After Two Antidepressants: What the STAR*D Trial Tells Us (and Its Limits)

STAR*D, which NIMH describes as the largest and longest study ever conducted to evaluate depression treatment, found that the chance of remission falls sharply after the first two medication steps. Remission rates were 36.8% and 30.6% at steps one and two, then 13.7% and 13.0% at steps three and four. People who needed more steps also relapsed more often during follow-up.

The study's famous headline, a 67% cumulative remission rate, has been challenged: a 2023 reanalysis using the protocol's own primary measure put it at 35.0%, and the original investigators disputed that. Either way, the practical lesson for patients and clinicians is the same. After two adequate antidepressant trials have not worked, a third or fourth medication becomes less likely to succeed, and it is reasonable to consider other kinds of treatment, including TMS.

What STAR*D was

STAR*D (Sequenced Treatment Alternatives to Relieve Depression) was funded by the National Institute of Mental Health and enrolled 4,041 adults with nonpsychotic major depression. It ran in 41 psychiatry and primary care centers and, unlike most trials, enrolled people who were seeking care and who had common medical and psychiatric conditions, to make the results generalizable. Everyone started on citalopram. Anyone who did not remit or could not tolerate a step was encouraged to move to the next.

The treatments at each level are as described by NIMH, and the remission rates are from the 2006 summary report, where remission meant a QIDS-SR16 score of 5 or less. In that report, 3,671 people received step one and only 123 received step four, which shows how many dropped out or stopped along the way. NIMH reported withdrawal of 21% after level 1, 30% after level 2 and 42% after level 3.

The 67% headline and the 2023 reanalysis

The 2006 report stated that "the overall cumulative remission rate was 67%." That figure became the standard reassurance that most people eventually get well if they keep trying medications.

In 2023, Pigott and colleagues reanalyzed the patient-level data. They argued that the summary report used a clinic-administered, non-blinded self-report scale instead of the protocol-specified blinded Hamilton scale, and that it included 99 patients who already scored as remitted at study entry and 125 who scored as remitted when starting their next level. Using the protocol's primary measure and excluding those patients, the cumulative remission rate was 35.0%. The original STAR*D investigators published a reply, "The STAR*D Data Remain Strong," disputing the reanalysis.

This guide does not try to settle the dispute. A reasonable reading is that 67% is an optimistic upper estimate drawn from an open-label study with heavy dropout, and that the true cumulative remission rate in routine care is probably lower.

What both readings agree on

• Each additional medication step yields less. The drop after level two is steep in the original data, not only in the reanalysis.
• Getting well later is less durable. Those who required more treatment steps had higher relapse rates in the 12-month follow-up, and those who entered follow-up in remission relapsed less than those who did not.
• Persistence still matters. Some people did remit at levels three and four, so a third medication is not pointless; it is simply one option among several, with modest odds.

The limits of STAR*D

STAR*D answers a narrow question well and leaves others open.

• It was open-label with no placebo arm, so it cannot separate drug effects from time, attention and natural recovery.
• Its well-known remission rates rely on a self-report scale, which is the core of the 2023 critique.
• It tested medication strategies available in the early 2000s. It did not test TMS, esketamine or ketamine, and cognitive therapy was offered only as an option at level two.
• It enrolled nonpsychotic depression only, so it says nothing about psychotic depression, for which the VA/DoD guideline gives ECT a strong recommendation.

It is also a population average. Individual factors, such as side effects that ended a trial early, a dose that was never adequate, or an undiagnosed medical contributor, can make a "failed" trial less informative than it looks. That is why a medical workup before calling depression treatment-resistant is worth doing, and why some clinicians use pharmacogenomic testing in selected cases.

Where "two antidepressants" comes from

Most current definitions of treatment-resistant depression center on two adequate trials that have not worked, which matches where STAR*D's curve bends. The definitions differ in detail:

The FDA labels set a lower bar than most insurers, and the 2018 NNDC consensus notes that FDA does not restrict rTMS to patients with only one medication failure. In practice, coverage rules rather than clinical evidence usually decide when TMS becomes available. The patient-side explanation of these rules is on Mind Body Seven: how many medications before TMS.

Reading your own treatment history

Before concluding that two medications have "failed," it is worth checking whether they were real tests. Questions a psychiatrist will usually go through include:

• Was the dose adequate, and was it taken consistently for long enough to judge?
• Did a trial end early because of side effects rather than lack of effect? Intolerance and non-response lead to different next steps.
• Was an add-on strategy ever tried, or only switches from one drug to another?
• Has evidence-based psychotherapy been part of treatment, and was it the right type and intensity?
• Is the diagnosis still right? Unrecognized bipolar disorder, a thyroid or sleep problem, substance use or an ongoing stressor can all look like treatment resistance. The differences are explained in bipolar disorder vs depression.

The answers do not just decide eligibility for TMS. They shape which next step makes the most sense, and they are the same details insurers ask for in a prior authorization request, so gathering pharmacy records and old notes early saves time later.

When to consider neuromodulation

STAR*D does not prove that TMS works better than a third medication; it never compared them. What it shows is that the expected benefit of each additional medication step is small, which changes the balance in favor of treatments that work by a different mechanism.

