A, Light micrograph of an unstained6-µm paraffin section showing the presence of large globular green pigment(original magnification ×100). B, Electron micrograph of a treated regionof the same tattoo shown in A. There are large lysosomal structures that containround pigment particles (arrowheads) of varying sizes. Most of these particlesare much smaller than the pigment particles found in the untreated areas (barindicates 0.2 µm). All particles retain a round shape. C, Light micrographof an unstained 6-µm paraffin section showing large green crystallinestructures (original magnification ×100). D, Electron micrograph ofa treated region of the tattoo shown in C. Note the large electron-dense structuresas well as the smaller similarly dense structures around them (arrowheads).C indicates crystal (bar indicates 0.2 µm).
Micrographs from a biopsy specimenof a treated tattoo that changed from flesh color to black on treatment. A,Light micrograph of an unstained 6-µm section (unseen epidermis [E]at top) showing changed pigment in the upper dermis and unchanged pigmentin the deeper dermis (bar indicates 20 µm). B, Electron micrograph ofan "unstained" (osmium tetroxide only) thin section from superficial dermisshowing great variability in the size of the pigment particles (arrowheads)within the lysosomes. Most are, however, much smaller than the particles seenin C (bar indicates 0.2 µm). C, Electron micrograph of the tattoo pigmentfrom the deeper dermis showing unchanged pigment particles (bar indicates0.2 µm).
Ross EV, Yashar S, Michaud N, Fitzpatrick R, Geronemus R, Tope WD, Anderson RR. Tattoo Darkening and Nonresponse After Laser TreatmentA Possible Role for Titanium Dioxide. Arch Dermatol. 2001;137(1):33-37. doi:10.1001/archderm.137.1.33
Copyright 2001 American Medical Association. All Rights Reserved.Applicable FARS/DFARS Restrictions Apply to Government Use.2001
To examine relationships between chemical composition, biopsy findings,and clinical outcome in laser-treated tattoos.
Observational nonblinded retrospective study.
University-based dermatology clinic and private practice.
Twenty patients who underwent biopsy of laser-treated tattoos.
Main Outcome Measures
Biopsy specimens were analyzed after laser treatment, and the depthsof changed particles were recorded. Ultrastructure of the changed particleswas examined by electron microscopy. Presence of inorganic chemicals was determinedby x-ray diffraction. Correlation between x-ray diffraction, microscopy, andclinical response was attempted.
Of the 20 tattoos, 7 lightened, 9 failed to change, and 4 darkened afterlaser treatment. There was a significant association between presence of titaniumdioxide and poor response to laser therapy. Microscopic studies showed variablechanges in the ink particles, but there was a trend toward residual deep greenpigment in the resistant tattoos. Also, round dark stippling was observedsuperficially in the darkened specimens.
Titanium is overrepresented in tattoos that respond poorly to lasertreatment. Further studies are necessary to show whether this metal is theprimary cause of this poor response.
MANY TATTOOS are resistant to laser treatment, particularly those containingblue, purple, yellow, green, and flesh-toned dyes.1,2For example, while some green tattoos clear completely after only 2 to 3 treatmentsessions, others do not lighten or may even become darker.
The mechanism by which tattoos darken after laser treatment is not completelyunderstood. One factor may be the laser-induced reduction of metallic compoundsused in certain dyes. A potential offender is titanium dioxide (TiO2), which is an increasingly popular white ink used to enhance the brillianceof tattoos.3 Titanium is most commonly foundin green, white, and flesh-colored tattoos; however, it has been identifiedin tattoos of almost any color.4 To explorea possible association between TiO2 and tattoo response to lasertreatment, we examined the metallic composition of a series of treated tattoosin this pilot study.
Biopsy specimens were obtained in this retrospective analysis from consentingpatients participating in tattoo treatment protocols with the use of Q-switchedruby (Spectrum Medical Technologies, Natick, Mass) or Nd:YAG (Con-Bio, Livermore,Calif) lasers. Each patient was treated by an experienced dermatologist witha protocol expected to achieve an adequate clinical response. Green tattooswere treated with the ruby laser with fluences of 4 to 7 J/cm2and spot sizes of 5 to 6 mm. Red tattoos were treated with a frequency-doubledNd:YAG laser with fluences of 2 to 4 J/cm2 and spot sizes of 2to 3 mm. Black tattoos were treated with an Nd:YAG laser at 1064 nm with fluencesof 4 to 8 J/cm2 and a spot size of 2 to 3 mm. Patients receivedat least 6 treatments with 1 or a combination of lasers before biopsy. Tattoocolors are summarized in Table 1.A total of 20 patients were included in the study.
