Figure 1. Kaplan-Meier curves representing the cumulative rate of patients reaching therapy success, defined as removal of the tattoo, with transient hypochromia or darkening being the only adverse effects, in smokers vs nonsmokers.
Figure 2. Kaplan-Meier curves representing the cumulative rates of patients reaching therapy success, defined as removal of the tattoo, with transient hypochromia or darkening being the only adverse effects, in relation to tattoo color.
Figure 3. Kaplan-Meier curves representing the cumulative rates of patients reaching therapy success, defined as removal of the tattoo, with transient hypochromia or darkening being the only adverse effects, in relation to the interval between treatment sessions.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Bencini PL, Cazzaniga S, Tourlaki A, Galimberti MG, Naldi L. Removal of Tattoos by Q-Switched LaserVariables Influencing Outcome and Sequelae in a Large Cohort of Treated Patients. Arch Dermatol. 2012;148(12):1364–1369. doi:10.1001/archdermatol.2012.2946
Objective To analyze variables affecting the treatment course and prognosis of Q-switched laser tattoo removal.
Design Observational prospective cohort study.
Setting The study was carried out in a referral center for surgery and laser surgery in Milan.
Participants Of 397 consecutive patients initially enrolled from January 1, 1995, to December 31, 2010, 352 patients (201 men and 151 women; median age, 30 years) were included in the analysis.
Intervention All patients were treated by the same investigator with Q-switched 1064/532-nm Nd:YAG laser and Q-switched 755-nm alexandrite laser according to tattoo colors. Laser sessions were scheduled at 6-week or longer intervals.
Main Outcome Measures Successful therapy was defined as removal of the tattoo, with no adverse effects other than transient hypochromia or darkening.
Results The cumulative rates of patients with successful tattoo removal were 47.2% (95% CI, 41.8%-52.5%) after 10 sessions and 74.8% (95% CI, 68.9%-80.7%) after 15 sessions. Smoking, the presence of colors other than black and red, a tattoo larger than 30 cm2, a tattoo located on the feet or legs or older than 36 months, high color density, treatment intervals of 8 weeks or less, and development of a darkening phenomenon were associated with a reduced clinical response to treatment.
Conclusions To our knowledge, this study is the first to formally assess prognostic factors for effective tattoo removal by Q-switched laser. Several variables influence response rates and should be considered when planning tattoo removal treatments.
The prevalence of decorative tattoos is rapidly increasing in Western countries, especially among adolescents and young adults.1 As many as 22% of students in the United States have at least 1 tattoo.2,3 Although popular, tattoos are often regretted later in the individual's life. Body image and lifestyle may change and a tattoo, once wanted and liked, becomes embarrassing. The demand for removal is becoming an emerging social trend. In the United States, 28% of adolescents regret their tattoos within the first year4 and 50% of those with tattoos choose to undergo removal procedures as adults.
Various methods, such as surgical excision or salabrasion, have been attempted, but are associated with high scarring risk and unsatisfactory outcomes.5 More recently, the advent of Q-switched lasers (QSLs) has made tattoo removal easier, with a higher rate of cosmetic success.6 Targeted destruction of tattoo ink by selective absorption of a specific wavelength emitted by a nanosecond high-intensity pulse laser forms the basis of QSL tattoo removal.7 After light absorption, ink molecules are partially destroyed and broken into smaller fragments by photoacoustic and pressure waves, as well as by the quick conversion of laser high-energy pulse into heat. Subsequently, ink fragments undergo phagocytosis by macrophages and are removed via the lymphatic vessels.8 Additionally, after QSL irradiation, changes in the optic properties of ink particles occur via thermal and photochemical mechanisms, with an overall lightening of the pigment. These mechanisms usually produce progressive tattoo clearing with little damage to the surrounding skin.6
Despite a sound scientific basis for laser treatment, the clinical results of tattoo removal vary greatly among patients, and it is not possible to guarantee complete clearing of a tattoo in any given patient. Predictive variables affecting the course and prognosis of treatment are poorly defined. The quality and type of pigments; multicolored inks; tattoo layering, size, location, and duration; skin phototype; and personal habits could influence the outcome, but their role remains unclear. The aim of this study was to analyze the course of tattoo removal and prognostic factors for clinical response in a cohort of patients who underwent QSL tattoo removal at an Italian referral center.
