How Long Do Antidepressants Take to Work?

Home > Depression > How Long Antidepressants Take to Work

Most of the improvement an antidepressant will produce arrives in its first two weeks. Across 47 placebo-controlled trials of established antidepressants, 60.2% of the total improvement seen on active medication occurred during the first two weeks of treatment, and the drug-placebo difference was largest in that same window (Posternak and Zimmerman, Journal of Clinical Psychiatry, 2005). A separate meta-analysis of 28 SSRI trials covering 5,872 adults found that SSRIs separated from placebo by the end of the first week, with improvement continuing at a slowing rate for at least six weeks (Taylor and colleagues, Archives of General Psychiatry, 2006).

The familiar advice to wait four to six weeks before judging a medication remains useful, although it describes when the full effect has arrived rather than when the effect begins. Patients who feel no change at all by the end of week two have a low probability of remitting on that dose, and that finding changes how I run a medication trial.

The first two weeks carry most of the improvement

For decades, clinicians taught that antidepressants had a built-in delay of several weeks before any true effect appeared. The trial data read differently. When Posternak and Zimmerman pooled weekly Hamilton ratings from 47 double-blind trials run between 1981 and 2000, 60.2% of the eventual improvement on medication and 61.6% on placebo occurred in the first two weeks. Importantly, the gap between drug and placebo was widest in those two weeks and narrowed stepwise afterward.

Taylor and colleagues tested the question formally, fitting early-onset and delayed-onset models to 28 randomized SSRI trials. The early-onset model fit better. SSRI treatment was associated with a 64% higher chance of a 50% symptom reduction by week one than placebo (relative risk 1.64, 95% CI 1.2 to 2.25), and the benefit kept accumulating through week six.

In the exam room, early change rarely feels like a lifted mood. It tends to show up first as sleeping through the night, eating more normally, or spending less of the day in rumination. Those shifts likely reflect the earliest pharmacologic effects, and they count as improvement even when the patient still describes the mood as flat. This is one reason I measure symptoms with a rating scale at every visit.

Week two is the decision point

Early improvement predicts the end of the story with surprising accuracy. Szegedi and colleagues pooled 41 double-blind trials with 6,562 patients and defined early improvement as a 20% or greater reduction in the 17-item Hamilton score within two weeks (Szegedi and colleagues, Journal of Clinical Psychiatry, 2009). Early improvement predicted stable response with a sensitivity of 81% or higher and stable remission with a sensitivity of 87% or higher. Negative predictive values ranged from 82% to 100%, while positive predictive values were far lower, at 19% to 60%.

In other words, nearly everyone who eventually remitted had already improved by week two. Yet improving by week two did not guarantee remission, because many early improvers later plateaued. The signal runs in one direction: the absence of any improvement at two weeks is strong evidence that the current dose will not produce remission. The authors concluded that changes in management should be considered earlier than conventionally thought.

Two cautions apply. First, the trials came from a single manufacturer's database and compared mirtazapine with other agents, so the heterogeneity across modern SSRIs and SNRIs is not fully represented. Second, trial populations exclude many of the patients I see, including adolescents and adults with comorbid ADHD or substance use. Nevertheless, the week-two finding is broadly consistent with the Taylor and Posternak analyses, and the 2022 NICE depression guideline now advises a review two to four weeks after starting medication, stating that if an antidepressant is going to work, it usually does so within four weeks (NICE, NG222, 2022).

A patient asked me recently, "It's been three weeks and I don't feel anything yet. Should I just stop?" Stopping is the one option I'd set aside. Three weeks of nothing on a starting dose tells me to look at the dose, at side effects that may be limiting adherence, and at whether the diagnosis is complete. It does not yet tell me the medication has failed.

What an adequate trial means

An adequate trial has two parts, dose and duration, and most trials that get called failures in practice have fallen short on one of them. Three weeks on a starting dose tests the starting dose. The medication has been tested once the dose has been raised to a therapeutic level and held there for roughly six to eight weeks, with symptoms measured along the way.