For most of the tattoos, only posttreatment biopsy specimens were available.Biopsy samples were handled in 1 of 2 ways. Some tissue was processed forparaffin embedding and stained with hematoxylin-eosin. Representative sectionswere left unstained and were used to assess the density of ink particles anddepth of microscopic lightening induced by laser. The rest of the tissue wasprocessed for electron microscopy. Specimens were fixed in 4% glutaraldehydein 0.1-mol/L cacodylate buffer, dehydrated, and embedded in epoxy resin. Postfixationin osmium tetroxide was omitted to enhance the x-ray analysis. Sections 1µm thick were examined either unstained or stained with toluidine blueO. Thin sections were stained with saturated uranyl acetate and Sato leadstain or left unstained and examined in an electron microscope (CM10; Phillips,Hillsboro, Ore).
All specimens underwent x-ray diffraction studies for detection of metallicelements. Paraffin-embedded specimens were cut at 10 µm and mounteddirectly on an aluminum stub for use in scanning electron microscopy. Paraffinwas removed from the sample with xylene, and a layer of carbon was appliedto the surface of the section. These samples underwent x-ray microanalysison a scanning electron microscope (Amray 1400; Amray, Bedford, Mass) withan x-ray detector (Kevex, Valencia, Calif). Samples processed in epoxy resinwere cut at 90 nm and placed on a carbon and polyvinyl formal–coatedgrid for examination with a transmission electron microscope (Philips Bio-twin;Phillips) equipped with an x-ray detector. Both methods produced spectralsignatures for elements present within the tissue.
A Fisher test was used to determine if there was a significant correlationbetween presence of titanium and tattoo resistance and/or darkening. We excludedtattoos found to contain iron oxides from the data analysis, as they are knownto darken with treatment.
To observe the direct effects of laser irradiation on titanium, a 5%TiO2 cream (Ti-Screen Natural; Pedinol, Farmingdale, NY) was irradiatedwith a Q-switched Nd:YAG laser (Schwartz Electro-optics, Orlando, Fla) witha fluence of 7 J/cm2.
Tattoo colors and metal compositions are summarized in Table 1. Of the 20 tattoos studied, 7 lightened significantly throughoutthe whole tattoo, 9 were at least partly resistant, and 4 became darker. Ofthe 13 tattoos that responded poorly, all except 2 contained titanium, whereasin the 7 tattoos that lightened considerably after laser treatment, only 2contained titanium. With the use of a Fisher exact test, there was a significantassociation between the presence of titanium and a poor response to lasertherapy (P = .02).
In general, tattoos that failed to lighten showed higher ink densitythan those that responded favorably (Table2). All samples of colored (nonblack) pigment appeared to be heterogeneousby light microscopy, showing mixtures of colored and indeterminate brown-blackgranules. Within the group of resistant green tattoos, 2 types of pigmentgranules were seen: (1) a large and rounded globular granule (2-5 µmin diameter; Figure 1A) that tendedto break down into smaller round particles of various sizes when treated (Figure 1B) and (2) crystalline granules (5-10µm in diameter; Figure 1C)that showed a tendency to "splinter" with therapy (Figure 1D). Changes in granule shape were less conspicuous withincreasing depth of the samples. In both types of specimens after treatment,there was a transition from translucent brownish granules (0.5-1 µm)with small black inclusions (like a stippling) located superficially, to larger,more opaque bright-green granules (1-3 µm) appearing unchanged by thelaser starting about 500 µm deep in the dermis (Figure 1A). Scattered among the deeper granules were stippled aggregatesof smaller 1-µm granules similar to those noted in the superficial biopsyregions. The number of these smaller dark dots appeared constant from 500to 1500 µm deep in the skin. However, more superficially, there wasa progressive decrease in their number such that they were rarely observedin the uppermost sections of the specimen (200 µm deep).
By transmission electron microscopy, resistant green tattoos showedsome cells with lysosomes containing pale material with a few dark particlesin treated areas (Figure 1B). (Note:In this study, we defined particle as the smallestidentifiable structure on electron microscopy. This is distinguished from granule, which we defined as the smallest structure observedon routine light microscopy, usually 0.5 to 10 µm in diameter, ie, theparticle-containing lysosome.) Other specimens showed more crystalline particlesevenly distributed in the cytoplasm (Figure1D). In both cases (globular and crystalline), areas from the deeperdermis, the presumed untreated regions, showed larger, more homogeneous particles.
