Lasers in Surgery and Medicine 45:283–295 (2013)

Evaluation of a Hot-Wire Hair Removal DeviceCompared to Razor Shaving

Brian S. Biesman, MD*Department of Ophthalmology, Dermatology, Otolaryngology, Nashville Centre for Laser and Facial Surgery, VanderbiltUniversity Medical Center, Nashville, Tennessee

Background and Objectives: We describe a blinded,controlled, prospective clinical study of a hot-wire devicepromoted for hair removal and the reduction or delay ofhair regrowth (no!no!, Radiancy, Inc., Orangeburg, NY)compared to a shaving control.Study Design/Materials and Methods: Twenty-twosubjects were treated by trained clinical staff with the hot-wire device according to its Instructions for Use on thelower leg two times per week for 8 weeks. An adjacent sitewas shaved with a razor blade on the same schedule toprovide a control. Subjects were followed for 3months afterthe last treatment to study the durability of the results.Standardized high-resolution photographs were made atbaseline, once a week during treatment, and monthlyduring the post-treatment follow-up period. Micro-tattooswere used to ensure treatments and photographs werereliably made in the same anatomical location from visit tovisit. Both active and control sites were shaved prior tobaseline and allowed to regrow for a fixed period of timebefore first treatment to provide a consistent and well-defined baseline hair condition. Quantitative hair countswere made by a third party from the photographs andstandard statistical analysis was performed to look fordifferences between the active and control sites. Visualassessments and quantitative analysis was also performedon the photographs to see if there were any differences inhair thickness (diameter) and hair color between the activeand control sites.Results: The results show that shaving and the hot-wiredevice are indistinguishable in short-term or long-termeffect, based on both visual assessment of the photographsand statistical analysis of the hair counts. The control(shaving) had a mean baseline hair count of 79.4, whichremained stable (74.8–84.3) during the 8 week-treatmentphase and climbed substantially after stopping treatmentto 98.8, 100.1, and 104.6 at 1, 2, and 3 months post-treatment, respectively. The active (hot-wire device) had amean baseline hair count of 86.0 which remained fairlystable (81.7–95.1) during the treatment phase and thenclimbed substantially after stopping treatment to 104.0,106.4, and 109.0 at 1, 2, and 3 months post-treatment,respectively. The difference in hair counts between thecontrol and shaving showed that (a) in the treatmentphase, shaving was slightly more effective at hair removalthan the hot-wire device with weak statistical significance(P < 0.05 at 5 of 7 time points) and (b) in the follow-up

phase, shaving and the hot-wire device were statisticallyindistinguishable (P ¼ 0.252, 0.0972, and 0.230 at 1, 2,and 3 months, respectively). Likewise, the difference inpercentage change from baseline in hair counts (whichnormalizes to baseline values) between the shaving controland hot-wire device is close to zero at every time point(�4.9% to þ4.9%) and the t-test P-values are high(0.154 < P < 0.890 over all the time points in the studyand 0.360 < P < 0.890 during the 1, 2, and 3month follow-up period), indicating no detectable difference betweenshaving and the hot-wire. In terms of hair characteristics,no difference in hair color or hair thickness was seenbetween the shaving control and the hot-wire sites in thetreatment or follow-up period.Conclusions: Relative to shaving, the hot-wire (no!no!)device does not produce lessened hair density, decreasedhair re-growth rate, greater duration of effect, nor inducechanges in hair thickness and color. We conclude that thehot-wire device does not offer any benefit as compared toshaving. Lasers Surg. Med. 45:283–295, 2013.� 2013 Wiley Periodicals, Inc.

Key words: home-use; hair removal; hair reduction; hot-wire; consumer; over-the-counter

INTRODUCTION

The no!no! (Radiancy, Inc., Orangeburg, NY) is a hot-wire device marketed to consumers for hair removal. Thedevice is used by rolling it across the skin’s surface, causinga heated wire in the device to move into position about halfa millimeter above the skin, where it can come into contactwith surface hair. The device has been the subject of oneprior journal article, which suggested the device wasclinically efficacious [1].While it is logical that the hot-wiredevice is capable of severing hairs and thereby affectingtemporary surface hair removal similar to a standardrazor, the ability of this device to induce long term changes

Contract grant sponsor: Tria Beauty, Inc.�Correspondence to: Brian S. Biesman, MD, 345 23rd AvenueNorth, Suite 416, Nashville, TN 37203.E-mail: bsbiesman@drbiesman.com

Accepted 17 April 2013Published online 5 June 2013 in Wiley Online Library(wileyonlinelibrary.com).DOI 10.1002/lsm.22144

� 2013 Wiley Periodicals, Inc.

in hair follicle growth has not been clearly establishedsince the prior work had a number of limitations, includingno shaving control, poorly controlled initial hair countconditions, unclear hair counting methodology, variationsin or poor quality photography, unclear means of indexingthe treatment locations, limited reporting of the raw data,lack of statistical analysis, minimal follow-up, low samplesizes, non-quantitative assessment methods, lack of blind-ing, and similar defects. Thus, the benefits of the hot-wiredevice beyond that of a shaving effect cannot be inferredfrom these studies.

The purpose of the present study is therefore to performa scientifically rigorous, independent measurement of theefficacy of the hot-wire device.

METHODS

Study Design

This was a prospective, single-center, controlled clinicalstudy using objective hair-counts. Its objective was tomeasure the efficacy of a commercially available hot-wirehair removal device versus razor-blade shavingcontrol. Endpoints assessed included short-term hairremoval and long-term suppression, reduction, or delayin hair regrowth, and evaluation of the color and thicknessof any recurrent hairs. The protocol was approved by aninstitutional review board (IRB), and the study wasperformed in accordance with the current internationalstandards for good clinical practice, including the Interna-tional Conference on Harmonization Good Clinical Prac-tice Consolidated Guideline, Declaration of Helsinki, andUS federal regulations 21 CFR Part 312 and Part 812.

