What Fat-Freezing Research Really Shows

Fat-freezing research supports the possibility of local contour changes, but it does not support treating every advertised percentage as a personal forecast. The useful question is more precise than whether it is “clinically proven”: proven to change which measurement, using which device, compared with what, and after how long?

Cryolipolysis uses controlled cooling to target fat beneath the skin. It is a contouring procedure, not a treatment for obesity or a route to meaningful overall weight loss. The FDA’s patient information makes that distinction and notes that people may not achieve their desired effect.

First, find out what the number measures

Fat-layer thickness is the depth of tissue at selected points, often measured with ultrasound. It is not a count of fat cells throughout the body. As a hypothetical example, a layer changing from 20 mm to 16 mm is 20% thinner at that measurement site. That does not mean losing 20% of your body fat, weight or waist measurement.

Circumference is a tape measurement around an area. Position and technique need to be consistent; it does not isolate the fat layer. Photographs assess visible contour, so matched posture, lighting and camera position matter. Satisfaction records the participant’s opinion. Someone can value a modest change, while another person finds a measurable change disappointing.

These distinctions are practical, not academic. In the inner-thigh study by Zelickson and colleagues, researchers used ultrasound, repeated tape measurements, standardised photographs and questionnaires separately. An advert combining those outcomes into one impressive “success rate” would hide what was actually tested.

A positive result still needs its denominator

That inner-thigh study treated 45 women with a particular CoolSculpting applicator. Its efficacy analysis included 42: three were excluded because their weight changed beyond the permitted range. At 16 weeks, average ultrasound thickness had fallen by 2.8 mm and average thigh circumference by 0.9 cm. After adjustment using untreated ultrasound control sites, the thickness reduction was 2.6 mm. These were local measurements after one cycle per inner thigh, not whole-body changes. Read the study.

The same paper reported 91% correct identification of pretreatment photographs by blinded reviewers. That percentage describes photograph judgements, not the proportion of fat removed. Although the study used useful measurement safeguards, participants were not randomly assigned to treated and untreated groups. Its specific applicator, selected participants and 16-week endpoint limit how widely the results can be applied.

“One session” can conceal several treatment cycles

A separate abdominal ultrasound study by Ponga-Manso illustrates why the treatment description matters. It retrospectively examined 30 women selected from 90 screened patients. At six months, the reported average local thickness reduction was 46.6%, with measured thickness falling from an average 9.4 mm to 5.0 mm. Weight and overall fat mass did not change significantly.

Crucially, the single appointment involved four to six cycles using specified CoolSculpting applicators. It was not one applicator placement. There was no untreated comparison group, and the study came from one centre. A striking percentage from that protocol cannot establish what one placement on another device will achieve. The paper also disclosed manufacturer-provided medical writing support, information worth reading alongside its methods and results.

What changes when there is a comparison group?

In Falster and colleagues’ randomised trial, 38 women were allocated to abdominal cryolipolysis or no treatment; 34 completed the study, 17 in each group. The treated group received one application using the Crio Top Body Redux device. A blinded assessor found no significant between-group benefit in ultrasound thickness, skinfold or circumference at 30, 60 or 90 days.

The circumference findings explain why a comparator matters. At the navel, the treated group’s average measurement fell from 85.79 to 84.23 cm by day 90; the untreated group’s fell from 84.79 to 83.36 cm. Reporting only the treated group’s decrease would suggest an effect that the comparison did not support.

This small trial also had limitations: participants knew their allocation, there was no sham procedure, and four did not complete follow-up. Its negative finding applies to the tested protocol and population. It challenges a universal promise of effectiveness without proving that every device or treatment plan is ineffective.

A review combines evidence; it cannot remove every weakness

A 2025 systematic review and meta-analysis included 30 studies and 3,158 participants overall. However, each pooled outcome used a smaller subset. For abdominal circumference, 211 participants contributed to an average reduction of 3.56 cm at around 12 weeks compared with baseline. Several other pooled measurements, including thigh circumference, did not show statistically significant reductions.

That abdominal number is a pooled before-and-after change, not proof of an equivalent advantage over untreated people. Combining studies does not automatically correct differences in devices, participant selection, additional interventions or measurement methods. A study average also conceals variation between people; statistical significance alone cannot tell you whether a change will feel worthwhile.

Safety evidence needs enough time

Common temporary effects include redness, bruising, discomfort and numbness. A more significant complication is paradoxical adipose hyperplasia, or PAH: treated fat enlarges and hardens instead of shrinking. It can emerge months later, does not normally resolve by itself and may need surgery. A growing firm bulge warrants assessment. FDA safety information.

A 2025 PAH review identified 29 affected patients among 13,078 across 28 studies, estimating a pooled incidence of 0.22% with a 95% confidence interval of 0.10–0.47%. Evidence certainty was low, and only four studies reported follow-up of at least 16 weeks. This patient-level estimate is neither a risk per cycle nor a measured rate for Birmingham. A small study reporting no cases cannot establish zero risk.

Take these questions into a consultation

• Which device, applicator and body-area studies support my proposed plan?
• Does the quoted percentage measure local thickness, circumference, photographs or satisfaction?
• How many people were analysed, what was the comparator, and when were results measured?
• How many placements or cycles does “one session” include?
• What would count as a worthwhile change for me, and how will it be recorded?
• What follow-up is planned, and who assesses delayed concerns such as a growing firm bulge?

You can use these questions when contacting the Birmingham team. Bring the advert or study that prompted your interest: a useful consultation should connect its evidence to your proposed treatment, including the uncertainties.

Sources and further reading

• Zelickson et al.: inner-thigh measurements and photographic assessment.
• Ponga-Manso: abdominal ultrasound after a multi-cycle session.
• Falster et al.: randomised abdominal trial.
• 2025 review of outcomes, adverse events and satisfaction.
• 2025 systematic review of PAH incidence.
• FDA: non-invasive body contouring technologies.