In the 4 tattoos that underwent darkening, light microscopy showed largegranules (1-3 µm) deeper in the dermis that appeared unaltered by lasertreatment. Exclusively in a thin band (about 100 µm thick; Figure 2A) in the superficial parts of thespecimens (about 200-300 µm in depth), smaller (1 µm), darkerround bodies appeared as stippling within these granules. In the more superficial("treated") areas of the specimens, these small bodies were more loosely distributed,as compared with a more aggregated, denser, clumped distribution in the untreateddeeper regions of the sections. By transmission electron microscopy, the appearancewas similar, with more heterogeneity seen superficially, where specimens showeda mixture of apparently unchanged particles (approximately 500 nm in diameteron electron microscopy) and smaller (50-100 nm) and "changed" darkened particles(Figure 2B and C).
Mercury was found in the one red specimen that darkened. The histologicfindings were similar to those of the other 3 darkened tattoos, with the exceptionthat in the superficial (ie, treated) dermis there were both larger reddishgranules and smaller, presumably changed, granules.
Microscopically, in treated tattoos that lightened considerably (mostof which were black), there was smudging and a light-brown color superficially,with little or no peppering. Deeper in the dermis, homogeneous black granuleswere noted. By transmission electron microscopy, treated parts of the specimensshowed a mixture of electron-dense and electron-lucent particles, with manyof the electron-lucent (changed) particles showing slightly increased size(60-100 vs 40 nm) compared with their unchanged electron-dense (native) counterparts.
Laser irradiation of the 5% TiO2 cream resulted in a dramaticimmediate color transformation from bright white to bluish-black.
Mechanisms for tattoo removal after Q-switched laser irradiation arenot clearly understood. It is presumed that the absorption of short-pulseenergy produces high temperatures in the ink particles, resulting in deathof ink-laden phagocytic cells.5 Tattoo clearingmay occur via particle fragmentation, phagocyte cell death and subsequentegress via lymphatics, transepidermal elimination, or intrinsic optical propertychanges in the pigment granules.6- 8Resistance to laser treatment may be related to early rephagocytosis of particlesby fibroblasts, excessive amounts of ink, depth of particles, failure of temperatureand pressure stresses to alter particles, and electrochemical changes in thevalence state of inorganic dyes subjected to high-power laser irradiation,as described below.
Tattoos containing ferric oxide, a brown-red ingredient widely usedin red, pink, and flesh-colored tattoos, have been reported to result in ablack discoloration when treated with the Q-switched ruby laser. The mechanismis thought to involve the reduction of ferric oxide, which is rust colored,to ferrous oxide, which is jet black.9 A similarphenomenon may be involved in white and other iron-free inks that containtitanium. In the untreated tattoo, titanium is in the TiO2 form,which is bright white. High-intensity laser irradiation has been shown toresult in the reduction of Ti4+ to Ti3+, which is responsiblefor the blue color.10 We demonstrated this colorchange by irradiating a titanium-enriched sunscreen.
Tope et al11 studied tattoo ink gels containingTiO2 and iron oxide and found that changes were both wavelengthand pulse duration dependent. Most important, they were unable to induce tattooink darkening in tattoos with pulse durations greater than 1 millisecond,suggesting that a threshold power density is required for tattoo ink darkening.
Interestingly, there is evidence to support spontaneous bleaching ofink darkening in tattoos containing titanium.12This observation is supported in vitro, where, on exposure to air, there is"bleaching" of the blue discoloration after 3 to 4 months.10
White ink, composed of about 95% TiO2, is commonly used tobrighten green, blue, yellow, and purple tattoos. The resulting mixtures aremade such that green inks, for example, are typically composed of about 30%to 40% TiO2. (These data were obtained in a recently completedpilot study with a newer energy dispersive x-ray microanalysis device. [A.L. Timko, MD, C. H. Miller, MD, F. B. Johnson, MD, E.V.R., unpublished data,1999]). Therefore, it is possible that tattoos with a large titanium fractionturn deeply black with Q-switched laser treatment, whereas other tattoos withsmaller titanium "burdens" darken so little with laser irradiation that theyappear grossly not to lighten. With repeated treatments of green tattoos,one possible scenario is that that the green "organic" portion, at least superficially,is being eliminated, whereas the titanium portion is darkening.
