Review of Alster TS et. al. Effect of a novel low-energy pulsed-light device for home-use hair removal

June 2, 2009

Dermatol Surg. 2009 Mar;35(3):483-9.

Effect of a novel low-energy pulsed-light device for home-use hair removal

Alster TS, Tanzi EL.

Review copyright 2009

This appears to be the first clinical study of the Silk’n home use intense pulsed light for hair removal.  The authors call this “the beginning of a major trend in laser dermatology.”  Given the nicely performed study, and promising results presented, they may be right.

In this study, hair removal was evaluated in one or more treatment areas of twenty women, while matched control areas were left untreated.  The subjects generally had lighter skin and darker hair, and treated areas included underarms, forearms, legs, and bikini lines.  Hair counts were performed in treatment areas and control areas, using a reliable method of manual hair counts (average of 3 manual counts in a 2cm2 or 3cm2 template).  Follow-up measurements were performed at 3 and 6 months, which is generally accepted as the appropriate time frame to assess permanent hair reduction.

The results were quite impressive – virtually no change in hair counts of the control sites at any time, but significant reductions in hair counts (36% to 53% depending on body location) at six months.

While small, this study was appropriately powered and appropriately controlled, and used objective, quantitative outcome measures.  While it would have been nice if the study included multiple sites to eliminate the potential for site bias, a single site study is reasonable for the early stage of clinical work and small size of the study.  (According to a company press release in March 2008, more than 150 patients were treated at 4 total centers, where the results of this first study were successfully replicated.  AestheticDeviceReview hopes these results are published soon.)

A curious point of this study is the 2-week intervals in the series of 3 treatments in the study.  Typically, laser hair removal is performed with at least 4-week intervals between treatments, to maximize the amount of hair follicles in the anagen phase, which is thought to provide the best results.  Perhaps, in addition to their novel device, this novel application of bi-weekly treatments contributes to the results.  Further study is needed here.

The study’s author, Dr. Alster, purchased stock options in the company subsequent to the study.  Given these results, and the reported results from the expanded data set, we can expect these options to become quite valuable.


Review of Coleman, SR et. al. Clinical Efficacy of Noninvasive Cryolipolysis and Its Effects on Peripheral Nerves

May 3, 2009

Aesthetic Plast Surg. 2009 Mar 19

Clinical Efficacy of Noninvasive Cryolipolysis and Its Effects on Peripheral Nerves

Coleman SR, Sachdeva K, Egbert BM, Preciado J, Allison J.

Review copyright 2009

Coleman and colleagues present the first clinical report on the ‘love handles’ of 10 patients using the Zeltiq non-invasive fat reduction device.  The investigators used repeatable data collection methods, and proper data analysis procedures, resulting in a highly credible paper.

During a single patient visit, the Zeltiq treatment was administered to one of two love handles, with two to three applications of the cooling device, so as to sufficiently cover the area to be treated.  The opposite love handle was left untreated and used as the control. Follow up visits for fat thickness measurements were performed at 2 months and 6 months.

Fat thickness reduction (from ultrasound images) was used as an quantitative measure of efficacy, at various follow-up points compared to baseline.  The investigators went to great lengths to ensure measurement accuracy, using a transparency to align ultrasound measurements at the various time points, and collecting up to 12 evenly spaced ultrasound measurements at each location.  Generous amounts of ultrasound coupling gel were used to prevent the ultrasound probe from compressing the dermis and fat.  Overall fat layer thickness reduction was calculated by subtracting the ‘fat thickness change on the control side’ from the ‘fat thickness change on the treated side,’  where the fat thickness change for each patient was the average of the up-to-12 individual measurements.  The authors properly used this internal control, unlike so many papers in this field.

Additionally, efficacy was also measured by photographic assessment of treated versus control area, using standardized photography setup at a series of viewing angles every 22.5 degrees from 0 degrees to 360 degrees.  Three blinded reviewers compared baseline and 4-month post-treatment photos for five subjects for whom four month follow-up was available.  Again the investigators used blinding properly, to produce reliable results.

