NIHR DC Discover

Man taking antidepressant drug

NIHR Signal A fifth of people, who have no improvement on antidepressants at four weeks, respond if given more time

Published on 7 August 2018

doi: 10.3310/signal-000632

An adult with acute depression not yet responding to an antidepressant drug has a 1 in 5 chance of substantial symptom reduction between 5 and 8 weeks if they continue taking it. In those unresponsive after eight weeks, 1 in 10 will respond between 9 and 12 weeks.

Changing treatment plans too early can mean needlessly discarding first choice anti-depressants. This is the first systematic review to calculate the proportions of people with a delayed but positive response at different time points.

The study combined results from nine double-blind randomised controlled trials of nearly 3,500 people. The selective serotonin reuptake inhibitors tested are the category of drug that NICE recommends alone or in combination with psychological treatments for moderate or severe depression.

These findings broadly support clinical guidance to wait for 3 to 8 weeks for antidepressants to work. This includes increasing support in the interim rather than switching to a less preferred antidepressant.

Share your views on the research.

Why was this study needed?

Depression affects between 4 and 10% of adults at least once during their lives. A King’s Fund study in 2006 estimated that, between 2007 and 2026, total UK service costs resulting from depression would be £3 billion, and over £12 billion taking into account lost employment. Untreated depression greatly increases suicide risk.

Primary and secondary health services prescribe antidepressants alongside cognitive behavioural therapy and other psychological treatments. After three months of treatment, a half to two-thirds of people will be much improved, compared with around a third taking placebo tablets.

Deciding for how long to wait when an antidepressant does not work quickly can be difficult. Delayed therapeutic effects may lead to people withdrawing from treatment, especially if experiencing side effects. This systematic review aimed to estimate the probability that a person will gain benefit in comparison to placebo if treatment is continued beyond four weeks up to 24 weeks. 

What did this study do?

This systematic review included nine trials of 3,466 adults with clinically diagnosed depression receiving either a single antidepressant or placebo. These studies recorded depression symptoms every four weeks from four weeks until at least 12 weeks after the start of treatment.

Trials covered different decades (the 1990s to 2010s), diagnostic criteria, measurement scales and antidepressants. Participants may have received the antidepressant as ‘first-line’ treatment, or been previously treated with an alternative drug or intervention. The sample included adults with additional psychiatric disorders and medical conditions.

The review was well conducted, using Cochrane Collaboration guidelines. Six trials were of good methodological quality, and publication bias was not likely, meaning we can have confidence in the findings.

What did it find?

  • Between 5 and 8 weeks after the start of antidepressant treatment, just over a fifth of unresponsive people had at least a 50% drop in depressive symptoms compared with just over an eighth receiving placebo tablets (22% vs 13%; 5 trials, 1,671 people). Symptoms were measured on clinician-administered rating scales such as the Hamilton Depression Rating Scale.
  • Between 9 and 12 weeks, one-tenth of people who had not responded by the end of eight weeks had this clinically significant level of response compared with very few on placebo (10% versus 2%; 5 trials, 1,671 people).
  • The overall response rates for people on antidepressants were 42% after four weeks, 55% after eight weeks and 59% after 12 weeks. These were all higher than the rates for people on placebo; 29% after four weeks, 38% after eight weeks and 39% after 12 weeks. According to two studies, few people began to respond to antidepressants after taking them for 12 weeks.

What does current guidance say on this issue?

NICE’s 2011 quality standard on depression states that treatment should be reviewed if people have not responded adequately within 6 to 8 weeks.

The NICE 2009 depression guideline gives details for antidepressants. With no or minimal response after 3 to 4 weeks of a therapeutic dose, frequency of support should be increased, and consideration given to dose increase or switching to a different drug, taking side effects into account. If there is partial improvement by four weeks, treatment should be continued for another two to four weeks. If there is no further response, side effects or the person prefers, then switching antidepressants should be considered.

The quality standard is under review.

What are the implications?

Shared individualised decisions are key. The probability of a delayed response may differ according to the specific drug and (if also used) psychological therapy, individual characteristics, clinical features, or concurrent conditions. 

