is there. Someone losing a kilogram from a 30-centimetre waist loses a smaller visible share than someone losing the same kilogram from a 20-centimetre one.
• You cannot direct fat loss to or away from a body part through targeted exercise. The 2013 leg-press trial found bilateral fat reduction from unilateral training, not localised reduction.
• The contested lipolysis signals from trained tissue are real but too modest to show up on body-composition scans. Training a region does not meaningfully accelerate its fat loss.
• South Asian bodies carry more visceral and central fat at lower BMIs, which is a risk factor independent of appearance. Tracking waist-to-hip ratio alongside weight gives a more complete picture than the tape alone.

Studies that scan people before and after weight loss find regional differences every time, and no consistent order. In overweight and obese postmenopausal women losing about 8% of body weight over six months, both android fat at the waist and gynoid fat at the hips and upper thighs fell. The gynoid region held roughly twice as much fat to begin with, about 7.5 kg against 3.5 kg, and lost more absolute mass. The ratio between the two fell only when exercise was added to the diet.

The pattern flips in other designs. In a six-month periodised training programme in 31 healthy women, published in 2000, total fat mass fell by about 10%, but the arms lost around 31% of their fat mass, the trunk about 12%, and the legs barely changed despite gaining lean tissue. Of roughly 2.6 kg of fat lost, 1.3 kg came from the trunk and 1.1 kg from the arms.

The physiology explains a spread rather than a sequence. Abdominal subcutaneous fat cells respond 10 to 20 times more strongly to catecholamine-driven lipolysis than gluteal cells, with about twice the beta-adrenoceptor density, while gluteal and femoral fat carries more of the receptors that suppress fat release. Oestrogen favours lower-body storage in women, part of why hips and thighs often hold on while central fat is already moving.

This is what resolves the contradiction between the scans and the mirror. A region holding the most fat can lose the most fat in kilograms and still look least changed, because the same loss is a smaller proportion of a bigger starting store, and because skin, posture and overlying tissue mask it.

Two datasets make that concrete. In Army Ranger candidates losing about 13% of body weight over eight weeks of high-volume training, truncal fat made up roughly 41% of total fat at baseline and accounted for approximately half of all the fat lost, a pattern the researchers described as abdominal fat being mobilised alongside limb fat rather than preserved until last. In the postmenopausal cohort above, android fat was only 7 to 9% of total fat and still fell meaningfully.

Perception does the rest. In a 16-week weight-loss intervention in overweight and obese women, improvements in body image were driven mainly by perceived physical change, especially perceived reductions in body fat, which explained 12 to 37% of the variance. Measured anthropometric change played a smaller part. Systematic reviews of attentional bias find that women with high body dissatisfaction direct more visual attention towards weight-related body parts. Watching one region closely makes its slow proportional change the most salient fact about the process, which is the condition under which "it is going last" feels like an observation rather than an impression.

The direct tests of spot reduction are unilateral training studies, where one limb works and the other does not, so each participant is their own control.

A 2013 trial put 11 young adults through 12 weeks of unilateral leg-press training at 960 to 1,200 repetitions per session. Whole-body fat fell about 5.1%, with no change in either the trained or the untrained leg, while upper-extremity fat fell about 10.2% and trunk fat about 6.9%. A small trial of 24 sedentary adults doing seven abdominal exercises five days a week for six weeks on an isocaloric diet found no significant change in body weight, total body fat, waist circumference or abdominal subcutaneous fat thickness.

The contested signals deserve stating rather than skipping. Acute work using one-leg knee extension shows subcutaneous fat next to contracting muscle has measurably higher blood flow at low intensity and higher local lipolysis at high intensity than the resting side, so local fat release is genuinely augmented during exercise. Some small, older studies using calipers or ultrasound reported modest reductions over exercised regions, particularly in men. Those effects tend to disappear once DXA, CT or MRI and adequate controls are applied, and the quantity mobilised locally is small against total energy expenditure. Local lipolysis is real. A chronic reduction in regional fat mass from training that region has not been demonstrated. Three of the trials here reach this article through a secondary research summary rather than the primary papers, so their details are reported as summarised.

Facial fat is part of the same systemic store, and the clearest imaging comes from an unusual population. In a facial MRI study of 24 patients with HIV-related lipoatrophy, wasting or weight change, superficial facial fat fell by a median of 5.2 mL in those with lipoatrophy and 6.0 mL in those losing weight, and rose by about 20.2 mL in those gaining more than 10% of body weight. Serial facial imaging through ordinary weight loss in general populations is scarce, so the rate at which any individual sees facial change is not quantified.

Facial exercise claims do not survive controlled trials. In a control-group study of 18 adults doing daily facial exercises for seven weeks, only the upper-lip area differed significantly, and the authors concluded facial exercises could not be considered effective. A systematic review reached the same verdict on a weak base of few, small, largely subjective trials. Commentary on a facial-muscle device trial noted reductions of about 0.5 mm at the midface and 1.4 mm at the jawline after eight weeks, judged clinically insignificant. Chewing training raises masseter thickness and bite force without documented contour change.

Localised fat-reduction procedures are a different category, assessed here as claims rather than offered as options. Deoxycholate preparations hold US Food and Drug Administration approval for contouring submental fat on trials showing superiority to placebo, with adverse events including pain, swelling, bruising and, rarely, nerve injury. A review of four trials and one case series, about 101 patients, found cryolipolysis for submental fat produced statistically significant local reduction with mainly mild, transient side effects. Up-to-date Indian approval status for specific formulations and devices is not consolidated in accessible sources and would require direct consultation of CDSCO databases. These set the boundary: localised fat removal, where it happens at all, is drug- or device-mediated, and any decision about one belongs with a registered medical practitioner.

Indian readers start from a distribution that makes this effect larger. Comparative DXA work across European, Maori, Pacific Island and Asian Indian adults found that after adjusting for age, height and weight, abdominal fat mass was higher in Asian Indians in both sexes, and abdominal fat as a proportion of total fat was highest in Asian Indians and lowest in Europeans.

A larger central store means a given loss is a smaller fraction of it, so the tape moves less than the same person would expect from a Western before-and-after photograph. Serial-imaging studies following Indian adults through weight loss are limited or not publicly accessible, so no Indian cohort documents this directly, and no agreed regional order of fat loss exists in any population. Claims about exact sequences, or centimetres lost per kilogram, cannot be traced to a primary source. A more useful reading than the waistband is a measure you can repeat consistently, such as the waist-to-hip ratio, interpreted with a registered medical practitioner alongside your weight trend and blood results rather than against a mirror.

Want to understand how your fat distribution affects your risk and your rate of change? Speak to a Voy clinician.

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.