Table A1.
Characteristics of studies included in the meta-analysis (k = 34).
| Studies | Design | Population | Stimulation Settings | Method of Craving Assessment and Other Outcome Measures 2
* Craving as Primary Outcome |
Main Results | |||
|---|---|---|---|---|---|---|---|---|
| No. of Sessions 1 Total Pulses/Session |
Frequency (Hz) Intensity (% of RMT) |
Stimulation Site Method for Locating Target |
Coil Type of shamTMS |
|||||
| Depressant group | ||||||||
| Alcohol | ||||||||
| Mishra et al., 2010 [106] | RCT, SB | 45 M, 30 real/15 shamTMS, detoxified | 10 1000 |
10 Hz 110% |
R DLPFC NS |
Figure-of-8 ShamTMS coil |
ACQ-NOW *, before, after, and 1 month after last session | Significant post-rTMS reduction |
| Höppner et al., 2011 [81] | RCT, SB | 19 F, 10 real/9 shamTMS, detoxified | 10 1000 |
20 Hz 90% |
L DLPFC 10–20 syst |
Figure-of-8 45° angulation and shifting |
OCDS * Mood (HDRS, BDI) Attentional blink |
No significant effect on craving and mood |
| Herremans et al., 2012 [99] | RCT, SB, | 31 inpatients (36 included), 15 real/16 shamTMS, 21 M/10 F, detoxified | 1 1560 |
20 Hz 110% |
R DLPFC Neuronavigation |
Figure-of-8 90° angulation |
OCDS * Before, after, and the 3 days following rTMS in natural setting |
No significant effect on craving (immediate or delayed) |
| Herremans et al., 2013 [100] | RCT, SB, crossover, 1 week washout | 29 inpatients (50 included), 19 M/10 F, detoxified | 1 1560 |
20 Hz 110% |
R DLPFC Neuronavigation |
Figure-of-8 90° angulation |
OCDS Go/NoGo test |
No significant effect on craving Effect on IIRTV on Go/NoGo |
| Herremans et al., 2015 [101] | Phase 1: RCT, DB, 1 session Phase 2: NC |
26 (Phase 2: 23), 13 real/13 shamTMS, 17 M/9 F, detoxified | Phase 1: 1 Phase 2: 15, for 4 days 1560 |
20 Hz 110% |
R DLPFC Neuronavigation |
Figure-of-8 90° angulation |
Craving cue-induced: TLS *, before rTMS, day1 and 7 General craving: AUQ and OCDS, before rTMS, day 7 fMRI, before rTMS, day1 and 7 |
No significant effect on cue-induced craving Significant reduction in general craving after phase 2 |
| Ceccanti et al., 2015 [92] dTMS |
RCT, DB | 18 M, 9 real/9 shamTMS, detoxified for 10 days | 10 1500 |
20 Hz 120% |
MPFC 5 cm |
H coil ShamTMS coil |
VAS (cue-induced) Alcohol consumption, cortisolemia, and prolactinemia Before, after rTMS and each month to 6 months |
Significant reduction in craving (maintained at 1 month), alcohol consumption, cortisolemia, and prolactinemia |
| Del Felice et al., 2016 [94] | RCT, SB | 17 inpatients (20 included), 8 real/9 shamTMS, 13 M/4 F During detoxification |
4 (2/week, 2 weeks) 1000 |
10 Hz 100% |
L DLPFC 10–20 syst |
Figure-of-8 ShamTMS: wooden panel under coil |
VAS Alcohol intake, EEG, Stroop, and Go/NoGo tasks Before and after rTMS sessions, at 1 month |
No effect on craving and alcohol intake Significant reduction in fast EEG frequencies Effect on Stroop and Go/NoGo |
| Addolorato et al., 2017 [91] dTMS |
RCT, DB | 11, 5 real/6 shamTMS (12 M/2 F included) |
12 (3/week, 4 weeks) 1000 |
10 Hz 100% |
Bilat DLPFC 5.5 cm |
H coil ShamTMS: blank session |
OCDS DAT by SPECT TLFB, STAI, Zung self-rating depression scale |
No effect on craving Decrease in DAT availability and alcohol intake, effect on STAI state |