Transient numbness is a common, non-serious complication of the Zeltiq treatment.  Sensory neurologic assessment of the treated area was performed in nine patients at baseline and weekly post-treatment for 2 months. In one patient, nerve biopsies were collected at 3 weeks and 6 weeks post-treatment, at the treated site and a contralateral control, and evaluated for signs of sensory impairment.

There were no serious adverse events.  One patient had a non-serious adverse event during treatment.  After treating the median area of the love handle, the applicator was applied to the left anterior area.  After several minutes the subject reported pain and the treatment was aborted.  Other than redness that lasted six hours there was no other effect.  This subject did not complete the treatment, so no follow-ups were performed.

Erythema was observed in all 25 treatment locations of 9 patients who completed study. Numbness was reported at 24 of 25 treatment locations, which improved at a 1-week phone follow-up and resolved completely by the 2 month follow-up visit. Three of nine subjects did not complete the 6 month follow-up ultrasound, and two subjects did not complete the 6-month follow-up photographs.

Unlike prior reports of non-invasive body contouring, the authors in this paper present their results as a percent reduction, rather than a centimeter reduction.  AestheticDeviceReview speculates that the authors found a correlation between starting fat thickness and the amount of reduction, and so chose to report a percentage rather than an absolute amount.  In the six patients with six-month data, ultrasound fat thickness reduction was 20.4% +/- 4.9% in the treated side versus control at two months, and 25.5% +/- 9.5% at six months.  All patients had at least 11.5% reduction at 2 months, and 10.7% reduction at 6 months.  These are impressive results for a single treatment,demonstrating that results last for at least six months. 

For the blinded photographic assessment, the blinded reviewers were able to correctly pick the follow-up versus the baseline image 93% of the time.  If there were no treatment effect, we would expect the reviewers to be correct 50% of the time.  Again, these are impressive results, and objectively indicate cosmetic improvement.

For the nine subjects who completed the battery of sensory tests, 3 subjects had no change from baseline during the study on any test, and the other 6 showed steady improvement over six weeks with full resolution at week seven.  Nerve biopsy results on one patient showed completely normal nerves at 6 weeks, suggesting cryolipolysis does not cause any long-term damage.  Given the small number of patients, these results do not completely rule out the possibility of long-term nerve damage in a larger population of patients.  We need to await further results.

Transient hypo-esthesia seems to be a common complication of the Zeltiq technique.  It is not explained why one patient felt pain and could not complete treatment.  Fat thickness measurements were performed in a well controlled manner, and the treatment side was properly analyzed versus controls.  Unusually, fat thickness reductions were reported as a percentage reduction, and centimeter reduction was not reported.  There was no correlation between the fat thickness change and weight change in this group.  Blinded photographic assessment shows that the clinical changes are visible to trained observers.  It is not clear if the clinical changes are evident to patients, especially since the changes take a long time to become visible. Patient satisfaction was not measured, and it is not known if the patients were offered subsequent treatment of the contralateral side.  It is not known why there were so many drop-outs for follow-up; possibilities include the large number of follow-up visits, the long duration between the treatment and the 4 and 6 month follow-up visits, or a lack of subject interest in the result.  Nevertheless, the main results – sustained fat thickness reduction and cosmetic improvement – are credible and compelling.

Review of Boulos PR et. al. In the eye of the beholder–skin rejuvenation using a light-emitting diode photomodulation device

April 12, 2009

Dermatol Surg. 2009 Feb;35(2):229-39.

In the eye of the beholder–skin rejuvenation using a light-emitting diode photomodulation device

Boulos PR, Kelley JM, Falcão MF, Tremblay JF, Davis RB, Hatton MP, Rubin PA

Review copyright 2009

In an unusual article, Boulos and colleagues present compelling evidence that the GentleWaves LED device simply does not work, and that patient satisfaction is simply the result of a placebo effect.