STAR*D, the largest real-world antidepressant study ever run, shows what that looks like. In the first step, 2,876 outpatients received flexibly dosed citalopram for up to 14 weeks under measurement-based care, and the mean exit dose was 41.8 mg per day (Trivedi and colleagues, American Journal of Psychiatry, 2006). Remission occurred in 28% by clinician rating and 33% by self-report, and the authors noted that a substantial portion of those who responded or remitted did so at or after eight weeks. Patients who later entered the augmentation step had spent a mean of 11.9 weeks on citalopram at a mean final dose of 55 mg per day before the trial was judged inadequate (Trivedi and colleagues, New England Journal of Medicine, 2006).

Accordingly, the week-two and week-six findings fit together. No improvement at week two means I adjust, usually by raising the dose once tolerability allows, and sometimes by addressing a missed diagnosis such as bipolar disorder or untreated ADHD. No meaningful response at six to eight weeks on a therapeutic dose means I change course. Both decisions are individual, and both are made with the patient, which is the core of what psychiatric medication management involves.

Notably, NICE draws the line at four weeks of no response at all on a recognized therapeutic dose. At that point the guideline asks the prescriber to review adherence, side effects, stressors, and coexisting conditions before moving to a different treatment. That sequence matches my practice.

After a failed trial: optimize, switch, augment

When an adequate trial has not produced remission, the options come in a standard order. First, I confirm the picture: adherence, alcohol or cannabis use, thyroid function, sleep, and whether the depression is part of a bipolar illness or sits alongside ADHD or an anxiety disorder that has gone untreated. Second, I optimize the dose if there is room. Third, I either switch to a different antidepressant or add a second agent.

STAR*D tested both paths. Among 727 patients who did not remit on citalopram and chose to switch, remission by clinician rating was 21.3% with bupropion, 17.6% with sertraline, and 24.8% with venlafaxine, with no significant differences in outcome or tolerability (Rush and colleagues, New England Journal of Medicine, 2006). Roughly one in four remitted, and switching within the SSRI class did as well as switching to a different class. Among 565 patients who chose augmentation instead, remission was 29.7% when bupropion was added to citalopram and 30.1% when buspirone was added, although bupropion produced a larger drop in self-reported symptoms and fewer dropouts for intolerance (Trivedi and colleagues, New England Journal of Medicine, 2006).

Those two numbers, about 25% for switching and about 30% for augmenting, are often read as a head-to-head comparison. They aren't one. Patients chose which path they were willing to be randomized within, and the augmentation group had generally tolerated citalopram and improved partially. The pattern is broadly consistent with practice: partial responders who tolerate a drug tend to do better with an add-on, while non-responders or those with side effects are better served by a switch.

Adding structured psychotherapy belongs in this sequence as well. For a patient who has had one partial medication response, combining medication with therapy frequently outperforms either alone, and the evidence on that question is laid out on the medication versus therapy versus both page. When a patient has completed two adequate trials without remission, the problem is better framed as treatment-resistant depression, and the options widen to lithium, thyroid augmentation, atypical antipsychotics, and ketamine or esketamine.

What STAR*D showed, step by step

The cumulative STAR*D numbers are the ones I quote to patients who ask what their odds are. Of 3,671 adults who entered the first step, remission by self-report was 36.8% at step one, 30.6% at step two, 13.7% at step three, and 13.0% at step four (Rush and colleagues, American Journal of Psychiatry, 2006). The theoretical cumulative remission rate across all four steps was 67%.

Two features of that sequence matter clinically. First, the yield per step holds up through the second trial and then drops by more than half, which is why a fully adequate first and second trial are worth more than a quick succession of partial ones. Second, patients who needed more steps relapsed more often during the 12-month follow-up, so those reaching step three or four warrant closer monitoring after they improve.