Subjects

The willing subjects were recruited from the localcommunity. Enrollment was targeted at 25 with anobjective of 20 ormore completing the protocol. Completionwas defined as undergoing the baseline visit, all treatmentvisits, and at least two of the four follow-up visits includingthe final follow-up visit at 12 weeks post final treatment.The sample size was chosen, as, on the basis of previousexperience, it was determined that the N value was highenough that the studywould be sufficiently powered from astatistical perspective.

Primary inclusion criteria required subjects to be 18–55years of age, to have sufficient hair density in the studyarea (the lower leg) for accurate hair counts, to bewilling tohave hair removal in this area, to have sufficient contrastbetween skin and hair color for good photography, and toagree to abstain from shaving or using any other hairremoval method in the study area during the entire studyperiod. Primary exclusion criteria were cuts, abrasions,infections or other potential confounding or unsafe skinconditions in the study area; plucking, tweezing, waxing,chemical depilatories, or any method of temporary hairremoval other than shaving in the study area within12 weeks of the study start; any prior laser hair removal,electrolysis, or any other permanent hair removalmethodsin the study area; bleaching or any other hair lightening

product in the study area within 6 months of enrollment;a history of keloid scar formation; having ever takenspironolactone or undergone chemotherapy or expecting todo so within the study duration; and being pregnant orexpecting to become so within the study duration.

Materials

The device studied was the current version of the no!no!hot-wire hair removal system Model 8800 including thedevice body and its replaceable “Thermicon” tips thatcontain the hot-wire component, Instructions for Use (IFU,or, User Manual), instructional video, power supply, skinbuffing pads, and other miscellaneous components thatcome in the retail packaging. For sanitation purposes andto ensure that these materials were not overused, eachsubject had a unique tip and buffing pad that werecarefully stored at the study site between visits.Photography was performed with a Nikon D300 high-

resolution (4,288 by 2,848 pixels) digital SLR camera. Toprovide consistent, high-quality standardized photo-graphs, the camera was mounted to a custom fixturethat consists of a transparent acrylic plate that ispositioned directly against the skin and a camera mountthat fixes the camera 25 cm above the plate such thatthe camera images the skin through the plate. Imageswere saved on the camera flash memory cards, the studysite computer, and a password-protected, encrypted on-line repository.Other relevant materials included: micro-tattoo ink

(SkinCandy Tattoo Supply, Burbank, CA) and needleswith applicators (SofTap Inc., Livermore, CA); electric hairtrimmer with 2 mm trimmer heads (Oster); disposablerazors and shaving gel; and alcohol prep pads for skincleansing and skin moisturizer (Cetaphil).

Treatments

The study schedule was as follows:

• Visit 0: Consent, screening, enrollment, and pre-treatment procedures

• Visits 1–16: Treatments and photography• Visits 17–20: Follow-up visits and photography

At the initial visit and after informed consent, screening,enrollment, and pregnancy testing, the study staff identi-fied two generally symmetric 3 cm � 3 cm areas on theinside lower leg, each having a minimum of 15 hairs, asthe study sites. Sites were adjacent but separated by 3 cm.The upper area (closer to the knee) was delineated as thecontrol (shaving) site and the lower or distal area as theactive (treatment) site. To provide visit-to-visit reproduc-ibility, the sitesweremarked at this visit with two invisibleink (visible under black light only) microdots located at thetwo extreme corners of the treatment site. A rigid templatewas used to precisely locate these two microdots. At eachtreatment visit, the two microdots were visualized withblack light and then marked with visible ink. A rigid

284 BIESMAN

templatewas then used tomark the corners of the two sitesbased on the two reference microdots.Following the tattooing and site marking, both sites

were (a) photographed to document the pre-study condi-tion; (b) clipped if necessary to achieve a hair shaft lengthof 2 mm for consistent photographs and accurate haircounts and in accordance with the no!no! hot-wireinstructions, (c) photographed to document the post-clipping condition, (d) shaved to provide a controlledbaseline condition for both sites, and (e) photographedagain to document post-shaving skin conditions. Thebaseline hair counts (for the control site and the activesite) were taken from the pre-treatment photos at thesubsequent visit, which occurred 4 days later. This ensuredthat both control and active sites had a well-defined initialhair state for all subjects, with currently telogen (dormant)hairs shaved flush with the skin and currently anagen(active) hairs having 4 days of stubble or regrowth, suchthat the effect of subsequent shaving and hot-wire treat-ments can be accurately compared without bias (SeeDiscussion Section).Twice a week treatment visits then occurred for the next

8 weeks. Each of these treatment visits had identicalprocedures, as described below, except that photographswere taken only once a week. At each treatment visit, thestudy sites were (a) evaluated for adverse events caused bythe previous treatment, (b) located and marked using therigid template indexed with the black-light micro-tattoos,and, if photographs were being taken, (c) clipped ifnecessary to 2 mm hair length for consistent photographsand in accordance with the no!no! hot-wire instructions,and (d) photographed to provide the pre-treatment haircount. Following this, the clinical staff shaved the upper(control) site using a disposable razor and performed thehot-wire treatment on the active (lower) site. If aphotography day, digital photographs were then takenagain to document the post-treatment skin condition. Thestudy sites were then cleaned with an alcohol prep pad toremove the visible ink and a pea-sized amount ofmoisturizing lotion was applied to each site.The hot-wire treatments were performed according to

the device’s Instructions for Use and the followingprocedures: the “wide” tip was used (as recommended forlarge areas like the leg); the treatment level was set to“high” (as recommended for best results); the device wasslid across the skin from a starting point at least 5 cmbelow the marked site to ensure that the filament wire hadheated up prior to reaching the treatment site; four passesover the site were made (the IFU recommends multiplepasses); and, the treatment area was buffed with thesupplied buffing pads to remove any remaining crystal-lized or singed hairs (as recommended in the IFU).At the follow-up visits, which occurred 4 days, 4, 8, and

12 weeks after the last treatment, the same steps wereused as in the treatment visits, however, no shaving or hot-wire treatments were performed. The visits were solely toobserve for adverse events, locate and mark the active andcontrol sites, and photograph them for later assessment ofhair count and characteristics.