In this study, 36 patients were treated with the GentleWaves device weekly for 8 weeks, as suggested by several other researchers.  Standardized photography was performed pre-treatment, immediately post-treatment, and 1 month post-treatment.  The unblinded investigator also performed clinical assessments at the same time points.

Thirty raters were asked to select the “after” photo from a randomly ordered before/after pair, looking at one of five aspects of the treatment – redness, wrinkles, dark spots, skin roughness, and general appearance.  Unfortunately, the results were not different from chance.

The data were re-evaluated by using a consensus of three raters for each aspect.  Still, the results were no different from chance.

Next, statistical power was increased by using raters’ confidence in their own judgements to weight the data.  In other words, this analysis more highly weighted the decisions that the raters felt sure about, and underweighted the decisions where the raters felt unsure.  Nevertheless, results again were not different from chance.

Finally, the unblinded investigator’s clinical assessments showed that “there was essentially no effect, or if anything, a negative effect.”

Despite the lack of any objective evidence of improvement whatsoever, patients were quite pleased with the results.  The authors conclude that the treatment is a classic placebo, where patients feel better because they believe that the treatment will work, even though it does nothing.  At AestheticDeviceReview, we would like to see more devices subjected to this type of study.

Review of Hantash BM et. al. Two articles on the Primaeva Bipolar Microneedle Radiofrequency Device

March 29, 2009


Lasers Surg Med. 2009 Feb;41(2):87-95.

Pilot clinical study of a novel minimally invasive bipolar microneedle radiofrequency device.

Hantash BM, Renton B, Berkowitz RL, Stridde BC, Newman J.


Lasers Surg Med. 2009 Jan;41(1):1-9.

Bipolar fractional radiofrequency treatment induces neoelastogenesis and neocollagenesis.

Hantash BM, Ubeid AA, Chang H, Kafi R, Renton B.

Review copyright 2009

In a pair of articles published a month apart, Hantash and colleagues present the first descriptions of the histological and immunohistochemical effects of the bipolar radiofrequency needle device from Primaeva. Several of the authors are affiliated with Primaeva, and the study was funded by Primaeva. AestheticDeviceReview takes a closer look at these two papers.

Three studies were performed on human abdominoplasty or face-lift patients prior to surgery. In the first study of five patients, a broad range of dosages were applied by varying tissue temperature and duration. Histological results from this group were used to select a narrower range of dosages for examination in a 10-patient second study. Finally, in the third study, a single dosage of 72C for 4 seconds was applied to 22 patients, and both histological and immunohistochemical testing was performed.

Briefly, the radiofrequency device consists of an RF generator attached to a handpiece, which includes a single-patient-use array of 10 needles arranged in 5 pairs. The 250-micron needles are spaced 1.25mm apart, and each needle-pair is independently powered by the generator. Each needle is 6mm long, with the top 3mm insulated and the bottom 3mm exposed to allow electrical current flow. The needles are inserted at a shallow 20 degree angle to the epidermis, such that when properly placed, the tip of the needle is 2mm from the epidermis. Insertion is done by spring-loaded injection. When properly inserted, the exposed portion of the needle is positioned nearly parallel to the skin surface, at 1mm to 2mm below the surface. Each needle pair has a temperature sensor, which is used to control the power so that constant temperature can be maintained. An impedence measurement from each needle-pair also provides feedback on the depth of needle insertion. In this study, a separate skin cooling device is held against the skin after the needles are inserted, maintaining 15C to protect the epidermis from conducted heat from the RF needles. In this prototype configuration, this cooling device requires a 2nd operator, and makes the technique somewhat more difficult to apply.

Because current flows between the two needles in each pair, we expect the RF energy to create five damage zones (one between each pair), and if the heat is applied long enough, temperature conduction should expand the damage zone in all directions, resulting in one contiguous zone of damage. The surface cooling device should prevent the conducted heat from damaging the epidermis. This is exactly what the authors show. In the example shown in the paper, by setting the device at 70C for 1 sec, five zones of about 2mm by 1.25mm x 2.5mm were created. At 70C for 4 seconds, one contiguous zone of damage of about 12mm x 1.5mm x 2.75mm was created. Unfortunately, while many patients were studied, the authors did not present quantitative data on the mean and variation of the size of the damage-zones, so we are left to wonder about intra-patient variation and inter-patient variation.

