The goals of treatment are to eradicate symptoms of depression, improve daily functioning and quality of life,[60] improve workplace functioning,[61] reduce suicidality, minimise treatment adverse effects, and prevent relapse.[62] Treatment modalities include antidepressants, other pharmacotherapies, psychotherapies, supportive interventions, and electroconvulsive therapy (ECT). For patients with depression undergoing outpatient treatment, significant benefits are associated with the collaborative chronic care model that incorporates patient training, organisational support, community resources, and other multidisciplinary interventions.[63][64] [ Cochrane Clinical Answers logo ]  Collaborative care appears to be effective both for patients with depression alone and for those with depression and comorbid chronic physical conditions.[65] Issues yet to be resolved in the effective deployment of collaborative care models include the education of providers, reimbursement, and communication.[66] The use of internet- and mobile-based interventions have also been shown to reduce depressive symptoms.[67][68]

Severe depression

Patients with severe depression include those who are psychotic, suicidal, catatonic, or have severe psychomotor retardation impeding activities of daily living, or severe agitation. These patients are at increased risk for suicide, impulsive and potentially self-destructive behaviour, and health complications due to poor self-care and immobility.

Consultant referral, hospitalisation, constant observation, tranquilisation, and/or ECT may be required to keep the patient safe until definitive antidepressant therapy can take effect. The pharmacological and non-pharmacological treatment options used in these patients, once the risks have been stabilised, are discussed in the section on 'Moderate depression' (below). 

Consultant referral is indicated and hospitalisation should be considered if patients:

  • Have significant suicidal ideation or intent and lack adequate safeguards in their family environment

  • Have intent to hurt others

  • Are unable to care for themselves and adhere to their treatment

  • Have psychotic symptoms

  • Have uncontrolled agitation accompanied by the risk of impulsive behaviour.

Suicide risk management

  • Suicide risk assessment is critical, especially as the risk may increase early in treatment. Routinely asking patients about suicidal ideation and reducing access to lethal means (especially firearms) can reduce the risk of suicide.[69] Close telephone follow-up by a trained psychiatrist may help reduce the risk of death by suicide after a previous suicide attempt.[70]


  • General principles of prescribing antidepressants are described in the section below on 'Moderate depression'. Antidepressants alone may not effectively address psychotic symptoms, such as delusions or hallucinations;[71] therefore, clinicians should have a lower threshold for adding an antipsychotic to antidepressant treatment in severe cases under several circumstances. 

  • For patients who have agitation as a depressive symptom, antipsychotics can directly tranquilise the distress associated with this form of severe depression. Agitated patients may require symptomatic treatment with a benzodiazepine, or possibly both an antipsychotic and a benzodiazepine, until definitive antidepressant therapy takes effect. 

Electroconvulsive therapy (ECT)

  • Although most patients referred for ECT have tried other antidepressant treatments, ECT may be considered as first-line treatment in certain patients with severe depression. It may be used early in treatment for psychosis, suicidality, or catatonia, or later in treatment for people with refractory depression or intolerance to antidepressants. ECT is often the treatment of choice for severely depressed older patients; it is effective,[72] and avoids complications that may arise from pharmacological intolerance and drug interactions associated with treatment for comorbid physical conditions. 

  • ECT is performed under general anaesthesia, 2 or 3 times a week for a total of 6 to 12 treatments.

  • Patient and clinician must be fullly informed of the potential risks, so that the patient can provide informed consent. The mortality rate of ECT is estimated to be about 2 deaths per 100,000 treatments,[73][74] meaning that it is one of the safer procedures performed under general anaesthetic. The risk of death may be increased in patients with coronary heart disease, due to a theoretical increased risk of ischaemia during the induced seizure. According to one systematic review, the majority of patients report adverse cognitive effects during and shortly after treatment, most commonly memory loss (both anterograde and retrograde amnesia).[75] This impairment seems to be short-lived according to objective assessment,[76] although a significant proportion of patients report persistent memory loss following ECT.[75] This potential risk must be balanced against the evidence in favour of its efficacy, especially in patients with severe depression.