One of the difficulties in this study is correlating the electron microscopyand light microscopy results with the clinical findings. For example, we havenoted that, in resistant green biopsy specimens, there is considerable remaininggreen ink at the base of the gross specimen, with less gross green color superficially.This was confirmed histologically, as we saw more large green granules deeperin the treated specimens. This observation suggests that simply the applicationof more surface fluence would remove the remaining deep green pigment. Higherfluences would presumably overcome depth-related attenuation of the laserbeam, as well as any competition for light absorption by the TiO2intermixed in the deeper remaining green pigment. Accordingly, the deepergreen pigment would be eliminated while any mild ink darkening (graying) shouldbe grossly undetectable.
Presumably, whatever is responsible for ink darkening would tend tooccur higher in the specimen, where the subsurface fluence is greatest. Inthe tan tattoo that darkened, we did note increased stippling on light microscopy,which might explain the gross darkening; perhaps this was TiO2being altered, as we did not find this stippling deeper in the specimen. Onthe other hand, we were unable to determine a reliable histologic correlatefor resistance in green tattoos, as stippling was noted in both superficialand deep areas of the specimen and was more prevalent deeper in the specimen.Without knowing where the TiO2 is ultrastructurally (and, justas important, its ultrastructural appearance before and after treatment),one can only speculate as to actual physical changes occurring within thespecimen during laser irradiation.
The limitations of the study are several. First, the tattoos were notall subjected to the same treatment protocol; however, every tattoo was treatedby a highly experienced dermatologist with a protocol expected to result inoptimal clearing. Second, tattoos were not matched for anatomic site or ageof the patient, skin type, age of tattoo, depth of pigment, density of pigment,or ink color. Third, although our analysis detected the presence of all constituentmetals in each tattoo, it was not designed to find the relative amount ofmetal present in each specimen. Optimally, one would search for matched controlsbetween the 3 groups (resistant, darkening, and lightening) with respect tolocation, age, and, most important, ink color. In this retrospective study,we initially limited enrollment to tattoos that were either resistant or darkenedwith treatment. Subsequently, we added a "good responder group" to serve asa control. Unfortunately, the good responder group could not be matched forcolor, as we observed that a disproportionate number of green tattoos respondedpoorly and that black tattoos generally responded well. Thus, color is a confoundingvariable; unfortunately, most green tattoo inks contain TiO2. Itis interesting that clinicians generally note that green tattoos are the mostdifficult to treat. Often a green tattoo lightens about 50% after 2 to 4 treatments,only to fail to lighten considerably more with even up to 20 additional treatments.The lack of response in green tattoos is interesting, as there is evidencethat red light (alexandrite or ruby laser) is well absorbed by green inks.13 It is unknown, however, whether TiO2 isplaying a role in resistance or, alternately, whether there is simply a higherdamage threshold for the organic azo-dyes commonly used as green tattoo inks.
Green tattoos are not singularly resistant to treatment; other ink colorsare also characteristically resistant. For example, we have found that purpleand yellow inks are difficult to treat, even though they, too, show at leastfair absorption of one or another commonly used Q-switched laser; these inkcolors also usually contain TiO2. Probably, green has become sucha nuisance color largely because of its prevalence. It is our experience thatthe 3 most common tattoo colors are black, green, and red. Of these, red andblack respond well in most instances, and these inks typically contain notitanium.
The ultimate test of TiO2 as an independent risk factor forresistance would be a comparison of green inks with and without TiO2. However, although our study incorporated 6 colors, we believe thatthe overrepresentation of TiO2 in poor responders and darkeningtattoos is due at least partly to elemental composition and not solely tattoocolor. Moreover, at least for the darkening tattoos, the microscopic correlation(increased stippling and heterogeneity of granules after treatment) and presenceof TiO2 make a compelling argument for titanium's role in the grossobservations. The results of this study should encourage other investigationsof tattoo ink composition and its relation to laser treatment success. A larger,more systematic study is necessary to individually assess the roles of titanium,ink densities, ink depth, and ink color. Experiments in animal models withthe placement of controlled tattoos with known concentrations of inks andmetal components would be invaluable. If additional compounds are identifiedthat darken with laser irradiation, it may be prudent for manufacturers tosearch for alternative dyes so that "permanent" but "laser-friendly" tattoosmight be available in the future.
Accepted for publication July 27, 2000.
Corresponding author and reprints: E. Victor Ross, MD, Departmentof Dermatology, Naval Medical Center, 34800 Bob Wilson Dr, Suite 5, San Diego,CA 92134-1005 (e-mail: email@example.com).