This description makes it clear that the device does not provide a fractional treatment, which would be characterized by 50 or more micro-thermal-zones of damage per square centimeter. Instead this device creates one to five milli-thermal-zones of damage per square centimeter, so the term “fractional” should not be used. AestheticDeviceReview believes that the relatively large volume of RF-needle-created tissue coagulation may even turn out to offer a clinical benefit compared to fractional treatments, so the fractional label may be unwise as well as unwarranted.

Interestingly, while large amounts of tissue were coagulated, the authors show that important dermal structures such as blood vessels, hair follicles, sebaceous glands, and sweat glands were undamaged by the treatment. Men getting wrinkles treated will not have to worry about bald patches in their beards. Similarly, needles placed into the subcutaneous fat did not cause fat damage, as electrical energy appeared to flow through the interstitial collagen (fibrous septae?). AestheticDeviceReview sees some potential for this technology as a treatment for cellulite.

In the immunohistochemical study, the authors nicely show that the tissue damage is not necrosis, but a zone of thermal coagulation that proceeds through a wound healing response over several weeks, consisting of sequential phases of inflammation, proliferation and ultimately remodelling. Again this should come as no surprise, as this response is similar to that seen with several other radiofrequency devices that have been developed for shrinking collagen-rich soft tissue. For example, temperature-controlled RF needle technology is used to shrink soft palate tissue as a treatment for snoring. See company white paper at:

Unlike non-ablative fractional laser treatments, all patients experienced swelling and focal edema which resolved in 48 hours following treatment. Despite the use of skin cooling, erythema appeared on the surface above each thermal damage zone, and this resolved within 8 hours of treatment. Several patients experienced purpura as well, which resolved in less than 7 days. So, overall “downtime” is greater than non-ablative fractional and less than ablative fractional laser treatments.

Is the downtime worth the result? Here, the authors leave us hanging. No data is presented on tissue shrinkage. Why didn’t the authors place tattoos prior to treatment, and measure the effect of treatment on tattoo distance? No data is presented on any change in dermal thickness. Why were so many patients studied, if quantitative histological analysis was not going to be performed? We at AestheticDeviceReview can only wait – perhaps we will see another study next month.

Review of Hruza et. al. Skin Rejuvenation and Wrinkle Reduction Using a Fractional Radiofrequency System

March 22, 2009

J Drugs Dermatol. 2009 Mar;8(3):259-65.

Skin Rejuvenation and Wrinkle Reduction Using a Fractional Radiofrequency System

Hruza G, Taub AF, Collier SL, Mulholland SR.

Review copyright 2009

In this paper, Hruza et. al. present the first report on Syneron’s MatrixRF technology, including both clinical and histological results.  While the inclusion of histological findings is welcome and helpful, both the clinical and histology studies have significant methodological flaws which diminish the value of the report.

The MatrixRF is available with two different “tips”, one with 64 “pins” in an 8×8 arrangement and one with 144 pins (12×12).  The description of each array is unclear, it appears that each pin is an electrode with a width of 200 microns.  The inter-pin spacing on the 64-pin tip is 1.5mm while the denser 144-pin tip has an inter-pin spacing of 1mm, and RF energy is apparently conducted though tissue between adjacent pins. The 64-pin tip is described as providing 5% coverage, and the 144-pin tip is described as 10%.  Delivered energy may be varied by the user.