The hairs were counted from the photographs by atrained, independent third party. A careful methodologywas followed whereby the raw images were set toconsistent size using Microsoft Picture Manager softwareso that the 3 cm � 3 cm square fills the program windowfrom top to bottom (60% zoom). Manual counting wasconducted in Microsoft Paint software by transferringscreenshots from Picture Manager. Each hair was markedwith the yellow pencil tool as it was identified and talliedwith a digital counter. A 1900 monitor with 1,280 � 1,024resolutionwasused. An example photographprinted at thecounting resolution is shown in Results Section for thereader’s reference.

The photographs were also analyzed for visual changesin the hair fineness (or thickness or gauge) and colorintensity (or lightness or darkness) as well as apparentdensity. The same third party graded the hairs for finenessand color intensity by comparing image pairs consisting ofa baseline photo and a follow-up photo (at 4 days and3 months after the last treatment), where the two photoswere randomly positioned on the “left” or “right”. Theimage pairs were created for both the control (shaving) andthe active (hot-wire) sites. The grader was asked to selectwhich image of the pair had hair that was thinner andlighter. Answers were “left”, “right”, or “cannot tell”.

In addition, the photographs were digitally analyzed byCanfield Scientific to measure the diameter, color value,and number of hairs at the same two time points (4 dayspost-treatment and 3 months post-treatment) compared tobaseline.

Outcome Measures

Efficacy was assessed by quantitative hair counts basedon the photographs at each time point, as well as visualassessment of the photographs for changes in hair gauge(fineness or thickness) and color intensity (lightness ordarkness). The primary efficacy endpoint was the percentchange in hair counts from baseline, which is definedas %cfb ¼ (count � baseline)/baseline � 100. The sameassessments weremade for the control and active sites andresults are reported for control, active, and the differencebetween the two. Using a difference metric is a standardstatistical approach for comparing active and control inthis type of study, because it reduces intra-subjectvariability and provides greater sensitivity, since thedifference is calculated within subject (i.e., the differencebetween control and active is calculated for each subject,showing the relative effect for that subject, and then thestatistics are computed for the subjects as a group).

Statistics were analyzed by a professional, independentstatistician using industry-standard statistical methodsand the commercial SAS package.

RESULTS

Subjects

Twenty-five subjects were enrolled in the study and 22completed the study. Two subjects were discontinued afterthe third treatment due to inability to comply with the

HOT-WIRE HAIR REMOVAL DEVICE 285

required study visits, and one subject was discontinued atthe 4 weeks follow-up visit due to inappropriate behavioras judged by the principal investigator (BSB). No subjectmissedmore than two visits, and all subjects had at least 14treatments. The sample size at all measurement timepoints thus was no less than twenty. The small number ofdropped subjects and missed visits did not meaningfullybias the results.

The demographics of the sample are shown inTable 1. Ascan be seen in the table, 18 of 25 (72%) were women, themedian age was 39 with a minimum of 19 and a maximumof 54, the ethnicity breakout was 92% Caucasian, 4%African-American, and 4% Other.

Photographs

As discussed in Methods, high-resolution standardizeddigital photographs indexed by UV-ink micro-tattoos weretaken at regular intervals throughout the study. Anexample photograph for one subject is shown in Plate 1for the baseline visit. The photograph shows the controlsite (shaving) on the left and the active site (hot-wire) onthe right. The micro-tattoos are located in the top left andbottom right corners (not visible in the regular lightingused in the photograph) and were used to mark the3 cm � 3 cm treatment squares visible in the photograph.As is evident even in the small reprint in Plate 1, the hairsare clearly visible and countable. The large field-of-view isalso useful for verifying that the subject did not performunsanctioned shaving, in that visibly longer hair appearsin regions adjacent to the treatment site. Plate 2 shows asubsection of Plate 1, reprinted at the actual resolutionused for the hair counts for the reader’s reference.

Plate 3 shows a complete time history of photographs forone typical subject, including baseline, pre-tx 3 (meaningjust prior to treatment 3 and 4 days after treatment 2),pre-tx 5, pre-tx 7, pre-tx 9, pre-tx 11, pre-tx 13, pre-tx 15,4 days post-tx (meaning 4 days after the 16th and finaltreatment), 4 weeks post-tx, 8 weeks post-tx, and 12 weekspost-tx. Again, the control site is shown on the left and theactive site on the right. Even at the small size required for

Plate 3, it can likely be appreciated that there is nodiscernible difference between the hot-wire and shavingsites at any time point for this subject.Photographic results at key time points are shown in