Supportive care

  • Agitated patients require high levels of care because of their enhanced emotional distress and the risk of impulsive violence. Severe impairment of the activities of daily living due to catatonia or psychomotor retardation increases the severity of depression, as patients who are inert and bedbound, or not taking adequate sustenance run the risk of a deterioration in health while awaiting a response to pharmacotherapy. These patients may require supportive nursing care.


  • Patients with severe depression are unlikely to find other talking treatment effective, and it may worsen their outlook. Limit psychotherapy to the support necessary to manage the patient safely and to encourage the patient to accept definitive treatment. 

Moderate depression

Patients with moderate depression have severe symptoms, significant impairment but no psychotic symptoms, no suicidal ideation, and no severe psychomotor retardation or agitation. These patients are suffering and if not unable to perform their normal life tasks, they are finding it very difficult to do so. 

Antidepressants are necessary in these patients,[77][78][79] but are possibly not sufficient to improve patient outcomes. Moderately to severely depressed patients derive the greatest benefit from the combination of antidepressants and psychotherapy.[80] Evidence B  Close follow-up and at minimum supportive or educational interventions during the onset of treatment can improve medication adherence. They may also reduce the risk of self-injury or suicide that can emerge in the very early phases of recovery, when energy and arousal have increased but mood remains depressed. 

General principles of antidepressant treatment

  • The main antidepressant options include:

    • selective serotonin-reuptake inhibitors (SSRIs) (e.g., citalopram, escitalopram, fluoxetine, paroxetine, sertraline);

    • serotonin-noradrenaline reuptake inhibitors (SNRIs) (e.g., desvenlafaxine, duloxetine, levomilnacipran, venlafaxine);

    • bupropion (a dopamine-reuptake inhibitor);

    • mirtazapine (a 5-HT2 receptor antagonist);

    • vilazodone (an SSRI and partial 5-HT1A receptor agonist); 

    • vortioxetine (a serotonin-reuptake inhibitor with serotonin receptor modulation properties); 

    • agomelatine (a melatonin receptor agonist and 5-HT2c receptor antagonist); and

    • reboxetine (a noradrenaline-reuptake inhibitor). 

  • Selection of an antidepressant depends on factors other than the relative efficacy of different agents; no consistent differences in safety or efficacy have been demonstrated between antidepressants,[81] although some head-to-head differences between drugs in relation to their efficacy and acceptability have been revealed in one large meta-analysis, and might be considered as one of many factors to consider in treatment selection.[79] According to one different meta-analysis of individual patient data from 15 controlled trials of acute-phase treatment of major depression, mirtazapine may be a more rapidly effective antidepressant than SSRIs.[82] [ Cochrane Clinical Answers logo ]

  • Choice of drug should be based on patient preference, tolerability, and past evidence of effectiveness in the patient. Some clinicians may also have a lower threshold for considering treatment with an antidepressant that has been shown to be effective in a family member. One rationale for this is that a person who has seen a family member respond to a particular drug may find treatment with this drug more acceptable and may have increased expectations for recovery. This is, however, not an evidence-based approach; advances in pharmacogenomics (examining the impact of heritable genetic factors on treatment efficacy and tolerability for individual patients) may eventually provide clarification, but pharmacogenomic analysis is not yet recommended for routine use.[83]

  • Although the net result of antidepressant response is a significant reduction in suicidal ideation,[77][84] there is some evidence of increased suicidal behaviour in the first weeks of treatment, particularly in teenagers and young adults, and in those on relatively high starting doses.[85][86]  Some evidence suggests that teenagers and young adults are especially likely to experience an increase in thoughts of suicide and self-harm.[87][88] The results of one large meta-analysis suggest that in adults under the age of 25, the risk of both emergence and worsening of suicidality may be raised in weeks 3-6 of treatment (but not in weeks 1-2), which is later than has been suggested by other studies.[89] 

  • Follow up patients 1 to 2 weeks after initiating therapy, then monthly for the next 12 weeks. If you prefer systematic assessment, use the Patient Health Questionnaire-9 (PHQ-9) to assess changes in symptom severity. Titrate the antidepressant dose to the maximum tolerated in patients who experience a partial response after 2 to 4 weeks. Patients are likely to begin to show a response within the first 1 to 2 weeks of treatment; however, successful antidepressant therapy to the point of remission of all symptoms may take 6 to 8 weeks. A 50% decrease in symptom score constitutes an adequate response, and a 25% to 50% change in symptom score may indicate the need to modify treatment.