The histology study should tell us about immediate tissue effects as a function of tip and energy, as well as the uniformity of tissue effect across each tip array and the variations across patients.  In this report, the authors show that the electrodes ablate and coagulate tissue, and that greater energy-per-pin (either fewer pins for a given energy, or more energy) generally results in a deeper ablation. Histological images show that the maximum ablation depth for the highest energy-per pin reaches just into the dermis, and that more typical settings limit the effect to the epidermal layer.  Unfortunately, the uniformity across the array and the variation across patients is not adequately addressed.  The authors tell us that 7 patients were studied with “different combinations of electrode-pins aray densities and energies,” at 7 time points (ranging from immediate to 30 days).  The authors do not tell us which energies were applied to each patient, and in what arrangement.  Worse yet, data on all patients and all time points is never presented.  Instead, the authors provide a table of which presents the “average of maximal depth” of tissue effect, for broad and overlapping ranges of energy delivered.  For example, for the range of 6 to 16 J of energy delivered, with the 64-pin tip, the depth of tissue effect is listed as 250 microns.  We do not know whether this represents one pin in one subject, multiple pins in one subject, or multiple pins in multiple subjects.  Further, this energy range varies by almost a factor of 3. Are we to expect the same results at 6 J as at 16J?  Will every subject have the same response?  The authors also claim, in the paper’s discussion, that depth of tissue effect depends on inter-pin spacing, which is different between the two tips.  However, for a given energy, the 64-pin array has both a wider spacing and a higher energy-per-pin than the 144-pin array.  There is no analysis which tells us whether the increased damage with the 64-pin array is due to the wider spacing or the higher energy-per-pin.

Post-treatment, the histology study should tell us how the body responds to the injury.  Here the authors present a brief, three sentence, qualitative description of tissue healing from the ablative and thermal injury, stating that the epidermis begins to re-epithelialize immediately and that full healing occurs within 2 days.  Given that readers are already very familiar with this type of tissue effect, and the body’s response, perhaps a fuller description of results is unnecessary.

In the 35 patient clinical study, each patient received a series of 3 facial treatments separated by 3 to 4 week intervals, and returned for a one-month follow-up evaluation.  No rationale for the interval was provided, as the histological study showed full recovery at 2 days.  Treatment settings (tip type and energy level) were left to the discretion of the treating physician, and were not reported.  Subjects were evaluated via clinical assessment, photographic assessment, and the usual self-reported-patient-satisfaction survey.

Importantly, the study lacked adequate control.  No untreated facial areas were used to control for any non-treatment-related, concurrent changes to the skin, such as changes in patient hydration or sun exposure.  Consequently, it cannot be determined whether clinical changes were due to the treatment or to other factors.  Further, outcomes were evaluated by comparing the post-treatment clinical evaluation to the pre-treatment photographic record.  In addition to being unblinded, these readings are highly influenced by the quality of the photography.  Thus they cannot be considered reliable.  Self-reported patient surveys are also known to be unreliable, and while these results appeared to be similar to the clinician evaluations, no correlations were reported.  The analysis of this experiment included an interesting result: about half of the patients saw a 40% or greater improvement in clinician-evaluated brightness, tightness, and wrinkling.  If at least a 40% improvement is needed, for a patient to be considered a success, then even using these uncontrolled, unblinded results, about half the patients were not successful.

This type study design is unfortunately all-too-common in the aesthetic device literature.  How could it have been improved?  First, a randomized split-face design should have been used.  As it was unknown prior to the study whether the treatment would have results that differed from doing nothing, an untreated control was appropriate.  (A post-study cross-over could have been offered if patients desired.)  Second, at each study visit, three blinded clinicians should have evaluated the left versus right side of the face, using a standardized procedure.  Intra-observer agreement could be analyzed to determine whether these evaluations were reliable, and in cases where the clinicians did not agree, the data would be discarded.  Third, in the analysis, the difference between the treated and untreated sides would be evaluated for each of the pre-treatment visit, treatment visits, and follow-up visits.  The analysis would measure how this difference changed over time.  This is the proper analysis of an internal control.  Fourth, a minimum improvement in the difference between treated and control sides would be pre-specified as a rule to consider a patient successful.  The study should be sized to show that an appropriate percentage (e.g. 75%) of patients could achieve success.

