TMS success rate: what the evidence says

Anyone searching for "TMS reviews" usually ends up in forums: one account raves, the next one warns. Both are real, and both say little about what you can expect. Individual reports are not representative, and in Germany the Heilmittelwerbegesetz (§ 11 HWG) in any case prohibits providers from advertising with patient experiences. We follow that rule out of conviction: the best answer to the success-rate question comes from meta-analyses, large randomised trials and real-world outcome data. That is exactly what you will find here, each with its source. What TMS technically is and how a session runs is explained under rTMS and How does rTMS work?

Two terms matter: response means symptoms improved by at least half; remission means the depression has largely subsided by rating-scale score. The numbers differ by study type:

• Sham-controlled trials: A meta-analysis of 29 randomised, double-blind trials with 1,371 patients found, after high-frequency rTMS of the left prefrontal cortex, a response rate of 29 % versus 10 % under sham and a remission rate of 19 % versus 5 % (Berlim et al. 2014, PMID 23507264). The US pivotal trial showed a similar picture with roughly double the response rate of sham (O'Reardon et al. 2007, PMID 17573044); in the NIMH-funded OPT-TMS trial, 14 % reached remission under active stimulation versus 5 % under sham (George et al. 2010, PMID 20439832). These trials mostly treated for only 3 to 4 weeks and enrolled people with many failed medication trials.
• Real-world data: In a multisite observational study of 307 patients under routine conditions, clinician-rated response was 58 % and remission 37 % (Carpenter et al. 2012, PMID 22689344).
• Treatment-resistant depression: A meta-analysis specifically of people in whom at least two antidepressants had failed found about a threefold higher chance of response and a fivefold higher chance of remission than sham (Gaynes et al. 2014, PMID 24922485). More under treatment-resistant depression.

A network meta-analysis in the BMJ covering 113 trials confirmed that high-frequency rTMS, theta-burst and deep TMS are superior to sham (Mutz et al. 2019, PMID 30917990). In short: roughly one in three to one in two people treated respond, and roughly one in four to five reach remission, in people for whom medication had not worked well enough. What that looks like day to day is described in rTMS for depression.

The classic protocol stimulates at 10 Hz for 20 to 40 minutes. Intermittent theta-burst stimulation (iTBS) compresses that into just over three minutes. In the THREE-D trial of 414 patients, iTBS was non-inferior to 10 Hz: response 49 % versus 47 %, remission 32 % versus 27 % after 4 to 6 weeks (Blumberger et al. 2018, PMID 29726344). A recent network meta-analysis confirms the equivalence of theta-burst protocols (Kishi et al. 2024, PMID 38844532).

SAINT / SNT went a step further: ten iTBS sessions per day on five days, with MRI-based individual targeting and a high pulse count. In the open-label pilot, 19 of 21 participants with treatment-resistant depression reached remission (Cole et al. 2020, PMID 32252538). In the subsequent double-blind sham-controlled trial, the remission rate was 79 % under active stimulation versus 13 % under sham, but with only 29 participants (Cole et al. 2022, PMID 34711062). These numbers are impressive and led to FDA clearance of the SAINT procedure in the US; they come, however, from small trials by a single group and still need confirmation in larger independent studies. A meta-analysis of accelerated protocols overall shows good efficacy but considerably more heterogeneous results (Sonmez et al. 2019, PMID 31207865). We offer a SAINT-inspired, iTBS-based accelerated protocol, details under Accelerated TMS. SAINT® is a trademark of Magnus Medical; its FDA clearance applies to the US only.

Obsessive-compulsive disorder (OCD): In the pivotal multicentre trial of deep TMS, 38 % responded versus 11 % under sham (Carmi et al. 2019, PMID 31109199). Meta-analyses of sham-controlled rTMS trials confirm a moderate effect, with differences by target (Perera et al. 2021, PMID 33775927; Steuber et al. 2023, PMID 37343662). The evidence is solid but narrower than in depression; in Europe rTMS for OCD is mostly an off-label use.

Chronic pain: The European guidelines give high-frequency rTMS of the motor cortex the highest recommendation level (Level A) for neuropathic pain and a medium level for fibromyalgia (Lefaucheur et al. 2020, PMID 31901449). The Cochrane review is more sober: it describes short-term, small effects on pain intensity and overall low-quality evidence (O'Connell et al. 2018, PMID 29652088). Realistic is moderate pain relief in a proportion of those treated, not freedom from pain.