This was an observational prospective cohort study. All consecutive patients treated for professional tattoo removal at a referral center for surgery and laser surgery in Milan, Italy, from January 1995 to December 2010 were eligible for entry. The study was reviewed and approved by the Review Board of Centro Studi GISED (Gruppo Italiano Studi Epidemiologici in Dermatologia).
Exclusion criteria included amateur or traumatic tattoos, tattoos applied less than 3 months before study entry, autoimmune diseases, AIDS, diabetes mellitus, and immunosuppressive therapy within 6 months before study entry. Data on patients not completing at least 2 treatment sessions were excluded from the analysis. Patients were instructed about QSL treatment and its adverse effects, and written informed consent was obtained from each person before starting the treatment. All patients received laser treatment from the same investigator (P.L.B), with Q-switched 1064/532-nm Nd:YAG and 755-nm alexandrite laser (VersaPulse; Lumenis) according to their tattoo colors. Black or blue inks were treated with the Q-switched 1064-nm Nd:YAG laser (fluence, 5-5.5 J/cm2; spot size, 3 mm); red, orange, violet, brown, and pink inks were treated with the Q-switched 532-nm Nd:YAG laser (fluence, 5 J/cm2; spot size, 2 mm); and green, light blue, and white inks were treated with the Q-switched alexandrite laser (fluence, 6 J/cm2; spot size, 3 mm). Laser sessions were scheduled at minimum intervals of 6 weeks. The end of the treatment sessions was decided on the basis of the complete (100%) removal of the tattoo or the lack of further improvement as judged by the treating physician after 3 consecutive laser treatment sessions. In all patients, treatment after the laser sessions consisted of topical antibiotics and long-term photoprotection with total block sunscreen.
Standardized en face photographs were taken using a digital camera (EOS 350D; Canon) at the initial visit, before each laser session, and 4 weeks after the final laser treatment session. The initial and final photographs were evaluated by 2 independent observers who were unaware of the number of sessions completed to assess clinical response. Clinical response was classified into the following categories: no improvement, minimal improvement, fair improvement, good improvement, excellent improvement, and removal. For the purpose of this analysis, successful therapy was defined as removal of the tattoo, with the only adverse effects being transient hypochromia or darkening (ie, the change of the original ink color to black).
At study entry, data were collected from each patient, including age, sex, Fitzpatrick skin phototype,9 alcohol consumption, smoking status, and tattoo characteristics including the location, size, age, colors, layering, and density. The following potential sequelae were recorded: transient or permanent hypochromia, hyperpigmentation, textural changes, and coarse scarring. Data concerning tattoo characteristics, treatment modalities, adverse effects, and personal habits were updated regularly.
Color density was established with ×20 magnification using a digital dermoscope (FotoFinder Systems) and defined as sparse when the colors were unevenly distributed with evidence of normal skin within the area. Layering was defined as the coverage of an undesirable tattoo with a new, usually darker and larger, tattoo.