Plate 4 for three randomly-selected subjects. The plateshows results at baseline, 4 days after the last (16th)treatment, and at 3 months after the last treatment.Comparing the control site on the left and the active site onthe right, it can be seen qualitatively that there is nomeaningful visible difference between them in the numberor characteristics of the hairs. The number of hairs presentimmediately after completing the 8 weeks course oftreatments appears to be the same with shaving or hot-wire. Further, the number of hairs present 3 months aftercompleting the 8 weeks course of treatment is alsoqualitatively the same between shaving and hot-wire.Analysis of hair length 12 weeks post treatment revealedsubstantially identical outcomes between control andactive treatment zones. Lastly, it can be seen (see alsoPlate 5 for a zoomed view for one subject) that the hair is

TABLE 1. Demographic Characteristics

Characteristic N (%)

Age

Age range 19–54

Median age 39

Mean age 36

Gender

Female 18 (72%)

Male 7 (28%)

Race

Caucasian 23 (92%)

African American 1 (4%)

Other: Bi-racial 1 (4%)

Plate 1. An example photograph for one subject at one time point,showing the control site (shaving) on the left and active site (no!no!hot-wire) on the right. Micro-tattoos are located in the top left andbottom right corners (only visible under black light so not visiblehere) and were used to mark the 3 cm � 3 cm treatment squaresvisible in the photograph. The inset region is printed at thecounting size in Plate 2.

Plate 2. The inset region of Plate 1 reprinted at the size used forthe hair counts and visible assessments.

286 BIESMAN

qualitatively the same color and the same thickness in boththe control and active sites and appears unchanged inthese characteristics from the baseline. This suggests thatuse of the hot-wire device does not produce finer or lighterhair in either the short-term (while treating or just after atreatment course) or long-term (1–3 months post-tx). Aswill be shown in the statistical section below, these

conclusions exemplified in Plate 4 apply to the sample asa whole and in a quantitative manner, not just the threesubjects discussed qualitatively in this section.

Quantitative Hair Counts

Quantitative hair counts were taken from the photo-graphs and statistically analyzed as described in Methods

Plate 3. A complete time history of photographs for a typical subject. Control site (shaving) is on theleft and active site (no!no! hot-wire) on the right. Time points are (a) baseline, (b) pre-tx 3 (meaningpost-treatment 1 and 4 days post-treatment 2 and just prior to treatment 3), (c) pre-tx 5, (d) pre-tx 7,(e) pre-tx 9, (f) pre-tx 11, (g) pre-tx 13, (h) pre-tx 15, (i) 4 days post-tx (meaning 4 days after the 16thandfinal treatment), (j) 4weeks post-tx, (k) 8weeks post-tx, and (l) 12weeks post-tx.No difference isapparent between control and active sites.

HOT-WIRE HAIR REMOVAL DEVICE 287

Section. The results are listed in Table 2 and plotted inFigures 1–4. The tabulation lists the mean hair counts forthe control, active, and difference. Also tabulated are thepercent change from baseline in hair counts for the control,active, and difference. The differences are defined through-out as control minus active. The table includes the samplesize (N and %), mean and standard deviation (SD),median, min, and max as well as t test and WilcoxonSigned Rank for conveying statistical significance.

The mean raw hair counts are plotted in Figure 1 for thecontrol and active sites. The control (shaving) site has amean baseline hair count of 79.4, which remains stable(74.8–84.3) during the 8 week treatment phase and thenclimbs substantially after stopping treatment to 98.8,100.1, and 104.6 at 1, 2, and 3 months post-treatment,respectively. The active (hot-wire) site has amean baselinehair count of 86.0 which remains fairly stable (81.7–95.1)during the 8-week treatment phase and then climbssubstantially after stopping treatment to 104.0, 106.4,and 109.0 at 1, 2, and 3 months post-treatment, respec-tively. Thus, shaving and hot-wire exhibit identical effects,namely that they maintain the baseline hair count whileused regularly (2�/wk) but that substantial and equivalent

regrowth occurs beginning immediately after cessation ofuse. This increase in post-treatment hair counts for boththe treatment and shaving sites is readily explainable byreference to the hair grown cycle. Only anagen hairs (i.e.,hairs in active growth phase) would have been apparentand counted at baseline (since telogen hairs would havebeen shaved off) and throughout treatment (since hairsthat transitioned to telogen during the treatment phasewould have been shaved off or removed with the hot-wiredevice). However, once treatment ceased, the existinghairs would remain as they transition to telogen phasewhile new hairs would be transitioning from telogen toanagen phase, thus increasing the total hair counts overthe 3-month post-treatment assessment period.To better compare the two techniques, Figure 2 plots the

mean of the difference between the control and active sitesat each time point, in a box and whiskers plot. As can beseen, themean difference is near zero at every time point inthe study, indicating that no meaningful difference isdiscernible between the control (shaving) and active (hot-wire) therapy. This can be appreciated visually by noticingthat the mean falls well within the box at every time point,the box definingwhere 50% of the data lie and providing an

Plate 4. Photographic results at key time points for three randomly-selected subjects. Control site(shaving) is on the left and active site (no!no! hot-wire) is on the right. Time points are (a) baseline,(b) 4 days post-tx (meaning 4 days after the 16th and final treatment), and (c) 12 weeks post-tx.Nomeaningful differences can be discerned in quantity, length, color, or thickness of hairs betweenthe control and active sites.

Plate 5. A zoomed view of one image set from Plate 4 at the assessment resolution, illustrating nomeaningful difference in hair properties between treated and control sites. Control site (shaving) ison the left and active site (no!no! hot-wire) is on the right.