Useful improvement can occur in a fifth of people even if they haven’t improved by a month. This might encourage some patients to stick with their first anti-depressant for longer than implied in the NICE guideline. This research did not compare continuation of the same antidepressant with alternatives such as a dose increase, switching to or combining with a different drug, or continuing solely with psychological treatments.

Bibliography

Haddad P, Talbot P, Anderson J, McAllister-Williams H. Managing inadequate antidepressant response in depressive illness. British Medical Bulletin. 2015;115(1):183-201.

NHS Choices. Antidepressants. London: Department of Health; updated 2015, due for review 2018.

NICE. Depression in adults: recognition and management. CG90. London: National Institute for Health and Care Excellence; 2009, updated 2018.

NICE. Depression in adults. QS8. London: National Institute for Health and Care Excellence; 2011.

Why was this study needed?

Depression affects between 4 and 10% of adults at least once during their lives. A King’s Fund study in 2006 estimated that, between 2007 and 2026, total UK service costs resulting from depression would be £3 billion, and over £12 billion taking into account lost employment. Untreated depression greatly increases suicide risk.

Primary and secondary health services prescribe antidepressants alongside cognitive behavioural therapy and other psychological treatments. After three months of treatment, a half to two-thirds of people will be much improved, compared with around a third taking placebo tablets.

Deciding for how long to wait when an antidepressant does not work quickly can be difficult. Delayed therapeutic effects may lead to people withdrawing from treatment, especially if experiencing side effects. This systematic review aimed to estimate the probability that a person will gain benefit in comparison to placebo if treatment is continued beyond four weeks up to 24 weeks. 

What did this study do?

This systematic review included nine trials of 3,466 adults with clinically diagnosed depression receiving either a single antidepressant or placebo. These studies recorded depression symptoms every four weeks from four weeks until at least 12 weeks after the start of treatment.

Trials covered different decades (the 1990s to 2010s), diagnostic criteria, measurement scales and antidepressants. Participants may have received the antidepressant as ‘first-line’ treatment, or been previously treated with an alternative drug or intervention. The sample included adults with additional psychiatric disorders and medical conditions.

The review was well conducted, using Cochrane Collaboration guidelines. Six trials were of good methodological quality, and publication bias was not likely, meaning we can have confidence in the findings.

What did it find?

  • Between 5 and 8 weeks after the start of antidepressant treatment, just over a fifth of unresponsive people had at least a 50% drop in depressive symptoms compared with just over an eighth receiving placebo tablets (22% vs 13%; 5 trials, 1,671 people). Symptoms were measured on clinician-administered rating scales such as the Hamilton Depression Rating Scale.
  • Between 9 and 12 weeks, one-tenth of people who had not responded by the end of eight weeks had this clinically significant level of response compared with very few on placebo (10% versus 2%; 5 trials, 1,671 people).
  • The overall response rates for people on antidepressants were 42% after four weeks, 55% after eight weeks and 59% after 12 weeks. These were all higher than the rates for people on placebo; 29% after four weeks, 38% after eight weeks and 39% after 12 weeks. According to two studies, few people began to respond to antidepressants after taking them for 12 weeks.

What does current guidance say on this issue?

NICE’s 2011 quality standard on depression states that treatment should be reviewed if people have not responded adequately within 6 to 8 weeks.

The NICE 2009 depression guideline gives details for antidepressants. With no or minimal response after 3 to 4 weeks of a therapeutic dose, frequency of support should be increased, and consideration given to dose increase or switching to a different drug, taking side effects into account. If there is partial improvement by four weeks, treatment should be continued for another two to four weeks. If there is no further response, side effects or the person prefers, then switching antidepressants should be considered.

The quality standard is under review.

What are the implications?

Shared individualised decisions are key. The probability of a delayed response may differ according to the specific drug and (if also used) psychological therapy, individual characteristics, clinical features, or concurrent conditions. 

Useful improvement can occur in a fifth of people even if they haven’t improved by a month. This might encourage some patients to stick with their first anti-depressant for longer than implied in the NICE guideline. This research did not compare continuation of the same antidepressant with alternatives such as a dose increase, switching to or combining with a different drug, or continuing solely with psychological treatments.

Bibliography

Haddad P, Talbot P, Anderson J, McAllister-Williams H. Managing inadequate antidepressant response in depressive illness. British Medical Bulletin. 2015;115(1):183-201.