While the study design and analysis left much to be desired, sufficient data was presented to indicate that the MatrixRF probably provides a fractional version of a light peel.  However, a disclaimer may need to be given to each patient: “actual results may vary.”

~ AestheticDeviceReview

Review of Brown S. et. a. Characterization of Non-thermal Focused Ultrasound for Non-invasive Selective Fat Cell Disruption (lysis): Technical and Pre-clinical Assessment

March 15, 2009

S. Brown, PhD et al

Characterization of Non-thermal Focused Ultrasound for Non-invasive Selective Fat Cell Disruption (lysis): Technical and Pre-clinical Assessment

currently available on PRS Advance Online at

In this new paper, Spencer Brown MD et. al. performs four pre-clinical experiments to elucidate the acute biological effects of the Ultrashape device for non-invasive fat cell disruption.  Brown’s five co-authors are Ultrashape employees.  In general, the presented work appears to be careful and the results accurate.  Unlike the previously reviewed Zeltiq pre-clinical study, however, several important pre-clinical experiments were not performed, so we still do not know how the acute biological effects of the Ultrashape device are related to ultimate clinical outcomes.

In the first two experiments, the authors characterize the energy delivery of the UltraShape probe in water, which is a standard method for characterizing ultrasound energy fields.  Brown shows that the device focuses the ultrasound energy in a volume that has a diameter of about 8mm, and a depth that ranges from about 5mm to about 25mm from the probe.  Brown shows that the energy density (power per cm2) at the probe-water interface is very small, as desired.  Further, the authors showed that the ultrasonic energy created air bubbles in the focal region, consistent with a non-thermal cavitation effect.  Quantitative measures of ultrasonic power density were performed at 0mm and 14mm depth, and showed an absence of “hot spots.”  An improvement to the study would have included power density measurements at 1.5-2mm (approximately the depth of the dermal-fat junction) and 25mm (to characterize the extent of the ultrasonic energy transmission).

In the third experiment, the UltraShape probe was characterized in a gel phantom intended to simulate the ultrasound transmission properties of skin and fat.  In this case the focal volume was 9mm in diameter (slightly less focused than in water) and extended about 18mm in depth (the distance from the surface was not reported, but appears to extend from about 4mm to 22mm from the probe according to the figure).  Again, bubbles were seen in the focal region in this model, consistent with a non-thermal cavitation effect. 

In the fourth experiment, porcine skin was treated and then immediately evaluated with both frozen sections and histologically stained sections.  Untreated control skin was also evaluated to ensure that results were not due to processing artifact.  Importantly, no effect on skin color or skin appearance was seen on the animals receiving this treatment, and histology showed that the dermis and epidermis appeared to be completely unaffected by the treatment.  The subcutaneous fat, however, showed evidence of tissue injury in both the frozen sections and the histology.  Histological staining for LDH activity using NTBC (elevated levels of LDH indicate tissue breakdown) demonstrated a layer of adipocyte cell breakdown extending from about 15mm to 25mm of tissue depth.  In the treated tissue, but not the control tissue, frozen sections and two other histological stains (H&E and Masson’s Trichrome) indicated a “defined area of tissue destruction” extending from approximately 8mm to 18mm of tissue depth.  This region showed clear disruption of fat cells, while connective tissue, blood vessels and nerves remained intact.  No evidence of any thermal damage was seen in any treated tissue, again “consistent with initial cavitation followed by the mechanical destruction of cells.”  The authors state that fourteen animals were treated in this study, and the results were “consistent over time” despite the use of “multiple devices, [and] multiple transducers [by] numerous users.”  No quantification of subject-to-subject variability was provided.  For example, the authors should have measured the zone of tissue damage in each animal, and presented the results as averages with 95% confidence intervals.

