This randomized, placebo-controlled, double-blind study was performed at the Department of Psychiatry and Psychotherapy of the Ludwig Maximilian University Munich. Participants were randomized 1:1 to 5 sessions of sham or active tDCS and were followed up after 3 months.

The study was approved by the ethics committee of the Ludwig Maximilian University Munich (approval number 513-13) and was performed in accordance with the International Council for Harmonisation Good Clinical Practice guidelines and the principles of the Declaration of Helsinki.

A total of 36 individuals participated in the study (22 women and 14 men). The sample size was determined in accordance with the recommendation by Lancaster et al. [25] to include at least 30 individuals in a clinical study. Participants were recruited by newspaper advertisements in a local weekly newspaper and by information on the homepage of the specialized outpatient clinic for tobacco dependence at the Ludwig Maximilian University Munich.

The inclusion criteria were as follows: nicotine dependence according to ICD-10 criteria (F17.2); smoker for at least 1 year; 10 or more cigarettes/day; a nicotine dependence score greater than 4 on the Fagerström Test (FTND) [26, 27]; a carbon monoxide (CO) value greater than 10 ppm (measured in expired air by a Micro + Smokerlyzer [Bedfont Scientific Ltd.]); older than 18 years old; provided written informed consent; and no attempt at smoking cessation or drug treatment for smoking cessation for at least 3 months before the start of the study. Exclusion criteria were the clinical diagnosis of an acute mental disorder according to ICD-10/DSM-IV; having a legal representative; pregnancy; chronic mental disorder; acute risk of suicide; drug, medication, or alcohol abuse at the time of the study; dementia according to ICD-10/DSM-IV criteria (clinical diagnosis) [28]; history of severe traumatic brain injury; evidence of structural damage to the basal ganglia or brain stem; severe neurological diseases (such as prolapsed disc in the past 6 months; polyneuropathies; Parkinson syndrome; epilepsy; systemic neurological diseases; cerebrovascular diseases; history of stroke; gradually worsening, repeated cerebral ischemia; increased intracranial pressure; normal pressure hydrocephalus); severe medical diseases (such as manifest arterial hypertension, severe heart disease, pacemaker, respiratory insufficiency); any type of electronic implant; any type of malignancy, either previous or current; severe active infection; chronic and systematic dermatological diseases; and bone diseases (such as Paget disease, osteoporosis with spontaneous fractures, new fractures).

All participants provided written informed consent to participate in the study. Participants who completed the study received a one-time financial compensation of 100 euros.

Participants were assigned to active or sham tDCS in a 1:1 ratio by an independent investigator (UP) with a double-blind randomization procedure, Software RandList (Version 1.2, DatInf GmbH, Tübingen, Germany). This computer program uses a random number generator to create a list of 4-digit numbers and assigns each participants a number (pseudonym) from this list. All study participants and study investigators were blind to the group.

Participants received 5 separate tDCS stimulations with a CE-certified Eldith DC Stimulator (neuroConn GmbH, Ilmenau, Germany). The anode was attached over the left DLPFC in a position corresponding to the left electroencephalogram (EEG) F3 location (according to the 10–20 system), and the cathode was attached over the right supraorbital cortex in a position corresponding to the EEG Fp2 location. Stimulation was performed with a 2 mA current for 20 min, with additional ramp-in and ramp-out phases of 15 s each. The ramp-in and ramp-out phases were identical in active and sham tDCS. In sham tDCS, the stimulator was pre-programmed to generate a 20-min off interval between the ramp-in and ramp-out phases; the sensation of the off interval is identical to the sensation of active stimulation [29].

All persons performing the stimulation were trained in the use of the tDCS stimulator and blind to the type of stimulation. Active or sham tDCS was activated by entering the participant’s number from the randomization list into the apparatus.

For each participant, tDCS sessions were performed at 5 visits every other day over a 9-day period (V1, V3, V5, V7, and V9). Five sessions were considered to be sufficient because Boggio et al. [22] showed in a clinical study that 5 days of anodal active tDCS over the left DLPFC resulted in a significant reduction of smoking craving and a significant decrease in the number of cigarettes smoked, and the 5 sessions were performed every other day for organizational reasons.

Immediately after each of the 5 tDCS treatments, each participant attended a 10-min brief intervention for smoking cessation. To standardize the quality of the brief intervention, all interventions were performed by the same investigator. The content of the brief intervention was based on the brochure “The Smoke-free Program” (authors’ translation) and the accompanying instructor manual “Compact version. The Smoke-free Program,” and participants received also a written copy of the brochure [30, 31]. The topics discussed with participants at each brief intervention are shown in Table 1.

Before each tDCS stimulation (V1, V3, V5, V7, and V9), the CO content of expired air was measured with a Mikro-Smokelyzer (Bedfont Scientific Ltd.). In addition, a sterile salivette device (Salivette, Sarstedt AG & Co., Nümbrecht, Germany) was used to obtain a saliva sample for a cotinine test: Participants were instructed to place a cotton roll in their mouths, to chew on it for 1 min, and then to spit it into a tube without touching it. The salivettes were kept frozen in the laboratory until analysis.

After every stimulation, participants were asked the questions in the German version of the Questionnaire on Smoking Urges (QSU) [32]. The QSU comprises 32 items that are rated on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). The items are assigned to either the Factor 1 scale, which assesses the “desire and intention to smoke and the anticipation of a positive outcome of smoking” [32, 33], or the Factor 2 scale, which “describes the anticipation of immediate relief from nicotine withdrawal or relief from negative effect and the strong urge to smoke” [32]. In addition, participants completed the self-rated Comfort Rating Questionnaire (CRQ) to assess side affects of the stimulations [34].

At the 90-day follow-up (V6), 7-day and continuous abstinence were assessed with a study-specific questionnaire comprising 5 questions on tobacco use (see Appendix). In addition, abstinence was validated with a CO test, a saliva cotinine test was performed, the QSU was completed, and the number of cigarettes smoked per day was recorded.

The parameters assessed at each visit are summarized in Table 2.

Primary endpoints were the abstinence rate, i.e., the smoking status nonsmoker, and craving. Abstinence rate was assessed as the number of participants who were nonsmokers after each of the 5 intervention sessions and at the 3-month follow-up. The abstinence rate and smoking status nonsmoker were assessed at each of the 5 tDCS visits by recording the number of cigarettes smoked per day, the cotinine level in saliva, and the CO level in expired air. At follow-up, they were assessed by recording the continuous and 7-day abstinence rate with the questionnaire on tobacco use. Participants were classified as a nonsmokers if they smoked zero cigarettes a day, had a CO level less than or equal to 5, and showed a marked decrease in the cotinine level.

Craving for tobacco was assessed with the German version of the QSU at each of the 5 study visits and at follow-up.

The secondary endpoint was the number of cigarettes smoked per day, which was recorded at each of the 6 study visits by asking the participants.

Data were analyzed with SPSS version 22.0 (SPSS Inc., U.S.A.). A per protocol analysis was performed. Because the overall sample size was less than 50 (N = 36), the Shapiro–Wilk test was applied to test for normal distribution. Because of the small sample size, group differences in nominal data (sex, smoking status) were analyzed with the Fisher test, and group differences in metric data, with the Mann–Whitney U test. To analyze the interaction effects over the course of study visits 1–6, the active and sham values at the individual visits were analyzed separately by Friedman’s two-factor analysis of variance in related samples. A P value of <0.05 was considered as statistically significant. Bonferroni correction for multiple testing was considered for final analysis but omitted as all results were insignificant.