Continuous variables are presented as medians with ranges and categorical variables as numbers with percentages. A Kaplan-Meier product-limit estimator was used to compute the cumulative rates of patients reaching therapy success, ie, removal of the tattoo, during several treatment sessions. Univariate analysis with a log-rank test was used to assess survival differences according to specific categories of interest. Continuous variables were categorized using their tertiles as cutoff points and rounded to multiples of 5, unless otherwise specified. All variables with statistical differences of P ≤ .10 in univariate analysis were included in multivariate analysis to assess independent factors for outcome. Baseline variables and ongoing plus baseline variables were fitted to 2 different models using Cox proportional hazards regression with forward stepwise selection to identify the main factors. The influence of the factors on the outcome was expressed as a hazard ratio along with 95% CI and P values. Analysis of complications during therapy was limited to darkening because of the small number of cases in the other categories. The analysis showed that only patients with tattoo colors other than black developed darkening. Therefore, to determine unbiased estimates of risk factors for darkening, multivariate analysis was performed only on data from these patients. Multivariate logistic regression models with forward stepwise selection were used, and odds ratios with 95% CIs were calculated. Statistical differences were considered significant at P ≤ .05. All analyses were carried out using commercial software (SPSS, version 17.0; SPSS, Inc).
Of the 397 patients initially enrolled, 352 were considered for the analysis; 25 patients were ineligible because of comorbidities or immunosuppressive therapy, and 20 patients were excluded because they did not complete at least 2 treatment sessions. There were 201 men and 151 women, and the median age was 30 years (range, 18-53 years). Details about the tattoos and the treatments are reported in Table 1. The median age of the tattoos was 48 months (range, 6-360 months). The median size of the tattoos was 50.5 cm2 (range, 0.5-980 cm2). The median number of different colors per tattoo was 1 (range, 1-7). Tattoos were located on the face and neck (7.1%), trunk (44.9%), or limbs (48.0%). The median number of treatment sessions was 10 (range, 2-18). The cumulative rates of patients reaching therapy success, ie, removal of the tattoo, was 47.2% (95% CI, 41.8%-52.5%) after 10 sessions and 74.8% (95% CI, 68.9%-80.7%) after 15 sessions.
Results of univariate analysis for clinical response are reported in Table 1. All the variables examined in the univariate model were associated with differences of P ≤ .10 and were included in the multivariate analysis for selection of the main independent predictors. Results of the multivariate analysis are reported in Table 2. Smoking, the presence of colors other than black and red, a tattoo larger than 30 cm2, tattoo location on the feet or legs, a tattoo older than 36 months, a high color density, an interval between treatment sessions of 8 weeks or less, and development of a darkening phenomenon were associated with a reduced clinical response to treatment. Subanalysis of chromatic components in colored tattoos showed that green and yellow were related to a lower clinical response.
Figures 1, 2, and 3 present the cumulative rates of patients reaching therapy success as a function of the number of sessions in smokers vs nonsmokers, in participants with a black or black and red tattoo vs other colors, and according to the interval between treatment sessions, respectively. Table 3 lists the adverse effects observed with treatment. A darkening phenomenon was observed in 17 patients (4.8%), all with colored tattoos. In multivariate analysis, a white color in the tattoo was the only factor significantly associated with an increased risk of darkening (odds ratio, 13.75; 95% CI, 4.10-46.09). No single factor was identified as being associated with other specific adverse effects of treatment.
Requests for tattoo removal have increased steadily in the past 20 years, becoming a relevant medical problem.1-5 Tattoo removal with QSL is considered the criterion standard of treatment.6 Such treatment requires many sessions. Identifying predictors of clinical response may help in the selection of patients and should provide a basis for giving patients an estimate of the anticipated treatment outcome when obtaining their informed consent.
A major finding of this study was the effect of smoking on tattoo removal. The chance of achieving removal after 10 treatment sessions was decreased by 69.7% in smokers compared with nonsmokers. To the best of our knowledge, this is a new finding and will need to be confirmed in independent studies. Another new finding is that prolonged intervals between treatment sessions may improve the clinical response. One advantage of our study is that it examined several variables simultaneously, providing a more complete profile of factors that may affect clinical response after reciprocal adjustment. Some of the factors, such as the presence of colors other than black and red, tattoo size, color density, and the appearance of a darkening phenomenon during treatment, are known and usually considered when planning laser treatment.10 However, we were able to provide a quantitative estimate of the effect of the variables analyzed. For example, the chance of achieving removal after 10 treatment sessions was decreased by 79.5% for tattoos with colors other than black and red. Other variables were less straightforward and represent new findings or are supported by conflicting evidence. In addition to smoking and the length of the intervals between treatment sessions, these variables include a reduced response in older tattoos and in tattoos located on the feet or legs.