So far, the results are promising, with clear evidence of non-invasive damage to subcutaneous fat and no apparent impact to the dermis.  Unfortunately, the analysis stops there.  For example, it is clear from the presented images that not all fat cells in the treated region were disrupted, but the authors do not quantify the percentage of the treatment volume that was disrupted.  Further, the response of the animal to this treatment was not studied.  Biopsies of treated and control areas were not performed at meaningful time durations subsequent to treatment (such a 1 day, 1 week, 1 month and 3 months post-treatment).  Unlike the recent Zeltiq study, we have no idea how the skin and subcutaneous fat respond to these injuries.  Does inflammation occur?  While no changes to the histology of the dermis were seen immediately post-treatment, could an inflammatory response occur over time?  Are non-viable cells removed or replaced?  Does this treatment cause meaningful changes in fat thickness compared to control volumes over time, and if so, when do these changes occur?  Lastly, blood lipid profiles were not analyzed in this study.  We cannot know if release of lipids from the disrupted adipocytes has any systemic effect, either on blood lipids or the liver.

The authors state that these study “observations do not directly lead to predict clinical results,” and they recommend further clinical evaluation.  However, the real need is for further pre-clinical evaluation.  Perhaps this partly explains why this device, widely available in Europe and Canada, is not yet cleared by the FDA.

Review of Bousquet-Rouaud, High-frequency ultrasound evaluation of cellulite treated with the 1064 nm Nd:YAG laser

March 8, 2009

J Cosmet Laser Ther. 2009 Mar;11(1):34-44

High-frequency ultrasound evaluation of cellulite treated with the 1064 nm Nd:YAG laser.

Bousquet-Rouaud R, Bazan M, Chaintreuil J, Echague AV.

Dermatological Laser Unit, Millenium Clinic, Montpellier, France.


This article presents the first evaluation of the use of a high-power pulsed Nd:YAG laser for the treatment of cellulite.  Three of the four authors are employees of Candela Corporation, the maker of the laser.

The study at first appears to be well-designed.  Quantitative endpoints of dermis thickness and ultrasound density were chosen, and preliminary studies were performed to show that measurement methods were not subject to intra-observer or inter-observer error.  Twelve patients were treated, with one thigh randomized to treatment and the other randomized to control.  In addition to pre-treatment ultrasound measurements, follow-up measurements were performed at 1 month and 3 months after the final treatment of the series of three treatments.  Adjunctive photographs were taken before each treatment, and presumably at the follow-up visits.  The example photograph appears to be of high quality.  The treatments were appropriately standardized to minimize the impact of dosage variation on outcome.  Likewise, the patient population appears to be reasonably homogenous in terms of age, body mass index, and pre-treatment evaluation of cellulite severity, to minimize the impact of patient variation on outcome variation.

Given the apparent care taken in the study design, the analysis was extremely disappointing.  The analysis presents only the variation in ultrasound measurements of the treated thigh, and completely ignores the control thigh.  The authors claim that the mean dermal thickness of treated thighs decreases from the beginning to the end of the study, but this is a flawed claim.  Because they do not present data on the control thigh, we cannot know if this result stems from a treatment effect, or from a natural variation in the patient population over time.  For all we know, the dermal thickness of the control thigh improved more than the treated leg.  This is the whole point of having a control thigh.  No other control variables, such as BMI or weight, were presented.

Further, the authors include the raw data for the dermal thickness measurements in all 12 patients, but offer no explanation why some patients have increased thickness at 1 month follow-up which then decreases at 3 months, while other patients show decreased thickness at 1 month with subsequent increases at 3 months.  Presumably, the treatment effect would go in generally the same direction in this homogeneous group that recieved the same treatment regime.  This variation adds to the concern that something else was going on with these patients during the study that might effect dermal thickness.  The same comments apply to the raw data on dermal echogenicity.

The exclusion of measurements of the control thigh from the analysis renders any conclusions moot.  The authors further present patient satisfaction data.  This type of data is notoriously challenging to interpret, as patients often feel satisfied that they are taking steps towards treating a condition. In this case, on average the patients reported being somewhat satisfied, and no patients reported being very satisfied.  Readers of the article should not be satisfied at all.