  • Determine antidepressant dose based on the known target dose range. In a minority of patients, pharmacogenomic testing may indicate using a low-end or high-end dose depending on genotype. On balance, there is no evidence to support routinely increasing the dose beyond the established dose range.[90]

  • If the response to first-line therapy is inadequate, consider switching to an alternative antidepressant.[91][92] By the end of 4 different medication trials, 60% to 70% of patients are likely to respond to treatment. Switching may be appropriate if no improvement in symptoms has occurred within the first 2 weeks of treatment;[93][94] however, be aware that early response may be, but is not necessarily, a reliable indicator of continued response.[95][96] Continue treatment if there has been some improvement for at least the full 6 to 8 weeks, but do not continue prescribing a drug providing inadequate benefit indefinitely. Of note, in patients with no initial improvement to treatment with fluoxetine, one study showed the likelihood of converting to a positive response decreased the longer patients remained unimproved.[97]

  • Caution is required when switching from one antidepressant to another due to the risk of drug interactions, serotonin syndrome, withdrawal symptoms, or relapse. The timeframe required for safely switching depends on various factors including the pharmacokinetic properties of the drugs and possible interactions between them, as well as patient characteristics such as age, sensitivity to adverse effects, and the capacity to wait to begin a new course of treatment. In some situations, it is possible to give both drugs for a short period of time while the changeover is occurring; however, this should only be done under specialist guidance. In other cases, it is safer to take a more conservative approach. This involves slowly tapering the dose of the first drug before stopping it, and then waiting a period of time before starting the second drug (known as a washout period, which is usually five half-lives of the first drug). Drugs with longer half-lives (e.g., fluoxetine) require longer washout periods (e.g., up to 5 to 6 weeks) when combined with a drug with which the combination is contraindicated (e.g., monoamine oxidase inhibitors with fluoxetine). Specific recommendations for switching from one antidepressant to another, along with suitable washout periods, are available and local guidance should be consulted. When in doubt, in the absence of such guidelines, as a general principle perform a drug interaction check to be sure there are no absolute contraindications, and then 'start low and go slow' until safety can be ascertained. 

  • If there is an inadequate response to 2 (or more) full-dose and duration antidepressants, the patient’s depression might be considered treatment resistant or refractory, and warrants a more complex approach, as outlined in the 'Treatment-resistant/refractory depression' section below. 

  • Duration of treatment following the remission of symptoms depends on the prior course of illness. Evidence shows reduced risk of relapse when antidepressants are continued for over 6 months.[98][99][100] Evidence A  Continue successful antidepressant treatment for 9 to 12 months following remission.[99][100] Evidence A Continue maintenance treatment indefinitely if the patient has had multiple episodes and relapses, incomplete treatment response, or complicating problems, such as substance abuse, that might tend to promote recurrence.[101]

Psychotherapy and other non-pharmacological treatments

  • Psychotherapy in various forms has been shown to be both effective and cost-effective in reducing depressive symptoms.[102][103]

  • In addition to pharmacotherapeutic strategies, cognitive behavioural therapy (CBT) has shown greater efficacy than pharmacological placebo across levels of severity.[104] Treatment response to CBT is comparable with antidepressant response in some studies.[105] Staged treatment trials suggest that CBT may be particularly beneficial when used during the continuation phase of treatment; CBT reduces the risk of relapse/recurrence at least as well as, and perhaps better than, antidepressant continuation.[106][107][108] In pooled clinical trials, mindfulness-based CBT was found to be particularly useful in relapse prevention.[109] Adjunctive CBT has also been found to improve outcomes for depression treatment in the primary care setting.[110]