STAR*D has limitations that are easy to lose in the headline figure. It was open-label, and the 67% cumulative figure assumes no one dropped out. Nearly 80% of the sample had chronic or recurrent depression and most had comorbidity, which makes the sample representative of real-world practice and the remission rates lower than those in cleaner efficacy trials. Even so, two-thirds reaching remission with persistence remains the most useful single number in this field.

How long to continue after you feel well

Remission is the midpoint of treatment. In a systematic review of 31 randomized trials with 4,410 participants who had responded to acute treatment, continuing the antidepressant reduced the odds of relapse by 70% compared with switching to placebo (Geddes and colleagues, Lancet, 2003). The average relapse rate was 41% on placebo and 18% on continued medication. Most trials lasted 12 months, and the benefit appeared to persist for up to 36 months in the smaller number that ran longer.

The 2022 NICE guideline translates this into a minimum of six months of continued medication after remission, at the dose that produced the remission, with longer treatment considered for people at higher relapse risk. Risk factors NICE lists include a history of recurrent episodes, particularly within the past two years, residual symptoms, a history of incomplete response, and prior severe episodes.

For patients who have already been well for a long time, the ANTLER trial is the most informative recent evidence. Lewis and colleagues randomized 478 primary-care patients in the United Kingdom who had either two or more prior episodes or at least two years on medication, and who felt well enough to consider stopping, to continue their antidepressant or to taper onto placebo (Lewis and colleagues, New England Journal of Medicine, 2021). By 52 weeks, 56% of the discontinuation group had relapsed compared with 39% of the maintenance group (hazard ratio 2.06). Both numbers deserve attention. Stopping roughly doubled the hazard of relapse, and 44% of those who stopped stayed well through the year.

Stopping, when it happens, is a planned process. Prescribers typically reduce the dose in steps, taking smaller steps as the dose gets lower and moving more slowly with short half-life drugs, and they schedule follow-up through the months when relapse risk is highest. I don't give taper schedules on a website because the right pace depends on the drug, the duration of treatment, and how the first reduction goes. The decision belongs with your prescriber.

Who this practice fits, and when to go elsewhere

This fits adults, adolescents, and children in New York who are starting an antidepressant and want the trial run with measurement at each visit, as well as patients who have been on a medication for weeks without clear improvement and want to know whether the trial has been adequate. It also fits people who are well on medication and want a plan for how long to continue. Patients under care elsewhere who want a second opinion on a stalled trial are welcome.

Some situations belong elsewhere. Depression with psychotic features, mania, or an eating disorder with medical instability needs a higher level of care than an outpatient practice provides. Patients who want ketamine as a first-line treatment are better served by completing the standard sequence first, and I'll say so. An active suicidal crisis needs an emergency department today.

Cost and booking

Initial evaluations run 60 to 90 minutes, which is the time it takes to review prior medication trials with their doses and durations, screen for bipolar disorder and ADHD, and agree on a measurement plan. The practice is out-of-network with all insurance plans. You pay directly and receive a superbill to submit for reimbursement; many NYC plans with out-of-network benefits reimburse a meaningful share of the fee once the deductible is met. Fees are laid out at fees and out-of-network costs, and the questions to ask your insurer are on checking your out-of-network benefits.

Bring a list of every antidepressant you have tried, with the highest dose reached and how long you stayed on it. That history is the single most useful document for deciding the next step. Schedule a consultation or call (646) 893-4309.

Related

Depression · Treatment-resistant depression · Medication vs therapy vs both · Medication management · Ketamine for depression · Second opinions · What type of doctor treats depression · CBT in NYC · Emotional blunting on antidepressants · Antidepressant sexual side effects · Antidepressants and weight gain · Depression with ADHD: combining medications · Anxiety medication options · Lexapro vs Zoloft · SSRIs for teens · Stopping antidepressants · Wellbutrin vs SSRIs

When Depression Becomes an Emergency

The first weeks after starting or changing an antidepressant are a time to watch mood closely; NICE advises a review within one week for anyone aged 18 to 25 or with a particular concern for suicide risk. Suicidal thoughts short of a crisis still warrant same-day or next-day psychiatric evaluation.