288 BIESMAN

TABLE 2. Hair Count Statistics

VISIT Parameter

Raw score (N ¼ 23)

Percent change from

baseline (N ¼ 22)

Left Right Left–right Left Right Left–right

Baseline N (%) 22 (95.7%) 22 (95.7%) 22 (95.7%)

Mean (SD) 79.4 (36.08) 86.0 (38.93) �6.6 (18.62)

Median 81.5 81.5 �9.0

t test 0.1117

Pre Tx 3 N (%) 23 (100.0%) 23 (100.0%) 23 (100.0%) 22 (100.0%) 22 (100.0%) 22 (100.0%)

Mean (SD) 79.5 (35.60) 86.3 (40.23) �6.9 (19.40) 2.1 (9.72) 1.4 (13.27) 0.8 (12.76)

Median 79.0 74.0 �8.0 �0.5 0.4 0.7

t test 0.1036 0.3167 0.6349 0.7820

Pre Tx 5 N (%) 21 (91.3%) 21 (91.3%) 21 (91.3%) 21 (95.5%) 21 (95.5%) 21 (95.5%)

Mean (SD) 80.0 (35.18) 90.9 (41.74) �10.9 (21.37) 3.4 (12.85) 5.4 (17.94) �2.0 (14.51)

Median 84.0 89.0 �10.0 5.6 0.0 0.3

t test 0.0299 0.2332 0.1807 0.5382

Pre Tx 7 N (%) 22 (95.7%) 22 (95.7%) 22 (95.7%) 21 (95.5%) 21 (95.5%) 21 (95.5%)

Mean (SD) 80.8 (33.33) 90.2 (38.51) �9.4 (19.43) 3.3 (16.45) 5.6 (20.69) �2.3 (16.56)

Median 75.5 83.0 �10.5 3.0 0.0 1.3

t test 0.0338 0.3685 0.2263 0.5256

Pre Tx 9 N (%) 22 (95.7%) 22 (95.7%) 22 (95.7%) 21 (95.5%) 21 (95.5%) 21 (95.5%)

Mean (SD) 74.8 (33.77) 81.7 (36.35) �7.0 (17.86) 2.2 (17.10) 1.4 (17.31) 0.8 (14.69)

Median 65.5 75.5 �6.0 0.0 �2.7 �2.4

t test 0.0821 0.5679 0.7141 0.8140

Pre Tx 11 N (%) 23 (100.0%) 23 (100.0%) 23 (100.0%) 22 (100.0%) 22 (100.0%) 22 (100.0%)

Mean (SD) 80.6 (34.72) 91.4 (40.12) �10.8 (20.70) 7.7 (39.70) 10.8 (28.39) �3.1 (23.62)

Median 78.0 85.0 �7.0 �1.5 0.8 �6.1

t test 0.0204 0.3733 0.0893 0.5470

Pre Tx 13 N (%) 23 (100.0%) 23 (100.0%) 23 (100.0%) 22 (100.0%) 22 (100.0%) 22 (100.0%)

Mean (SD) 81.3 (34.79) 92.3 (40.82) �11.0 (20.59) 6.2 (17.29) 11.5 (22.70) �5.3 (16.92)

Median 73.0 78.0 �12.0 6.6 3.2 �7.4

t test 0.0178 0.1097 0.0271 0.1536

Pre Tx 15 N (%) 23 (100.0%) 23 (100.0%) 23 (100.0%) 22 (100.0%) 22 (100.0%) 22 (100.0%)

Mean (SD) 83.8 (35.46) 93.1 (39.73) �9.3 (22.98) 11.7 (25.88) 14.1 (25.14) �2.3 (19.03)

Median 78.0 81.0 �12.0 4.4 6.6 0.1

t test 0.0650 0.0453 0.0157 0.5690

4 Day FU post N (%) 23 (100.0%) 23 (100.0%) 23 (100.0%) 22 (100.0%) 22 (100.0%) 22 (100.0%)

Mean (SD) 84.3 (32.60) 95.1 (38.10) �10.9 (22.52) 14.4 (31.02) 19.3 (31.74) �4.9 (15.56)

Median 81.0 90.0 �12.0 7.4 9.4 �1.9

t test 0.0304 0.0411 0.0095 0.1537

4 Week FU Post N (%) 22 (95.7%) 22 (95.7%) 22 (95.7%) 21 (95.5%) 21 (95.5%) 21 (95.5%)

Mean (SD) 98.8 (35.24) 104.0 (35.31) �5.1 (20.46) 41.0 (40.77) 40.5 (41.33) 0.6 (17.96)

Median 92.5 98.5 �7.0 39.4 19.2 �1.3

t test 0.2522 0.0002 0.0002 0.8895

8 Week FU Post N (%) 21 (91.3%) 22 (95.7%) 21 (91.3%) 20 (90.9%) 21 (95.5%) 20 (90.9%)

Mean (SD) 100.1 (40.54) 106.4 (44.67) �7.5 (19.81) 28.0 (33.52) 30.9 (38.55) 1.4 (26.50)

Median 92.0 93.0 �4.0 32.8 25.4 �1.8

t test 0.0972 0.0014 0.0015 0.8183

12 Week FU Post N (%) 22 (95.7%) 22 (95.7%) 22 (95.7%) 21 (95.5%) 21 (95.5%) 21 (95.5%)

Mean (SD) 104.6 (38.25) 109.0 (43.86) �4.4 (16.86) 37.2 (32.25) 32.2 (31.74) 4.9 (24.18)

Median 94.0 99.0 �3.5 30.3 21.4 0.0

t test 0.2382 <.0001 0.0002 0.3596

Left is the control (shaving) and right is the active therapy (hot-wire).

Fig. 1. Hair counts for control (shaving) and active (no!no! hot-wire) at each time point. Box andwhisker plot with box, whiskers, and crossbar displaying the 25–75 percentile, 5–95 percentile, andmedian, respectively. The mean is plotted as a line overlay.