  • Therapists often use a combination of CBT Evidence B and interpersonal psychotherapy (IPT)[111] Evidence B  or problem-solving therapy (PST).[112][113]  IPT may improve interpersonal functioning, whereas CBT appears to have an enduring effect that reduces subsequent risk after treatment ends.[114] IPT is useful only if the patient has the capacity for psychological insight and is committed to longer-term therapy.[115] Frequency for both CBT and IPT is determined by the healthcare provider. The time to response is approximately 12 weeks. PST focuses on training in adaptive problem-solving attitudes and skills.[112][116] 

  • Bibliotherapy, a program of self-help by reading, may have long-term benefits for some patients.[117]

  • Behavioural activation is a less cerebral, more behavioural alternative to CBT. It actively promotes a return to functioning and has the advantage of not requiring doctoral-level therapists to administer it. One randomised controlled trial found it to be no less effective than CBT for adults with depression.[118]

  • ECT may be an option for those who have not responded to, or cannot tolerate, antidepressants; Evidence B the response rate is better for patients with severe major depression than for moderate or mild depression.

Mild depression

Patients with mild depression have low to moderate severity symptoms, partial impairment, no psychotic symptoms, no suicidal ideation, and no psychomotor retardation or agitation.

These patients do equally well with either CBT Evidence B or an antidepressant.[119] Evidence C Combination psychotherapy and medication offers no demonstrated short-term advantage in this group. However, continued psychotherapy with antidepressant management is an effective option when continued through both acute and ongoing phases of treatment.[120]

The initial choice of therapy should be guided by patient preference and includes:

  • Antidepressant treatment

  • Psychotherapy[121][122][123]

  • Supportive interventions: self-help books, yoga, relaxation training, light therapy, exercise, tai chi, music therapy, and acupuncture[124][125][126][127][128][129][130][131][132][133][134][135][136][137][138] [ Cochrane Clinical Answers logo ]  and

  • Computer-based treatment: CBT,[139][140][141][142][143][144] PST,[145][146] and stress management.[147]

Antidepressant treatment

  • An antidepressant may be preferable in some patients as it may offer a more rapid response than non-pharmacological therapies. The most commonly prescribed antidepressants, SSRIs and SNRIs, offer similar response rates and can be used first line as monotherapy in mild to moderate disease.[148] There is essentially no consistent evidence that any of the traditional antidepressants are superior to any other.[149]

  • Choice of drug should be based on patient preference, tolerability, and past evidence of effectiveness in the patient. Some clinicians may also have a lower threshold for considering treatment with an antidepressant that has been shown to be effective in a family member. One rationale for this is that a person who has seen a family member respond to a particular drug may find treatment with this drug more acceptable and may have increased expectations for recovery. This is, however, not an evidence-based approach; advances in pharmacogenomics (examining the impact of heritable genetic factors on treatment efficacy and tolerability for individual patients) may eventually provide clarification.[83]

  • Follow up patients 1 to 2 weeks after initiating therapy, then monthly for the next 12 weeks. Use the PHQ-9 to monitor symptoms over time. Patients who experience a partial response to antidepressants after 2 to 4 weeks should have the dose titrated to the maximum tolerated. Continue successful antidepressant treatment for 9 to 12 months following remission;[99][100] Evidence A however, some physicians recommend indefinite therapy for patients with frequent previous recurrences and relapses, and who respond successfully to antidepressant treatment.