Frequently Asked Questions

How long do antidepressants take to work?

Measurable improvement begins in the first one to two weeks. Across 47 placebo-controlled trials, about 60% of the total improvement on medication occurred within the first two weeks, and SSRIs separated from placebo by the end of week one in a 2006 meta-analysis. Improvement then continues at a slowing rate, so the full effect is usually judged at six to eight weeks on a therapeutic dose.

Why does my doctor say to wait four to six weeks?

Four to six weeks is roughly when the full effect of a given dose has arrived, and it remains the right window for deciding whether a medication has worked. Early improvement is a separate signal. A 20% drop in symptoms by week two predicts eventual response with high sensitivity, and no change at all by week two is a reason to review the dose rather than keep waiting.

What does an adequate antidepressant trial mean?

A therapeutic dose held for roughly six to eight weeks, with symptoms measured at each visit. In STAR*D, the first citalopram step ran up to 14 weeks with a mean exit dose of 41.8 mg per day, and many of those who remitted did so at or after week eight. Three weeks on a starting dose tests the starting dose; the medication is judged after the dose has been raised and held.

What happens if the first antidepressant doesn't work?

The sequence is usually to confirm the diagnosis and adherence, optimize the dose, then switch or augment. In STAR*D, switching to bupropion, sertraline, or venlafaxine after citalopram produced remission in roughly one in four patients, with no difference among the three. Augmenting citalopram with bupropion or buspirone produced remission in about 30%. Decisions are individual and made with the prescriber.

How long should I stay on an antidepressant after I feel better?

NICE recommends at least six months after remission, and longer for people at higher risk of relapse. Across 31 randomized trials, continuing medication cut relapse from 41% on placebo to 18%. In the 2021 ANTLER trial of patients who felt well enough to stop, 56% of those who discontinued relapsed within a year compared with 39% who continued. Stopping is a planned, gradual process worked out with the prescriber.

Should I stop if I feel nothing after three weeks?

Stopping on your own is the one option I'd set aside. No improvement at three weeks is a reason to contact your prescriber, who will typically check the dose, side effects, and adherence and often raise the dose rather than abandon the medication. If a therapeutic dose has been held for six to eight weeks without meaningful change, that is when a switch or augmentation is warranted.

Sources

1. Posternak MA, Zimmerman M. Is there a delay in the antidepressant effect? A meta-analysis. J Clin Psychiatry. 2005;66(2):148–158. PMID 15704999
2. Taylor MJ, Freemantle N, Geddes JR, Bhagwagar Z. Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and meta-analysis. Arch Gen Psychiatry. 2006;63(11):1217–1223. PMID 17088502
3. Szegedi A, Jansen WT, van Willigenburg AP, van der Meulen E, Stassen HH, Thase ME. Early improvement in the first 2 weeks as a predictor of treatment outcome in patients with major depressive disorder: a meta-analysis including 6562 patients. J Clin Psychiatry. 2009;70(3):344–353. PMID 19254516
4. Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry. 2006;163(1):28–40. PMID 16390886
5. Rush AJ, Trivedi MH, Wisniewski SR, et al. Bupropion-SR, sertraline, or venlafaxine-XR after failure of SSRIs for depression. N Engl J Med. 2006;354(12):1231–1242. PMID 16554525
6. Trivedi MH, Fava M, Wisniewski SR, et al. Medication augmentation after the failure of SSRIs for depression. N Engl J Med. 2006;354(12):1243–1252. PMID 16554526
7. Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163(11):1905–1917. PMID 17074942
8. Geddes JR, Carney SM, Davies C, et al. Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review. Lancet. 2003;361(9358):653–661. PMID 12606176
9. Lewis G, Marston L, Duffy L, et al. Maintenance or discontinuation of antidepressants in primary care (ANTLER). N Engl J Med. 2021;385(14):1257–1267. PMID 34587384
10. National Institute for Health and Care Excellence. Depression in adults: treatment and management. NICE guideline NG222. 2022. nice.org.uk/guidance/ng222