| Hanlon et al., 2017 ** [98] | RCT, SB, crossover, 7–14 days washout | 24 non-treatment-seeking alcohol-dependent, 17 M/7 F | 1 3600 |
cTBS 110% |
L FP 10–20 syst |
Figure-of-8 ShamTMS coil |
VAS (cue-induced) fMRI before and after cTBS (change in MPFC–striatal connectivity) |
No effect on craving Decreased evoked BOLD signal in left OFC, insula, and lateral sensorimotor cortex |
| Jansen et al., 2019 [82] | CT, SB | 39, 19 real/20 shamTMS, 26 M/13 F, Subjects sober for at least 3 weeks 36 HC |
1 3000 |
10 Hz 110% |
R DLPFC Neuronavigation guided by fMRI |
Figure-of-8 90° angulation |
AUQ before and after the emotional reappraisal task, after rTMS Emotion regulation and related brain activity using fMRI |
No effect on craving In AUD patients: reduced self-reported emotions to positive and negative images, reduced right DLPFC activity but no significant effect of rTMS on reappraisal-related brain function |
| Perini et al., 2019 [86] dTMS |
RCT, DB | 56 (45 finished sessions), 29 real/27 shamTMS (23/22 finished), treatment-seeking alcohol-dependent patients | 15 (5/week, 3 weeks) 1500 |
10 Hz 120% |
Insula bilat and overlaying areas excluding ant PFC | H coil, H8 ShamTMS coil |
AUQ cue-induced before each session; PACS during rTMS and follow-up (week 1, 2, 4, 8, 12) Consumption during rTMS, at the end of sessions and during follow-up; CGI and CRPS-SA during rTMS follow-up Structural and rsMRI before and after rTMS sessions |
Decrease in craving and drinking measures but with no difference between real and shamTMS rTMS Difference in rs insula connectivity after treatment between real and shamTMS groups |
| Cannabis | ||||||||
| Sahlem et al., 2018 [107] | RCT, DB, crossover, 1 week washout | 16 (2 subjects withdrew before first session), 13 M/3 F | 1 4000 |
10 Hz 110% |
L DLPFC Beam F3 method |
Figure-of-8 ShamTMS coil |
MCQ (cue-induced) prior, during, after, and 15 min after the completion of rTMS | No effect on craving Feasibility and safety validated |
| Opiate | ||||||||
| Shen et al., 2016 [108] | RCT | 20 M, 10 real/10 shamTMS, long-term addicts | 5 2000 |
10 Hz 100% |
L DLPFC NS, no MRI |
Figure-of-8 90° angulation |
VAS * (cue-induced) Before, after 1st and last session |
Significant effect on craving at day 1 and 5 |
| Stimulant group | ||||||||
| Nicotine | ||||||||
| Johann et al., 2003 [102] | RCT, DB, crossover, 2 consecutive days | 11, 2 M/9 F, motivated to quit | 1 1000 |
20 Hz 90% |
L DLPFC NS |
Figure-of-8 45–90° angulation |
VAS * | Significant craving reduction |
| Amiaz et al., 2009 [93] | RCT, DB, 4 arms: real/shamTMS, smoke/neutral cue |
48, 26 real/22 shamTMS (smoke cue: 12/9), 21 M/27 F, ≥20 cig/day | 10 and maintenance phase: 6 over 1 month 1000 |
10 Hz 100% |
L DLPFC 5 cm |
Figure-of-8 ShamTMS coil |
Cue-induced craving * by VAS and sTCQ, Consumption and nicotine dependence (FTND) Before, after rTMS and at 6 months |
Significant effect on cue-induced-smoke craving, on consumption and dependence No difference at 6 months |
| Rose et al., 2011 [88] | RCT, crossover, 3 visits (1 Hz, 10 Hz, and shamTMS) | 15, 8 M/7 F, ≥20 cig/day, with good cue reactivity | 1 at each frequency 1 Hz: 450/10 Hz: 4500 |
1 and 10 Hz 90% |
SFG 10–20 syst |
Figure-of-8 ShamTMS: motor cortex stimulation |