  • Psychotherapy is also considered a first-line option in mild to moderate depression. Psychotherapy appears to have a positive impact on the quality of life of patients with depression, beyond measurable reductions in depressive symptom severity.[150] Mild depression treated with psychotherapy may be less likely to progress to severe depression.[151] Psychoeducation alone can achieve remission for some patients.[152]

  • Therapists often use a combination of CBT Evidence B and IPT[111] Evidence B or PST.[112][113] IPT may improve interpersonal functioning, whereas CBT appears to have an enduring effect that reduces subsequent risk after treatment ends.[114] IPT is useful only if the patient has the capacity for psychological insight and is committed to longer-term therapy.[115] Frequency for both CBT and IPT is determined by the healthcare provider. The time to response is approximately 12 weeks. Time-limited psychodynamic therapy has also gained empirical support as a treatment for major depression.[153] 

Supportive interventions

  • Self-help books are popular and bibliotherapy has demonstrated better efficacy than no treatment at all.[154]

  • Yoga may have a beneficial effect on depressive disorders, but there are significant variations in interventions, reporting, and feasibility.[155][156] Evidence C

  • Other supportive interventions include relaxation training, light therapy, exercise, tai chi, music therapy, and acupuncture.[124][125][126][127][128][129][130][131][132][133][134][135][136][137][138] [ Cochrane Clinical Answers logo ]  In non-remitted patients, higher remission rates were observed in a higher-dose exercise group plus continuation of SSRI treatment compared with low-dose exercise plus SSRIs.[157] Conversely, cessation of exercise may worsen depressive symptoms.[158][159] [ Cochrane Clinical Answers logo ]

  • St John's wort is a herb that is considered to be effective for the treatment of mild to moderate depression.[160][161] Evidence C It may also be used as an alternative therapy (as monotherapy only) if there is no response to first- and second-line treatments.[162][163] St John’s wort has an encouraging safety profile; however, numerous reports indicate the possibility of clinically significant drug-drug interactions, which must be taken into account before prescribing.[162][163] 

Computer-based treatment

  • Evidence supports the efficacy of computer-based CBT,[139][140][141][142][143][144] PST,[145][146] and stress management.[147]  However, high withdrawal rates are common.

Treatment-resistant/refractory depression

The majority of patients with depression do not respond adequately to their first antidepressant trial, but a substantial proportion of those will respond to a second antidepressant.[164] The general consensus is to consider a depressive illness that has not responded to two antidepressant trials of adequate dose and duration (preferably using two antidepressants with distinct mechanisms of action) to be treatment resistant or refractory.[165]

Although algorithms of the approach to treatment-refractory depression have been published, in practice algorithms are often altered or broken by variables unique to an individual patient. For example, adverse effects of medications, comorbid medical conditions, or affordability, along with psychosocial factors such as temperamental vulnerabilities, behavior patterns, and life circumstances, may all affect treatment. Clinical trials centered on individual patients may be feasible, but have not often been used.[166]


  • Reassessment can be useful after an apparently failed course of treatment, as some of the residual symptoms of depression (e.g., social avoidance, sleep/wake reversal, a demoralised attitude) can reflect behavioural adaptations to depression, rather than the disease itself. In such cases, symptoms may best be corrected through behavioural intervention or psychotherapy rather than a new medication trial. With intermittent, brief follow-up visits it is also easy to miss mood-cycling that may occur between sessions that would indicate a bipolar spectrum disorder rather than pure major depression.

Antidepressant treatment

  • Assuming major depressive disorder continues to be the most salient clinical problem, alternative options for treatment-resistant/refractory depression within the antidepressant class include monotherapy with a third (or fourth or fifth) SSRI, SNRI, or an atypical agent. Combined antidepressant therapy (i.e., an SSRI or SNRI plus bupropion or mirtazapine) may be indicated as a way of making optimal use of adverse effects (e.g., adding mirtazapine to an SNRI to facilitate sleep, or bupropion to an SSRI to try to improve sexual functioning); however, the evidence has not consistently supported a synergistic effect of combined antidepressants in alleviating depression.[167][168][169][170] There is some evidence that failure on one or several antidepressants does not preclude later success.[171][164] By the end of 4 different medication trials, 60% to 70% of patients are likely to respond to treatment. Although the general rule of thumb is to give antidepressants for at least 6 to 8 weeks, if there is no improvement at all in the first 2 weeks, switching may be appropriate at that point.[172]

  • When selecting a third (or fourth or fifth) medication to switch to, consider not only another SSRI, SNRI, or atypical agent (e.g., bupropion, mirtazapine), but also a tricyclic antidepressant (TCA) (e.g., amitriptyline, desipramine, doxepin, imipramine, or nortriptyline). Historically the first-line pharmacotherapy for depression, TCAs have fallen somewhat out of favour because of their adverse effects, the need for gradual dose increases, and their potential lethality in overdose. However, they remain effective and useful for many patients. Dose TCAs according to therapeutic blood monitoring. For most TCAs there is a minimum therapeutic level; for nortriptyline, uniquely, there is a therapeutic window delineating a range of effective levels. 