Fig. 2. Difference in hair counts between control and active (i.e., control–active) at each time point.Box and whisker plot with box, whiskers, and crossbar displaying the 25–75 percentile, 5–95percentile, and median, respectively. The mean is plotted as a line overlay.

290 BIESMAN

indication of the measurement range. The fact that themean difference is almost exclusively negative indicatesthat the hot-wire sites generally had more hair than thecontrol sites. This matched the study staff’s anecdotalobservations that it was generally more difficult to obtain“a perfect shave” using the hot-wire device than the razorand that multiple passes and buffing were required, assuggested in the IFU and as permitted by the studyprotocol. Based upon the t test, the control and active arestatistically indistinguishable (P > 0.05) at seven out of 12time-points. For the five points with statistical signifi-cance, the significance is weak (P ¼ 0.018–0.034) but thedata indicates that the control (shaving) is actuallysuperior to the active (hot-wire) treatment. These pointsall occur during the active treatment phase. In the follow-up phase, the p values are 0.252, 0.0972, and 0.230 at 1, 2,and 3 months, respectively, which indicates that shavingand hot-wire are not statistically different from oneanother in terms of regrowth at any point in time duringthe post-treatment phase of the study.Figure 3 plots themeanpercentage change frombaseline

in hair counts for the active and control sites, that is cfb(%) ¼ (count � baseline)/baseline � 100, which normal-izes the data. It can be seen from the figure that

• The control and active results overlay each other to ahigh degree at all time points.

• The degree of difference at any time point is smallcompared to the measurement range indicated by theboxes at that time point,

• shaving is generally slightlymore effective than the hot-wire device in removing hair during the treatmentperiod but the difference is small,

• both shaving and hot-wire essentially maintain theshaved baseline hair counts throughout the treatmentperiod, although there is a slight increase in hair in thelater phases of treatment,

• there is a substantial increase in hairs in the follow-upperiod after cessation of shaving and hot-wiretreatments,

• there is no measurable difference between shaving andthe hot-wire device in the follow-up period.

The percentage change from the baseline is generally notstatistically significant for either the control or activetreatment during the treatment phase, but after treatmenthas stopped, the hair regrows promptly and the percentagechange from baseline is strongly significant (generally,P < 0.001) for both shaving and the hot-wire device,indicating significant regrowth immediately after stoppingtreatment.

To better compare the normalized active and control toeach other, Figure 4 plots the difference in percentagechange from baseline in hair counts between the active andcontrol sites. This definitive figure shows that the meannormalized difference between the treatments is very closeto zero at every time point, ranging from only�4.9 to 4.9%points, and that any excursions from zero are smallcompared to the measurement range. The t-test P-valuesare high, ranging from P ¼ 0.154–0.890 over all the time

Fig. 3. Percentage change from baseline in hair counts for control (shaving) and active(no!no! hot-wire) Box and whisker plot with box, whiskers, and crossbar displaying the 25–75percentile, 5–95 percentile, and median, respectively. The mean is plotted as a line overlay.

HOT-WIRE HAIR REMOVAL DEVICE 291

points in the study and P ¼ 0.360–0.890 during the 1, 2,and 3month follow-up period. Thus, we can safely concludethat there is no statistically significant difference in haircounts between the hot-wire device and shaving in thisstudy, and, thus, that the hot-wire device did not suppress,inhibit, or reduce hair regrowth, change the rate of orotherwise delay hair regrowth, or effect permanent or long-term hair reduction.

It is also evident from Table 2 and Figures 1–4 that theresults cannot be explained on the basis that only somepersons will respond to the hot-wire treatment. In somecases, the mean can be misleading because certain types oftreatment work well on responders but poorly on non-responders. The data from the present study shows asimilarly even distribution of results between the shavingcontrol and the hot-wire device as evident from the similarmedians and the similarity of the box and whisker plots.The qualitative results were corroborated by quantitativedigital hair counts determined by analysis of digitalphotographs. The results of this analysis are provided inTable 4 which lists the mean percentage change frombaseline in hair count for the control (shaving) and active(hot-wire) sites at the two time-points, as well as theP-value describing the statistical significance of anydifferences between the control and active. As can beseen, the shaving and hot-wire sites both show asubstantial increase (23–52%) in hair counts after cessa-tion of treatment. TheP-values are high at both time points(0.2398 and 0.7148). Thus, the quantitative analysis

further confirms that shaving and hot-wire have nostatistically significant differences in the rate of regrowthof hair, either at the end of 8 weeks of treatment or at3 months after the last treatment.

Visual Hair Characteristics

As noted, the photographs were also analyzed for visualchanges in the hair fineness (or thickness or gauge) andcolor intensity (or lightness or darkness) as well asapparent density. The results of the subjective visualassessments are listed in Table 3, which shows that theblinded grader was unable to discern any significantdifference in either hair thickness or color betweenbaseline and post-treatment in the vast majority of theimage pairs (73–89%). The differences that were reportedbetween image pairs were distributed between the controland active with no statistical significance (x2 > 0.05 in allcases). That is, sometimes the control (shaving) was felt tobe thicker or darker and sometimes the active (hot-wire)was felt to be thicker or darker. Therefore, there was nooverall difference between the treatment (hot-wire) andcontrol (shaving) groups either at the end of the 8 weekstreatment period or the 12weeks post-treatment follow-up.This matched the study staff and principal investigator’slive observations of the subjects, in which they could notobserve any difference in the hair characteristics (or haircount) between the shaved control site and the hot-wiresites. Note that because the sites were adjacent (only a

Fig. 4. Difference in percentage change from baseline between control (shaving) and active(no!no! hot-wire). Box and whisker plot with box, whiskers, and crossbar displaying the 25–75percentile, 5–95 percentile, and median, respectively. The mean is plotted as a line overlay.