Shiffman–Jarvik questionnaire (cue-induced, neutral/smoke) * Cigarette evaluation questionnaire |
Cue-induced craving increase at 10 Hz but decrease if neutral cue |
| Li et al., 2013 [103] | RCT, DB, crossover, 1 week washout |
14 (16 included), 10 M/4 F, non-treatment-seeking | 1 3000 |
10 Hz 100% |
L DLPFC 6 cm |
Figure-of-8 ShamTMS coil |
QSU-B (cue-induced, neutral/smoke) * | Significant effect on craving, correlated with dependence severity |
| Pripfl et al., 2014 [87] | RCT, cross over, 1 week washout |
11 (14 included), 5 M/6 F, abstinent for 6 h | 1 1200 |
10 Hz 90% |
L DLPFC Neuronavigation |
Figure-of-8 ShamTMS: vertex stimulation |
5 points LS (cue-induced) * EEG |
Significant reduction in craving and EEG delta power |
| Diehler et al., 2014 [95] | RCT, DB Add-on a 6-sessions group CBT, rTMS at meetings 3 to 6 |
74, 38 real/36 shamTMS, 40 M/34 F | 4 (2/week) 600 |
iTBS 80% |
R DLPFC 10–20 syst |
Figure-of-8 ShamTMS: 45° angulation and 60% RMT |
QSU * before CBT and after last rTMS session Abstinence rate at 3, 6, and 12 months |
No effect on craving Significant effect on abstinence rate at 3 months |
| Trojak et al., 2015 [111] | RCT, DB Phase 1: rTMS + NRT (2 weeks) Phase 2: NRT only (4 weeks) Phase 3: follow-up (6 weeks) |
37, 18 real/19 shamTMS, 20 M/17 F, motivated to quit, at least 2 unsuccessful attempts to quit | 10 360 |
1 Hz 120% |
R DLPFC Neuronavigation |
Figure-of-8 ShamTMS coil |
VAS, FTCQ-12, and QSU Abstinence rate Before and after rTMS, week 6 and 12 |
No significant effect of add-on rTMS on craving but effect on abstinence rate, without lasting effect |
| Li et al., 2017 [104] | RCT, SB, crossover, 1 week washout | 11, 5 M/6 F | 1 3000 |
10 Hz 100% |
L DLPFC 6 cm |
Figure-of-8 ShamTMS coil |
VAS (cue-induced) Resting state fMRI |
No effect on craving Decreased fALFF in the right insula and thalamus and temporal connectivity between L DLPFC and L OMPFC |
| Cocaine | ||||||||
| Hanlon et al., 2015 [80] | RCT, SB, crossover, 7–14 days washout | 11, 9 M/2 F, non-treatment-seeking | 1 3600 |
cTBS 110% |
L MPFC 10–20 syst |
Figure-of-8 ShamTMS coil |
VAS * fMRI |
Significant decrease in craving and striatum and ant insula activity |
| Hanlon et al., 2017 ** [98] | RCT, SB, crossover, 7–14 days washout | 25, 12 M/13 F, non-treatment-seeking chronic cocaine users abstinent for 48 h | 1 3600 |
cTBS 110% |
L FP 10–20 syst |
Figure-of-8 ShamTMS coil |
VAS (cue-induced) fMRI before and after cTBS (change in MPFC–striatal connectivity) |
No effect on craving Decreased evoked BOLD signal in the caudate, accumbens, anterior cingulate, orbitofrontal and parietal cortex |
| Methamphetamine | ||||||||
| Li et al., 2013 [89] | RCT, SB, crossover, 1 h washout | 18, 10 MA dependent and 8 healthy controls, 4 M/14 F | 1 900 |
1 Hz 100% |
L DLPFC 6 cm |
Figure-of-8 45° angulation |
VAS (cue-induced, neutral/MA) * during rTMS session | For MA dependent only: significant craving increase |
| Liu et al., 2017 [85] | 5 arms (10 Hz P3 = shamTMS, 10 Hz L DLPFC, 10 Hz R DLPFC, 1 Hz L DLPFC, 1 Hz R DLPFC) | 50 M, detoxified for the last 2 months | 5 10 Hz: 2000 1 Hz: 600 |
1 Hz/10 Hz 100%RMT |
L DLPFC R DLPFC 10–20 syst |
Round coil ShamTMS condition = P3 |