Limitations of our study are that it did not consider any measure of patient satisfaction and the evaluation was the clinical judgment of 2 independent observers, without use of any validated criteria. However, this is a general limitation in this area of research. Another limitation is that we restricted our treatment and assessment to professional tattoos, without any comparison between different types of tattoos, including amateur and traumatic tattoos. A further limitation of the study, as well as of most studies in this field, is the lack of histopathologic examination of the tattooed areas before and after laser treatment. We considered such a procedure to be unethical.
The reduced response to treatment observed in smokers compared with nonsmokers is particularly worthy of further examination. Smoking has a long-term chronic effect on many important aspects of inflammation and immune responses.11 Suppression of neutrophil cell spreading, chemokinesis, chemotaxis, and phagocytosis have been described.12 Human macrophages interacting with extracellular matrix proteins modified by cigarette smoke dramatically downregulate the ability of the macrophages to phagocytose apoptotic neutrophils.13 Although the biological mechanisms of ink clearing during laser treatment are intricate, it is well known that ink fragments resulting from the rupture of ink molecules after laser light absorption are phagocytosed by macrophages and carried away via the lymphatic system.14 Approximately 4 weeks after laser-assisted treatment, the remaining particles can again be encapsulated intracellularly within dermal phagocytic cells, contributing to further lightening.14-16 Smoking probably influences clinical responses via the phagocytic system.
Another aspect worthy of consideration in our study is the direct relationship between the length of the intervals between treatment sessions and the clinical response. One reason for such a phenomenon could be the possibility of fully exploiting the mechanism of phagocytosis for ink removal with longer intervals between treatment sessions.
The reduced response observed with older tattoos may be partially explained by the natural history of dermal tattoo ink. The ink particles move deeper into the dermis over time as demonstrated by random biopsy studies17 and by the common clinical observation that tattoos become duller and more indistinct as they age. This migration of ink particles deeper into the dermis makes them more difficult to efficiently target with laser lights owing to their limited lengths of extinction. Moreover, the ink particles in older tattoos are found in dermal fibroblasts beneath a layer of fibrosis, which could represent an optical shield against the laser light, increasing scattering and deflecting the light from the target.17
To summarize, our study is, to the best of our knowledge, the first to formally analyze prognostic factors for the effective removal of tattoos by QSL. Smoking, the presence of colors other than black and red, larger tattoo size, tattoo location on the feet or legs, older tattoos, higher color density, appearance of a darkening phenomenon during treatment, and shorter treatment session intervals influenced the response and should be considered when planning tattoo removal treatments. The effects are substantial for some of these variables. Our data could represent the basis for developing predictive models of tattoo removal that are applicable in clinical practice.
Correspondence: Luigi Naldi, MD, Centro Studi GISED, Via Garibaldi 13, 24100 Bergamo, Italy (firstname.lastname@example.org).
Accepted for Publication: July 5, 2012.
Published Online: September 17, 2012. doi:10.1001/archdermatol.2012.2946
Author Contributions: Dr Naldi had full access to all study data and takes responsibility for data integrity and analysis accuracy. Study concept and design: Bencini and Naldi. Acquisition of data: Bencini, Tourlaki, and Galimberti. Analysis and interpretation of data: Bencini, Cazzaniga, Tourlaki, and Naldi. Drafting of the manuscript: Bencini, Cazzaniga, Tourlaki, Galimberti, and Naldi. Critical revision of the manuscript for important intellectual content: Bencini, Cazzaniga, and Naldi. Statistical analysis: Cazzaniga. Administrative, technical, and material support: Bencini, Tourlaki, and Galimberti. Study supervision: Bencini and Naldi.
Conflict of Interest Disclosures: None reported.
Create a personal account or sign in to: