Introduction

Tattooing, which dates back to 12,000 BC [1], is a rising sociocultural phenomenon. Contemporary studies estimate that as many as one fourth of young to middle-aged adults in the USA have at least one tattoo [2, 3]. However, many tattoos are applied impulsively and later cause the individual embarrassment, low self-esteem, and stigmatization [4]. As many as 17 % of individuals with tattoos consider removing them [2].

Tattoo removal is probably as old as tattooing itself and was first documented in 543 AD, using a process resembling salabrasion [5]. In 1965, Goldman first reported the use of lasers for removing a tattoo [6], and in 1967, the same group reported their 3-year experience with the Q-switched (QS) ruby laser for tattoo removal with minimal scarring [7].

The laser beam’s ability to remove tattoo pigment can be explained with two theories: (a) according to the conventional photothermal theory, light energy is absorbed by the target pigment, causing it to break up into microparticles. The optimal laser pulse duration needs to be less than or equal to the time it takes the heat to spread to the surrounding tissue from the target pigment particle [8]; (b) the newer photoacoustic theory suggests that the laser beam generates strong acoustic waves within the pigmented particle, leading to its breakup [9].

Since the 1980s, QS lasers, with a pulse duration of 5–100 ns, have been considered the mainstay of tattoo removal. However, most of the particles in tattoo inks measure approximately 100 nm in diameter [10], corresponding to a thermal relaxation time of less than 10 ns. Therefore, according to photothermal theory, picosecond-domain lasers would be expected to be more effective than nanosecond-domain lasers. This is also true according to the photoacoustic theory, since decreasing the laser’s pulse duration results in increased wave intensity [9].

The use of picosecond-domain lasers to remove tattoos has been described in several studies. The objective of this review was to systematically review the evidence for the effectiveness and safety of picosecond-domain lasers for tattoo removal.

Methods

A systematic review was conducted and reported in accordance with the PRISMA statement [11] and was registered with the PROSPERO international prospective register of systematic reviews (CRD42015023458).

Eligibility criteria

Studies that met the following criteria were included in the analysis:

  • Language: English.

  • Population: human patients and animal models with tattoos that underwent picosecond-domain laser therapy for tattoo removal.

  • Study design: All relevant studies were included, regardless of their study design.

Outcomes

  1. 1.

    Primary outcome

    • Over 70 % clearance of tattoo pigment.

      When a higher percentage of clearance was defined in the article, it was recorded as more than 70 % clearance and indicated as such in the results.

  2. 2.

    Secondary outcomes

    • 90–100 % clearance of tattoo pigment.

    • Number of laser sessions required.

    • Adverse effects, including acute bulla formation, hyperpigmentation and hypopigmentation, and pain level during the procedure.

Literature search

One reviewer (O.R.) searched the Cochrane Central Register of Controlled Trials (CENTRAL, from inception until July 2015), using the terms “laser” and “tattoo,” and PubMed, using the MeSH heading “lasers” and free text words. The PubMed search strategy is described in Appendix 1. The ongoing trials registry of the US National Institutes of Health (www.clinicaltrials.gov) was screened for additional trials that published results. Reference lists of included trials were searched for relevant publications. Authors were contacted for missing data and clarifications.

Study selection and data extraction

One reviewer (O.R.) screened the titles and abstracts of all retrieved articles. When the titles and/or abstracts suggested potential eligibility for the review, the same author screened the full texts and extracted the data into a predefined electronic form. A second reviewer (L.A.) checked the extraction from all included studies. Disagreements were resolved through discussion with a third reviewer (D.M.).

Risk-of-bias assessment

The risk of bias was assessed using the criteria generated by Downs and Black [12]. The question on power was simplified to a check of whether or not the study had conducted a statistical power calculation. The maximum achievable score for each of the subscales was 10 for reporting, 3 for external validity, 7 for internal validity—bias in measurement of intervention and outcomes, 6 for internal validity—confounding (selection bias), and 1 for power, for a maximum score of 27.