VAS (cue-induced) * prior to rTMS stimulation, 30 min after rTMS on day 1 and on day 5 |
Significant decrease in craving after either at left or right side, both high and low frequency rTMS, but not after shamTMS condition |
| Su et al., 2017 [109] | RCT, DB | 30 M, 15 real/15 shamTMS | 5 1200 |
10 Hz 80% |
L DLPFC 5 cm |
Figure-of-8 90° angulation |
VAS (cue-induced) * Cognitive functions HDRS, HARS, PSQI |
Significant decrease in craving, improvement in verbal learning, memory and social cognition |
| Liang et al., 2018 [84] | RCT, DB | 48 M, 24 real/24 shamTMS | 10 (12 days) 2000 |
10 Hz NS |
L DLPFC NS, no MRI |
NS | VAS (cue-induced) Withdrawal symptoms, quality of sleep, depression, and anxiety |
Significant craving reduction Improvement in sleep, depression, and anxiety Withdrawal symptoms reduction in both groups |
| Su et al., 2020 [110] | RCT, DB | 126, 70 real/56 shamTMS, 106 M/20 F | 20 (4 weeks) 900 |
iTBS 100% |
L DLPFC 10–20 syst |
Figure-of-8 180° angulation |
VAS (cue-induced) * at baseline and after each 5 sessions Sleep quality at baseline and after sessions, cognitive functions: at baseline, 1 month, and 12 months |
Significant craving reduction Improvement in sleep quality and cognitive functions |
| Yuan et al., 2020 [90] | RCT, DB | 73 M, 37 real/36 shamTMS | 10 600 |
1 Hz 100% |
L DLPFC 5 cm |
Figure-of-8 90° angulation |
Craving (cue-induced) Impulse inhibition (2-choice odd-ball task) After 1 session, 24 h after 10 sessions, and at 3 weeks follow-up |
Significant craving decrease and improvement in response inhibition, both lasting 3 weeks after treatment |
| Behavioral group | ||||||||
| Eating Disorder | ||||||||
| Van den Eynde et al., 2010 [112] | RCT, DB | 38, 17 real/20 shamTMS, 5 M/33 F BN or EDNOS-BN, fasting 2 h before |
1 1000 |
10 Hz 110% |
L DLPFC 5 cm |
Figure-of-8 ShamTMS coil |
VAS (cue-induced, FCT) * and FCQ-S No. of binges over the 24 h after rTMS |
Significant diminution of cue-induced craving and No. of binges but no effect on FCQ-S |
| Gay et al., 2016 [96] | RCT, DB | 47 F, 23 real/24 shamTMS BN |
10 1000 |
10 Hz 110% |
L DLPFC 6 cm |
Figure-of-8 ShamTMS coil |
VAS (cue-induced, FCT) before and after first and last rTMS session No. of bingeing and purging episodes in the 15 days following last session, MADRS |
No significant effect on cue-induced craving or binge episode |
| McClelland, 2016 [105] | RCT | 49 F, 21 real/28 shamTMS 28 AN-R, 21 AN-BP Mean BMI: 16.5 |
1 session 1000 |
10 Hz 110% |
L DLPFC Neuronavigation |
Figure-of-8 | After FCT, combination of AN-related experiences evaluated by VAS of which urge to restrict * Mood, temporal discounting (TD), and salivary cortisol Before, after rTMS, and 24 h following |
No significant effect of rTMS but a trend on core AN symptoms, maintained at 24 h, and on TD No effect on cortisol |
| Gambling disorder | ||||||||
| Zack et al., 2016 [113] | RCT, DB, 3 * 3 (rTMS, cTBS, shamTMS), crossover, 1 week washout | 9 M, non-treatment-seeking | 1 of each rTMS: 450 cTBS: 900 |
10 Hz et cTBS 80% |
rTMS: MPFC cTBS: R DLPFC Neuronavigation |
rTMS: double cone coil cTBS: figure-of-8 ShamTMS: vertex stimulation |
VAS (cue-induced, slot machine) Cognitive tests: DDT, Stroop test Amount of money and frequency of gambling |