NHS Choices. Antidepressants. London: Department of Health; updated 2015, due for review 2018.

NICE. Depression in adults: recognition and management. CG90. London: National Institute for Health and Care Excellence; 2009, updated 2018.

NICE. Depression in adults. QS8. London: National Institute for Health and Care Excellence; 2011.

Trajectories of Acute Antidepressant Efficacy: How Long to Wait for Response? A Systematic Review and Meta-Analysis of Long-Term, Placebo-Controlled Acute Treatment Trials

Published on 17 April 2018

Henssler, J.,Kurschus, M.,Franklin, J.,Bschor, T.,Baethge, C.

J Clin Psychiatry Volume 79 Issue 3 , 2018

BACKGROUND: In patients who are not responding to antidepressant pharmacotherapy, information regarding the future probability of response with the same treatment is scarce. Specifically, it is unclear at what point in time the probability to respond or remit ceases to increase, because few studies report data on response or remission at repeated time points beyond 4 or 8 weeks of treatment. Consequently, treatment recommendations in clinical practice guidelines differ widely. DATA SOURCES: We systematically searched MEDLINE, Embase, PsycINFO, and CENTRAL databases through March 2014 using generic terms for depressive or affective disorders, individual drug names, and placebo (Prospero Registration: CRD42014010105). STUDY SELECTION: We identified double-blind, randomized studies with continuous outcome reporting from 4 weeks up to at least 12 weeks that compared antidepressant monotherapy to placebo in adult patients suffering from acute depressive disorder. DATA EXTRACTION: Data extraction and synthesis followed Cochrane Collaboration guidelines. Primary outcome was response; secondary outcomes were remission and changes in rating scale scores in previously unresponsive patients, respectively. RESULTS: Of 6,043 articles screened, we selected 9 studies including 3,466 patients. Altogether, 21.6% (18.6%, 24.9%) of previously nonresponsive patients achieved response with ongoing antidepressant treatment between weeks 5 and 8, and 9.9% (7.5%, 12.7%), between weeks 9 and 12. Probability of response when taking placebo was 13.0% (9.9%, 16.5%) between weeks 5 and 8 and 2.4% (1.2%, 4.6%) between weeks 9 and 12. Differences in the probability of response between antidepressant and placebo translated into a number needed to treat of 11 after 4 weeks and 17 after 8 weeks. Heterogeneity was low to moderate, and results remained stable across subgroup and sensitivity analyses. CONCLUSIONS: In patients unresponsive to antidepressant pharmacotherapy, improvements in psychopathology can be expected with ongoing antidepressant treatment for up to 3 months. After 8 weeks of treatment, improvement with ongoing monotherapy is relatively small.

Selective serotonin reuptake inhibitors are the newer type of antidepressant drug commonly used in clinical practice. They have a lower overdose risk, and generally fewer side effects than other anti-depressants.

Clinical depression rating scales such as the Hamilton Depression Rating Scale measure the severity of symptoms at a specific point in time. They do not give information on long-term recovery, quality of life, or functional outcomes.

Expert commentary

The symptomatic response of depressed patients to antidepressant medicines takes time. It poses a practical problem for clinical decision making: how long to wait before calling it non-response?

This meta-analysis summarises the prospect of further response at different times. After four weeks, it is about 1 in 5, after eight weeks, 1 in 10 and after 12 weeks, negligible.

So 12 weeks is clearly too long. For the individual patient, we must weigh competing probabilities and preferences. The currently unmet need is for biomarkers to improve the choice of initial treatment or predict non-response earlier in the process than four weeks.

Guy M Goodwin, Emeritus Professor of Psychiatry, University Department, Warneford Hospital

Expert commentary

Stick or twist? It’s the age-old dilemma when prescribing antidepressants. A month is a long time when considering the effects of prolonged spells of depression, including brain plasticity (and possibly increased relapse risk), social and psychological disabilities, and risk.

Knowing that further gains are unlikely after 12 weeks on monotherapy helps. But whether knowing there are marginal improvements between 5 and 12 weeks (greatest up to eight weeks) affects practice is hard to predict, and guidelines in many countries direct more urgent action. Sadly, this will mostly help those who commission services with long waiting times for treatment review and psychological and social interventions.

Scott Weich, Professor of Mental Health, University of Sheffield