  • In cases where nothing else has worked and the patient can tolerate a washout period from their current antidepressant, a monoamine oxidase inhibitor (MAOI) (e.g., isocarboxazid, phenelzine, selegiline, tranylcypromine) can be uniquely effective, even though it is associated with a more severe adverse effect profile and recommended only when other options prove ineffective.[173] The washout period depends on the half-life off the antidepressant the patient is currently on and can range from 1 to 5 weeks. Do not use a MAOI without consulting a psychiatrist first. 

  • Some studies show that combinations of antidepressants with other classes of medication are better than just a combination of different antidepressants alone.[174] In patients who have not responded to conventional antidepressants, lithium augmentation remains the best evidence-based approach; however, it is ideally initiated by a psychiatrist because of its narrow therapeutic index and risks of inadvertent toxicity from excessive dosing and drug-drug interactions. While these limitations make lithium unwieldy as a first-line treatment, evidence has emerged from one Finnish cohort study suggesting that lithium monotherapy is not only effective at preventing rehospitalisation after severe depression, but is also more effective on its own than when combined with antidepressants.[175] The addition of an atypical antipsychotic to an antidepressant has been historically controversial;[176][177] however, augmentation with some agents is becoming more common practice and may improve outcomes.[178][179] Second-generation antipsychotic medications used in combination with antidepressants demonstrate efficacy and their use is becoming widespread.[180] [ Cochrane Clinical Answers logo ]  A commercially available olanzapine/fluoxetine combination has been shown to be superior to fluoxetine monotherapy and olanzapine monotherapy in producing early improvement in patients who have not responded to an antidepressant trial.[178] Aripiprazole, which the US Food and Drug Administration (FDA) has approved for antidepressant augmentation, was found to be slightly more effective as augmentation than switching to a different antidepressant in US military veterans with treatment-resistant depression.[179] Brexpiprazole, a novel serotonin-dopamine activity modulator, is approved by the FDA as an adjunctive treatment for major depressive disorder,[181] although the evidence for its efficacy derives from a relatively small number of studies.[182] Weight gain and akathisia are among the most commonly reported adverse effects, and small effects on glucose and lipids have also been noted.[183] While some benefit has been demonstrated in meta-analysis of clinical trial data, it is unclear whether benefits outweigh risks in people without psychosis.[176] Other augmentation strategies used by specialists include thyroid hormone, modafinil, ketamine, and pindolol.[184][185][186]

Non-pharmacological approaches 

  • Check and ensure that the patient has started psychotherapy if multiple pharmacological agents have been unsuccessful. 

  • When depression is severe enough to cause danger, significant distress, or functional impairment, the superior efficacy of ECT makes it a reliable and reasonable rescue treatment. The transient impact on memory and cognition, which may reduce functioning during active treatment, make ECT less desirable for milder cases. It is important to remember that the effects of ECT generally last only a few weeks, so pharmacotherapy is necessary to sustain its effects or act as maintenance therapy. 

Recurrent episodes

Recurrent episodes of depression should be treated with the same antidepressant that previously induced remission, provided that the recurrences do not occur while under adequate maintenance treatment with such medication. Consider maintenance therapy for at least 3 to 5 years or lifetime for patients who have had their third episode of depression.[100] Patients with a first recurrence and risk factors for further recurrences (including family history of bipolar disorder, recurrence within 1 year, onset in adolescence, severe depression or suicide attempt, and sudden onset of symptoms) may also benefit from long-term maintenance therapy. The selection and success of these treatments depends on the type and severity of depressive symptoms, but most often relies on trial and error. Psychotherapy in patients who suffer from recurrent episodes may be well aimed if it addresses the despair patients often feel when they see recovery as only a temporary respite from suffering, and if it educates patients about ways to cope with and possibly prevent recurrences.