292 BIESMAN

3 cm gap), the live comparison is quite simple to performand any cosmetically meaningful differences, had theyexisted, would have been readily apparent.Again, the results of the quantitative analysis of the

digital photographs corroborate the qualitative analysis.The results are provided in Table 4 which lists the meanpercentage change from baseline in thickness and intensi-ty for the control (shaving) and active (hot-wire) sites atthe two time-points, as well as the p-value describing thestatistical significance of any differences between thecontrol and active. As can be seen, the shaving and hot-wire sites show similarly insignificant changes in hairthickness (�1–8%) and hair color (�1–4%) from baseline.The P-values are high for both parameters at both timepoints (0.2398 and 0.7148). Thus, the Canfield Scientificanalysis confirms that the hot-wire device has no statisti-cally significant impact on hair thickness or hair color,either at the end of 8 weeks of treatment or at 3 monthsafter the last treatment.

Adverse Events

Two adverse events were reported during the study:(1) one subject had bruising on the lower leg outside thetreatment area from bumping into a table leg, and (2) onesubject had breast augmentation. Neither was judged to berelated to the study and thus there were no adverse deviceevents related to treatment observed by the study staff orreported by subjects. While burning of the skin by the hot-wire device has been reported on websites such asAmazon.com, these reports are unconfirmed and burnswere not encountered in this study.

DISCUSSION

Photographic analysis of the treated sites using severalparameters including number of hairs present at the end ofthe treatment period, number of hairs present at the end ofthe 12-week follow-up period, and hair color and thicknessgenerated data from which we can draw several con-clusions about the differences between the treatment andcontrol methods. First, the data indicates that the hot-wiredevice provides a shaving-like benefit in the short-term asthere was no difference in hair counts at the end of thetreatment period. That the hair counts were similar at the12 week post-treatment follow up indicates that the hot-

wire device did not inhibit the rate of regrowth or disablehair follicles compared to shaving. That hair length wassubstantially identical at the 12 week follow up pointfurther indicates that the hot-wire device did not inhibitthe rate of regrowth. Finally, analysis of hair color andthickness found no difference between treatment andcontrol groups, suggesting that use of the hot-wire devicedoes not produce finer or lighter hair in either the short-term (while treating or just after a treatment course) orlong-term (1–3 months post-tx).

The current results differ markedly from the only priorpublished study [1]. This can be readily explained byproblems and limitations in the prior study. Some of theselimitations are as follows. First, the prior study wasuncontrolled so it cannot be reliably concluded that the hot-wire device was responsible for the observed effect.Without a control, any number of systematic or randomfactors other than the treatment device (e.g., occasionalshaving by study subjects, difference in photographyquality, or difference in anatomical location) could havecaused the observed reduction in hair counts. Second,there are undefined initial hair count conditions, such thatsome anatomic sitesmay have had recent hair removal andothers none. In fact, the paper describes that clipping was,at least sometimes, required to reduce the hair length to3–4 mm, which indeed indicates that the hair at baselinehadnot been treated for a substantial period of time. This isa problem because it means that the baseline hair countwould include both anagen and telogen hair whereas thepost-treatment (for a razor or for hot-wire) would leavejust anagen hair (stubble) and therefore the count wouldbe just anagen hair. Thus, while it may appear that thetreatment reduced the hair count, in reality it only severeddormant telogen hairs at the surface (just like shavingwitha razor) that would not regrow until the follicle changedphase. Since a significant amount (�50%) of hair is intelogen at any given time, this effect alone could besubstantial, and indeed at least one article indicates thatsuch a shaving bias can still be substantial at 12 weekspost-shaving [2,3]. Third, this artifact is partially built-intothe protocol because of the 2-week hair removal “washout”,whereby the baseline hair counts would have at least2 weeks of growth compared to the first post-treatmentmeasurement which was made immediately after treat-ment, almost guaranteeing a “reduction” in hair counts at

TABLE 3. Subjective Assessments of Hair Characteristics

Can not tell any

difference

Control (shaving)

judged more

Active (hot-wire)

judged more x2Statistically significant

difference?

4 days post-tx

Finer 86.70% 2.20% 11.10% 0.102 No

Lighter 88.90% 2.20% 10.90% 0.102 No

3 months post-tx

Finer 75.60% 12.20% 12.20% 1.000 No

Lighter 73.20% 9.80% 12.80% 0.739 No

HOT-WIRE HAIR REMOVAL DEVICE 293

that time point. Fourth, there is no description of how thetreatment locations were indexed from visit-to-visit, anddespite the concerted effort made to reproducibly identifytreatment locations throughout the study, in fact theimages shown in the paper suggest that the positions werenot repeatable. Since there can bemeaningful variations inhair density at different anatomical locations, this cancause error. Lastly, there are a number of other potentiallimitations in the prior study that contribute to it beinginconclusive, including small sample size, no description ofthe hair countingmethodology, poor camera resolution andvariable photographic technique, unsupervised treat-ments, and no discussion of how it was verified that noshaving occurred during the study.It is also interesting to compare the present hot-wire