Guidelines reflect this. CANMAT 2023 rates rTMS as a first-line treatment for treatment-resistant depression and says it is typically recommended after first-line medication and psychotherapy have failed. It adds that rTMS can be a first choice when medications are poorly tolerated or when TMS worked in a previous episode, and that outside emergencies it should generally be considered before ECT. The Clinical TMS Society recommends that TMS be considered when antidepressants have not given satisfactory benefit or when intolerance precludes their use.

Timing may also matter for TMS itself. In the NIMH-funded OPT-TMS trial, most remitters had low antidepressant treatment resistance, and a pooled analysis of six trials found that greater treatment refractoriness predicted less improvement. The more steps a person has already been through, the lower the odds with any next treatment, which is an argument against waiting through many more medication trials once two have failed. The research on who responds to TMS covers this in more detail.

TMS is not the only option at this point. Augmentation with lithium or an atypical antipsychotic, a switch to a different class, psychotherapy, esketamine and, for severe or psychotic depression, ECT all have a place. The treatment-resistant depression options compared page lays these side by side. APA's 2010 guideline still lists ECT as the most effective therapy when depression has not responded to medication.

If you are thinking about suicide or are in crisis, call or text 988 (the 988 Suicide & Crisis Lifeline) or go to the nearest emergency room.

For how antidepressants fit alongside these options, see antidepressants for depression. The TMS guide gathers the related evidence, and the TMS therapy page describes evaluation for TMS.

Frequently asked questions

What does STAR*D stand for?

Sequenced Treatment Alternatives to Relieve Depression. It was an NIMH-funded study of up to four successive treatment steps for adults with nonpsychotic major depression, published in its main summary form in 2006.

Did STAR*D test TMS?

No. It tested medication switches and add-ons, plus a cognitive therapy option at level two. Its relevance to TMS is indirect: it shows how small the expected gain from each additional medication step becomes.

Is the 67% remission rate wrong?

It is disputed. A 2023 reanalysis using the protocol's blinded Hamilton scale and excluding patients who were already remitted at entry found 35.0%. The original investigators rejected that critique. Most readers treat 67% as an optimistic upper estimate.

Should I try a third antidepressant or TMS?

Both are reasonable, and the choice depends on side effects, past responses, severity, logistics and coverage. CANMAT 2023 rates TMS first line for treatment-resistant depression. Your psychiatrist can weigh the options with you.

Does it matter how long I wait before trying TMS?

Possibly. Studies have found that people with less treatment resistance tend to do better with TMS, so cycling through many more medication trials before considering it may lower the odds.

Sources

1. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. American Journal of Psychiatry, 2006. pubmed.ncbi.nlm.nih.gov
2. Questions and Answers about the NIMH Sequenced Treatment Alternatives to Relieve Depression (STAR*D) Study - All Medication Levels. National Institute of Mental Health, n.d.. nimh.nih.gov
3. What are the treatment remission, response and extent of improvement rates after up to four trials of antidepressant therapies in real-world depressed patients? A reanalysis of the STAR*D study's patient-level data with fidelity to the original research protocol. BMJ Open, 2023. pubmed.ncbi.nlm.nih.gov
4. The STAR*D Data Remain Strong: Reply to Pigott et al.. American Journal of Psychiatry, 2023. pubmed.ncbi.nlm.nih.gov
5. Predictors of antidepressant response in clinical trials of transcranial magnetic stimulation. International Journal of Neuropsychopharmacology, 2006. pubmed.ncbi.nlm.nih.gov
6. Daily left prefrontal transcranial magnetic stimulation therapy for major depressive disorder: a sham-controlled randomized trial. Archives of General Psychiatry, 2010. pubmed.ncbi.nlm.nih.gov
7. Canadian Network for Mood and Anxiety Treatments (CANMAT) 2023 Update on Clinical Guidelines for Management of Major Depressive Disorder in Adults. Canadian Journal of Psychiatry, 2024. pmc.ncbi.nlm.nih.gov
8. The Clinical TMS Society Consensus Review and Treatment Recommendations for TMS Therapy for Major Depressive Disorder. Brain Stimulation, 2016. pmc.ncbi.nlm.nih.gov
9. Consensus Recommendations for the Clinical Application of Repetitive Transcranial Magnetic Stimulation (rTMS) in the Treatment of Depression. Journal of Clinical Psychiatry, 2018. pmc.ncbi.nlm.nih.gov
10. VA/DoD Clinical Practice Guideline for the Management of Major Depressive Disorder. Department of Veterans Affairs / Department of Defense, 2022. healthquality.va.gov
11. Practice Guideline for the Treatment of Patients With Major Depressive Disorder, Third Edition. American Psychiatric Association, 2010. psychiatryonline.org
12. FDA 510(k) K160703 (NeuroStar TMS Therapy System). U.S. FDA, 2016. accessdata.fda.gov
13. LCD L33398 Transcranial Magnetic Stimulation (version 40). CMS Medicare Coverage Database (Wellpoint Federal), 2026. cms.gov
14. New York State Medicaid Coverage of Therapeutic Transcranial Magnetic Stimulation. NYS Department of Health, Medicaid Update, 2025. healthweb-back.health.ny.gov
15. Transcranial Magnetic Stimulation and Cranial Electrical Stimulation (Clinical Policy Bulletin 0469). Aetna, 2026. aetna.com

Educational information, not medical advice. Whether TMS or any treatment fits you is a decision to make with a psychiatrist who knows your history.