Data analysis and synthesis

Since the study designs, participants, interventions, and reported outcomes varied markedly, we focused on describing the studies, their results and limitations, and the qualitative data synthesis.

Results

Our search yielded a total of 536 articles (Fig. 1). Eight studies fulfilled the eligibility criteria, including two non-randomized controlled trials in animal models [13, 14] and clinical studies with 160 participants bearing 182 tattoos. The clinical studies included one non-randomized controlled trial [15], two non-controlled prospective trials [16, 17], two case series [18, 19], and one retrospective case-control study that only investigated side effects [20]. The trials’ characteristics are detailed in Tables 1 and 2. In general, the studies’ Downs and Black scores [12] were low to medium, ranging from 8 to 19/27 (Table 3). The definitions of the outcome measures varied among the studies.

Fig. 1
figure 1

Flowchart

Full size image
Table 1 Characteristics of included trials
Full size table
Table 2 Outcomes of included trials
Full size table
Table 3 Downs and Black grading for included trials
Full size table

Seven trials explored the use of either 1064 nm (Nd:YAG) [15], 755 nm [16, 18, 20], 795 nm [13], 758 nm [14], or frequency-doubled 1064/532 nm [17] lasers for the treatment of mainly black and blue tattoos. One of the trials explored the use of frequency-doubled 1064/532-nm lasers for the treatment of yellow tattoos [19]. One trial [15] used a laser with a pulse duration of 35 ps, while the rest of the trials used lasers with pulse durations of 350–900 ps. Follow-ups ranged from 1 day to 16 weeks after the last treatment.

Animal models

Hairless guinea pigs, which are used in studies involving human tattooing due to the superior tattoo quality and retention [21], served as the model for both animal studies included in this review. Both reported a higher effectiveness for picosecond-domain lasers compared to nanosecond-domain lasers in black tattoo clearance. This difference was either not quantitatively assessed or was inconclusive.

The first study [13] compared the effectiveness of a single treatment with either a 795-nm titanium/sapphire 500-ps laser or a 752-nm 50-ns Alexandrite laser for 24 black tattoos on four guinea pigs. The investigators reported a “significantly greater response” with the 795-nm picosecond laser. The clearance ratio was not specified in the full text, so the credibility of this finding cannot be determined. Fifty percent of tattoos treated with the 795-nm picosecond laser in that study showed “almost complete clearance” of the pigment.

The second study [14] compared the effectiveness of a single treatment with a 758-nm 500-ps laser versus a 755-nm QS 30–50-ns Alexandrite laser for the removal of 72 tattoos made of either black India ink (carbon) or iron oxide (midnight brown/black). Mean (±SD) clearance scores of the black ink tattoos treated with low, medium, and high-fluence picosecond lasers were, respectively, 7.83 ± 1.5, 7.81 ± 2.08, and 7.57 ± 1.91, corresponding to around 70 % pigment clearance. The value for the nanosecond-domain laser was 6.44 ± 1.65, corresponding to around 60 % clearance. This difference was statistically significant (p = 0.0015, 0.014, 0.028).

Treatment of iron oxide tattoos with the nanosecond-domain laser was associated with a mean (±SD) clearance score of 2.64 ± 1.93, corresponding to around 30 % pigment clearance. Even lower scores, corresponding to around 10–30 % clearance of tattoo pigment, were seen with the picosecond laser (2.33 ± 2.12, 2.08 ± 1.76, and 1.08 ± 0.94 for low, medium, and high fluence, respectively). The difference between nanosecond and picosecond lasers was statistically significant only for the high-fluence picosecond laser (p < 0.001). The difference in the ability of the lasers to clear different ink types was attributed to the fact that only carbon can undergo combustion and to the “steam-carbon” reaction generated by the laser.

Human trials

Over 70 % clearance

755-nm picosecond-domain Alexandrite laser

Two trials explored the effectiveness of the 755-nm picosecond-domain Alexandrite laser [16, 18]. The first [16] treated 12 patients with black and blue ink tattoos with an average of 4.25 sessions (range 2–10) at 6-week intervals, and the second [18] treated 12 blue or green tattoos, with 1–2 sessions. In both trials, all patients showed more than 75 % clearance of the tattoo pigment.