Significant effect of rTMS only on craving No effect on impulsive choice and decrease in control with rTMS and cTBS No effect on gambling behavior |
| Gay et al., 2017 [97] | RCT, DB, crossover, 1 week washout | 22, 14 M/8 F, treatment-seeking | 1 3008 |
10 Hz 110% |
L DLPFC Neuronavigation |
Figure-of-8 ShamTMS coil |
VAS (cue-induced) * Gambling behavior before and 7 days after rTMS |
Significant effect on craving, no effect on gambling behavior |
| Sauvaget et al., 2018 [89] | RCT, DB, crossover, 1–2 weeks washout | 30 | 1 360 |
1 Hz 120% |
R DLPFC Beam F3 method |
Figure-of-8 ShamTMS coil |
VAS (cue-induced) * GACS-desire factor, heart rate, blood pressure |
No effect on craving and other outcomes |
10–20 syst, International 10–20 System; ACQ-NOW, Alcohol Craving Questionnaire; AN-BP, anorexia nervosa binge/purge type; AN-R, anorexia nervosa restrictive type; AUD, alcohol use disorder; AUQ, Alcohol Urge Questionnaire; BDI, Beck Depression Inventory; Bilat, bilateral; BMI, body mass index; BN, bulimia nervosa; CBT, cognitive behavioral therapy; CCT, controlled clinical trial; CGI, Clinical Global Impressions scale; CO, carbon monoxide; CRPS-SA, Comprehensive Psychopathological Rating Scale; Self-Rate; cTBS, continuous theta burst stimulation; DAT, dopamine transporter; DB, double blind; DDT, delay discounting task; DLPFC, dorsolateral prefrontal cortex; dTMS, deep transcranial magnetic stimulation; EDNOS-BN, eating disorder not otherwise specified—bulimic type; EEG, electroencephalogram; F, female; fALFF, amplitude of low frequency fluctuation; FCQ-S, Food Craving Questionnaire-State; FCT, food challenge task; fMRI, functional magnetic resonance imaging; FTCQ-12, French version of the 12-item Short Form of the Tobacco Craving Questionnaire; FTND, Fagerström Test for Nicotine Dependence; GACS, Gambling Craving Scale; HARS, Hamilton Anxiety Rating Scale; HC, healthy control; HDRS, Hamilton Depression Rating Scale; IIRTV, intraindividual reaction time variability; iTBS, intermittent theta burst stimulation; L, left; LS, Likert scale; M, male; MADRS, Montgomery–Asberg Depression Rating Scale; MPFC, medial prefrontal cortex; NRT, nicotine replacement therapy; NS, non-specified; NSc, numerical scale; OCDS, obsessive-compulsive drinking scale; OFC, orbitofrontal cortex; OMPFC, orbital middle prefrontal cortex; PACS, Penn Alcohol Craving Scale; PFC, prefrontal cortex; PG-YBOCS, Yale Brown Obsessive Compulsive Scale adapted for Pathological Gambling; PSQI, Pittsburgh Sleep Quality Index; QSU-B, Questionnaire of Smoking Urges-Brief; R, right; RCT, randomized controlled trial; RMT, resting motor threshold; rsMRI, resting state magnetic resonance imaging; SB, single blind; SFG, superior frontal gyrus; SPECT, single photon emission computed tomography; STAI, State-Trait Anxiety Inventory Scale; sTCQ, short version of the Tobacco Craving Questionnaire; TLFB, timeline followback interview; TLS, Ten-point Likert scale; U, urinary; VAS, visual analog scale. cTBS, 3 burst at 50 Hz applied at 5 Hz; iTBS, 3 burst at 50 Hz, 2 s every 10 s. 1 session frequency detailed only when different from usual daily protocol. 2 time of measure detailed only when different from pre/post rTMS. ** same study.