Tinnitus: Here the data are mixed. A meta-analysis found short-term improvement in tinnitus distress after low-frequency rTMS (Liang et al. 2020, PMID 33228598); a randomised trial of 70 participants saw 56 % responders versus 22 % under placebo (Folmer et al. 2015, PMID 26181507), while other trials found no difference. The guideline therefore gives only a weak recommendation. More under rTMS for tinnitus. A cross-diagnostic meta-analysis in Lancet Psychiatry shows that rTMS of the left prefrontal cortex affects symptoms such as mood, drive and cognition across disorders (Kan et al. 2023, PMID 36898403), a hint that the affected network matters more than the diagnostic label.

The flip side of the numbers: about half of those treated do not respond, or only partly. Some patterns recur in the studies without allowing a reliable prediction:

• High degree of treatment resistance: The more medication trials have failed in the current episode, the lower the response rate; TMS works best when it is not the very last resort.
• Very long current episode: People with years of uninterrupted depression tend to respond less well than those with shorter episodes.
• Psychotic features, severe personality disorders, active substance dependence: Excluded from most trials; the evidence here is thin.
• High-dose benzodiazepines during treatment are associated with poorer response in several analyses.
• Late response: Some responders only improve in weeks 4 to 6. Anyone who stops after two weeks misses this group (Fitzgerald et al. 2016, PMID 27059158).

And then there are people in whom TMS does not work despite good prerequisites, for no identifiable reason. That is true of every depression treatment, including medication and psychotherapy. For us, three things follow. It starts with a structured assessment, with qEEG brain mapping where needed, so that target and protocol are not guessed. We look beyond the coil at what pushes down cortical excitability from outside: an untreated underactive thyroid, chronically disturbed sleep or ongoing inflammation can dampen the response, which is why lab work and sleep belong in the same plan as the stimulation. And we measure during the course, to recognise early whether a change of protocol or another method within the program makes more sense.

"TMS therapy negative reviews" is a common search, and it deserves a clear answer. There are two kinds of negative experience:

1. Side effects. The updated international safety guidelines rate rTMS as well tolerated (Rossi et al. 2021, PMID 33243615). Common are transient headache and scalp discomfort (in about a third, mostly in the first sessions), facial muscle twitching and tiredness. Rare are transient sleep disturbance, restlessness or mood swings. The most serious risk is a seizure, which with proper use occurs in well under one in a thousand people treated. In bipolar courses a switch into a manic phase is possible but rare. Without ear protection the loud clicking can strain hearing, which is why you always wear it with us. Full detail under Is rTMS painful?

2. Disappointment. The most common negative experience is not a side effect but the absence of an effect after four weeks of daily appointments and a four-figure sum. Online accounts that TMS "made everything worse" do not appear as a typical pattern in controlled trials, there, no more people deteriorate under active stimulation than under sham. Individual cases are nonetheless real and to be taken seriously, and so is the time commitment of a course. We address both openly before starting: the realistic probability of improvement, the cost, and the plan B if you do not respond. Is TMS therapy legit? Yes, it is approved by the FDA, recommended in European guidelines and backed by more than a hundred randomised trials. The numbers above say how often it works; they do not say it will work for you, which is why we measure the course.

A response is only worth much if it lasts. A meta-analysis of durability found that, among responders, 67 % remained in response at 3 months, 53 % at 6 months and 46 % at 12 months (Senova et al. 2019, PMID 30344109). In the one-year follow-up of the US real-world study, 63 % of responders maintained their improvement over 12 months, with about a third receiving booster sessions along the way (Dunner et al. 2014, PMID 25271871). Even without any maintenance treatment the effect remains detectable for several months in many trials (Kedzior et al. 2015, PMID 25683231).

In practice: about half to two thirds of those who respond benefit for a year; for the others, a repeat course or maintenance treatment with single sessions weeks apart is possible and effective in studies. We plan for that from the outset, and combine TMS, depending on your measured profile, with elements meant to stabilise the result: neurofeedback, sleep and activity regulation, psychotherapy coordinated with your existing treatment. Our aim remains to restore functional balance, as far as possible without drugs, we never change existing medication without your treating physicians. Whether TMS is an option for you is something we clarify in the free initial consultation.