1064-nm picosecond-domain laser

One trial [15] with 16 patients compared the effectiveness of black tattoo removal between the 1064-nm picosecond-domain laser and the 1064-nm nanosecond-domain laser. Half of each tattoo was treated during four sessions with one laser and the other half during four sessions with the other laser. More than 70 % clearance was observed in 11 tattoos using the picosecond laser (69 %) compared to none of the tattoos using the nanosecond laser (p < 0.001).

Frequency-doubled 1064/532-nm picosecond-domain laser

Two trials explored the use of a frequency-doubled 1064/532-nm picosecond-domain laser. One trial [19] treated six yellow tattoos with up to 10 sessions at 6–8-week intervals between them. They found that six of the six (100 %) tattoos had over 75 % clearance of yellow ink. A second trial [17] exploring the treatment of 31 mostly black (but also partially green, blue, purple, red, and yellow) tattoos did not report the proportion of tattoos achieving over 70 % clearance. Nevertheless, they found an average clearance of 79 ± 0.9 % after an average of 6.5 sessions. Different ink colors had different average clearance rates as follows: black 92 % clearance, green 65 %, purple 78 %, blue 43 %, red 80 %, and yellow 85 %.

90–100 % clearance

Clearance of 90–100 % was reported in four of the six human trials.

Two explored the use of the 755-nm picosecond-domain lasers. Clearance of 90–100 % was reported in 7/12 (58 %) of black and blue ink tattoos after 2–10 sessions [16] and in 8/12 (67 %) blue and green tattoos after 1 session [18]. In a comparative study on the use of 1064-nm picosecond-domain lasers for the removal of black ink tattoos [15], 90–100 % clearance rates were observed in 2/16 patients treated during four sessions with the picosecond-domain laser compared to none of the patients treated with the nanosecond laser [15]. Others reported that one to five treatments with the frequency-doubled 1064/532-nm picosecond-domain laser cleared 90–100 % of yellow ink tattoos in four of the six patients (67 %) [19]. A second trial [17] with the frequency-doubled laser at the same wavelength reported an average 91.6 ± 5.4 % clearance of black pigment after up to seven sessions.

Number of laser sessions required

The number of laser sessions varied greatly from one trial to another. Two trials used just one session [18, 20]. One trial used a fixed number of four sessions [15], and three trials allowed up to 7–10 sessions, with averages of 3.3 [19], 4.25 [16], and 6.5 [17]. This was not related to the tattoo color.

Adverse effects

Mild transient side effects such as edema, erythema, and pinpoint bleeding were frequently reported. No scarring was reported in any of the trials.

Acute bulla formation

There were variable reports of bulla formation for each laser session, ranging from 0 to 50 %. Rates per patient for the 755-nm picosecond laser in different studies were 0 [16], 10 % [18], and 32 % [20]. In one trial [20], adding a fractionated CO2 laser to the 755-nm laser treatment reduced the rate of bulla formation from 32 to 0 %. The frequency-doubled 1064/532-nm laser was associated with 0 [17] and 50 % [19] bulla formation. The 1064-nm picosecond laser was not associated with bulla formation [15].

Hyperpigmentation and hypopigmentation

Five of the six trials using the picosecond laser reported local pigmentation disorders. One of the trials reported several cases of transient dyspigmentation after treatment with a 755-nm picosecond laser [18], while another reported hypopigmentation in 3/12 patients (25 %) and hyperpigmentation in 2/12 patients (16.7 %) [16]. Treatment with the 1064-nm picosecond laser was associated with hypopigmentation in 1/16 patients (6.25 %), and treatment with the frequency-doubled 1064/532-nm picosecond laser resulted in transient hypopigmentation in 1/6 patients (16.7 %) in one trial [19] and dyspigmentation in 6/31 tattoos (19.3 %) in another [17].