device to alternative home-use light-based hair removalsystems. At present, there are three FDA-cleared home-use light-based systems that have an indication forpermanent hair reduction: the TRIA laser (TRIA Beauty,Dublin, CA), the Silk’n or Sens-epil flashlamp (Home-Skinovations, Yokneam, Israel), and the iLight flashlamp(Remington, Wadsworth, OH), as well as systems in theEuropean Union such as the Lumea flashlamp (Philips,Eindhoven, NV) and Smooth Skin flashlamp (Boots,Nottingham, UK). In terms of mechanism of action, thehot-wire device purportedly delivers heat to the follicle viaconduction down the hair shaft from the heat depositedwhen the moving hot-wire severs the surface hair. Incontrast, the light-based systems heat the follicle bydelivering light energy directly into the skin where it isabsorbed by melanin throughout the pigmented hair shaftand follicle to thermally injure the follicle. These light-based systems have limitations, such as only working onbrown or black hair and being safe only for lighter skintypes, but are known to deliver a substantial amount ofenergy to the follicle and have a growing body of peer-reviewed literature establishing safety and efficacy forindicated users (for example, [4–8]).As to whether conducting additional treatments beyond

the 8 week period of this study (which matches the longestperiod recommended in the hot-wire device’s instructionsfor use) might improve the clinical results achieved by thehot-wire device, the current study cannot address thisquestion conclusively. However, the study does show thatthere is no evidence for an inhibition in regrowth rateafter the initial regimen, and it seems unlikely that, if 16treatments produce no effect whatsoever, 24 or 48 or someother numberwould deliver the profound results claimed bythe manufacturer. Follicular injury is presumably neces-sary to affect a structural and/or functional change on thefollicle such that the regrowth rate is impacted. As has beendemonstrated theoretically and experimentally in electrol-ysis and light-based hair removal, (for example, [9–12])such follicular injury is presumably a threshold effect, thatis sufficient energy must be delivered to the follicle to atleast induce minimal injury that may then accumulateover time in repeated minimal injuries. If 16 treatmentsdemonstrate no detectable changes to follicular function, itsuggests that insufficient energy is being delivered to theT

ABLE

4.ObjectiveAssessmentofHair

Characteristics

Hair

count

Hair

size

Hair

intensity

AB

AB

AB

4days

post

12wks

post

4days

post

12wks

post

4days

post

12wks

post

4days

post

12wks

post

4days

post

12wks

post

4days

post

12wks

post

Mea

nch

angefrom

baseline(%

)

23

52

32

49

35

�18

�12

�14

ttest

P-value

(con

trol:active)

0.2398

0.7148

——

0.2202

0.3420

——

0.8723

0.3428

——

294 BIESMAN

follicle to produce even aminimal injury and therefore thatadditional treatments are not likely to eventually accumu-late to ameaningful benefit. It also seems unlikely that theaverage consumer would sustain treatments for such anextended period in any case, given that the treatments aremore time-consuming and cumbersome than shaving anddo not produce any noticeable benefit over shaving duringa long initial regimen.Lastly, the above results and discussion are not to

say that this device might not be a good alternative toshaving for some users. For example, for women who haveunwantedhair on their face andwho are stigmatized by thethought of using a razor blade and unwilling or unable toperform waxing, bleaching, laser hair removal or otheralternatives, the hot-wire device may be a viable option.

CONCLUSION

This clinical investigation of the hot-wire device demon-strated it to be a novel hair removal device that producesshort-term surface hair removal comparable to thatachieved using a standard razor. The hot-wire device didnot induce long-term reduction in hair counts, slow ordelay hair regrowth, alter the thickness or color of hairsthat do regrow, or offer any long-term benefit relative toshaving.

ACKNOWLEDGMENTS

Special thanks to Tobin Island, PhD for his assistance inmanuscript preparation.

REFERENCES

1. Spencer JM. Clinical evaluation of a handheld self-treatmentdevice for hair removal. J Drugs Dermatol 2007;6(8):788–792.

2. Kolinko VG, Littler CM. Mathematical modeling for theprediction and optimization of laser hair removal. Lasers SurgMed 2000;26:164.

3. Kolinko VG, Littler CM, Cole A. Influence of the anagen:telogen ratio on Q-switched Nd:YAG laser hair removalefficacy. Lasers Surg Med 2000;26(1):33–40.

4. Emerson R, Town G. Hair removal with a novel, low fluence,home-use intense pulsed light device. J Cosmet Laser Ther2009;11:98.

5. Alster TS, Tanzi EL. Effect of a novel low-energy pulsed-lightdevice for home-use hair removal. Dermatol Surg 2009;35:483.

6. Wheeland RG. Simulated consumer use of a battery-powered,hand-held, portable diode laser (810 nm) for hair removal: Asafety, efficacy and ease-of-use study. Lasers Surg Med 2007;39:476.

7. Roosen GF, Westgate GE, Philpott M, Berretty PJ, Nuijs T,Bjerring P. Temporary hair removal by low fluence photo-epilation: Histological study on biopsies and cultured humanhair follicles. Lasers Surg Med 2008;40:520.

8. Wheeland RG. Permanent hair reduction with a home-usediode laser: Safety and effectiveness 1 year after eighttreatments. Lasers Surg Med 2012;44:550–557.

9. Altshuler GB, Anderson RR, Manstein D, Zenzie HH,Smirnov MZ. Extended theory of selective photothermolysis.Lasers Surg Med 2001;29:416.

10. AndersonRR, Parrish JA. Selective photothermolysis: Precisemicrosurgery by selective absorption of pulsed radiation.Science 1983;220:524.

11. Grossman MC, Dierickx C, Farinelli W, Flotte T, AndersonRR. Damage to hair follicles by normal-mode ruby laserpulses. J Am Acad Dermatol 1996;35:889.

12. Margolis RJ, Dover JS, Polla LL, Watanabe S, Shea CR,Hruza GJ, Parrish JA, Anderson RR. Visible action spectrumfor melanin-specific selective photothermolysis. Lasers SurgMed 1989;9:389.

HOT-WIRE HAIR REMOVAL DEVICE 295