Pain level during procedure

Three out of six trials described the pain level reported by patients during the procedure on a scale of 1 to 10. Two studies reported pain levels of 1.08/10 (755-nm laser) [18] and 1.3/10 (1064/532-nm laser) [19] with local anesthesia. One trial reported a pain level of 4.5/10 (755-nm laser) [16] without any mention of local anesthesia.

Discussion

In this systematic review, we aimed to evaluate the effectiveness and safety of picosecond-domain lasers for tattoo removal. More importantly, we aimed to compare its effectiveness and safety with the mainstay laser methods of tattoo removal—the nanosecond-domain lasers. Compared to previous narrative reviews that included fewer studies, this is the first systematic review to include eight trials.

Lasers included in this review had wavelengths of 755, 758, 795, 1064, and 1064/532 nm and a pulse duration of mainly 350–900 ps, which could also be referred to as subnanosecond lasers. They were found to be effective mainly for black and blue ink tattoos in achieving over 70 and 90–100 % clearance of tattoo pigment. However, the response rate varied greatly between trials that utilized different picosecond lasers and different trial models. This was true especially for 90–100 % clearance rate.

QS lasers, with a pulse duration of 5–100 ns, are considered the mainstay of tattoo removal therapy. Previous reports exploring the use of nanosecond QS 1064-nm lasers for tattoo removal showed that 74–77 % of participants achieved over 75 % clearance of black ink after an average of 4–6.3 sessions [22, 23]. A more recent study [24] found a median of 75 % clearance after one to seven sessions. In our review, 69–100 % of patients achieved over 70 % clearance of black or blue ink tattoos after 1–10 sessions. Unfortunately, only one trial compared a 10-ns laser to a 35-ps laser using our predefined outcomes and found that the picosecond laser was superior in that it achieved >70 % clearance and >90 % clearance of black ink tattoos [15].

QS nanosecond lasers are all effective in removing black and dark blue ink tattoos [25, 26], yet when it comes to multicolored tattoos, one is forced to employ several types of QS nanosecond lasers [26] in order to target specific colors. In theory, using a picosecond laser enables a pronounced photoacoustic effect that may be effective for a wider variety of colors. The ability of picosecond lasers to remove multicolored tattoos was demonstrated for yellow [17, 19], green [15, 17], purple [17, 18], and red [17] colors. Yet, this was not directly compared to the effect of the traditional QS nanosecond lasers.

In general, picosecond-domain laser treatments at wavelengths of 755, 758, 795, 1064, and 1064/532-nm seem safe (although 758 and 795-nm are confined to research and have not been applied clinically). Minor and transient side effects were commonly reported after picosecond laser treatment, including erythema, edema, and pinpoint bleeding. Pigmentation disorders were reported in up to 25 % of patients and were probably related to patients’ skin type. The reports of acute bulla formation varied from 0 to 50 % in different studies. Since some studies did not indicate the spot size used, we can only assume that this variation can be attributed to a difference in fluence and energy used.

Previous reviews regarding nanosecond-domain laser treatments for tattoo removal reported similar findings. Transient side effects, including pinpoint bleeding and crusting, were common [27], while transient textural changes were reported in up to 50 % [28], and hypopigmentation was reported in 12–13 % of treatment sites [27, 28]. Hyperpigmentation and scarring were rarely reported [28].

This review has several limitations. Given the limited data published on picosecond laser treatment for tattoo removal, we elected to include all published trials regardless of their design and methodology. All included trials were subject to a high risk of bias. The studies varied in the type of laser investigated and in the outcome measures, and the vast majority of human studies did not compare their results to nanosecond-domain lasers. As a result, we were unable to conclude whether picosecond-domain lasers are superior, or at least not inferior, to nanosecond-domain lasers.

In conclusion, there is sparse evidence that picosecond-domain lasers have a higher effectiveness for the removal of mainly black and blue ink tattoos compared to their nanosecond-domain counterparts, when controlled for spot size and other variables. This treatment is well tolerated, with minor side effects. There is an urgent need for well-designed randomized controlled trials to compare the two treatment modalities.