Depression coinciding with pregnancy creates a significant clinical dilemma. On the one hand, the fetus is exposed to a potential for harm by the increased likelihood of maternal substance misuse, neglect of health, or suicide. On the other hand, all antidepressants cross the placental barrier, with the potential to cause iatrogenic harm to the fetus. Fortunately, studies of the safety of antidepressant use in pregnancy for the most part add up to minimal, if any, risk to the fetus.[187][188] Unfortunately, there is little controlled trial evidence.[189] Consistent data to support fully-informed decision-making are lacking.[190][191][192][193][194][195][196]

Antidepressant risks can extend to and beyond birth. The results of one systematic review and meta-analysis found that maternal SSRI use (but not depression without SSRI use) is associated with an increased risk of preterm birth.[197] Another systematic review and meta-analysis found that pregnant women with untreated depression have an increased risk of preterm birth and low birth weight compared with women without depression, suggesting that untreated depression itself may be a risk factor for early delivery.[198] Transient irritability and other symptoms reminiscent of antidepressant discontinuation syndromes affect a substantial proportion of neonates exposed to antidepressants in utero up to the time of delivery.[199] 

Evidence of a relationship between depression, antidepressant treatment, and autism spectrum disorders (ASD) is mixed, with some studies showing an association between maternal antidepressant use during pregnancy and a slightly increased risk of ASD in the child; other studies show increased risk of ASD in children of mothers with an antenatal psychiatric disorder and no antidepressant use.[200][201][202] 

It is fairly clear that women who stop their antidepressant are more likely to have a relapse of depression during their pregnancy.[203] Depression itself may negatively affect fetal development (e.g., causing hyperactivity and irregular fetal heart rate), increase infants’ cortisol levels, impact on infant temperament, and influence behaviour in later childhood and adolescence.[204] 

The best recommendation that might arise from all of these weak and/or contradictory data is for clinicians to carefully discuss the risks of remaining on antidepressant treatment during pregnancy, against the risks of stopping or avoiding antidepressants and exposing the fetus to the harmful effects of prepartum depression. Despite the lack of consistent evidence of harmful effects of antidepressants to fetal and infant health and development, caution is required. Updated information about potential harms from antidepressants and other pharmaceuticals can be found at various resources. UK Teratology Information Service external link opens in a new window

For women with severe major depression in pregnancy, ECT may be the treatment of choice as it does not expose the fetus to any known risk.[205][206] For moderate to severe episodes, there is little consistent controlled trial evidence that antidepressants should be contraindicated during pregnancy; the risk to the fetus from the potentially harmful effects of the mother’s untreated depression on her health apparently outweighs any detectable risk to the fetus from antidepressants.[207][188][189] Treat mild depression in pregnancy as you would any other, perhaps with a slightly higher threshold for using medication than you would with a non-pregnant patient. The risk/benefit balance may tip in favour of non-pharmacological therapies, particularly because many patients may have reservations about using medication when pregnant. CBT is associated with a moderate treatment effect for major depressive disorder during pregnancy. Interpersonal psychotherapy also appears to have a treatment effect, but to a lesser extent than CBT.[208]

There is evidence to support the use of counselling interventions, such as CBT and interpersonal psychotherapy, to prevent depression in pregnant and postnatal women who are at relatively high risk for depression due to family history, stressful life circumstances, and medical complications of pregnancy and delivery to mother and baby.[209]

Postnatal depression

Screen women with risk factors for postnatal depression to prevent or immediately treat postnatal depression. There is evidence from combined studies that CBT may be effective for both prevention and treatment of postnatal depressive symptoms.[210] Longer-term therapy may further enhance psychotherapeutic benefits to mothers and their offspring. Pharmacotherapy requires careful consideration. Many breastfeeding women choose not to take medication because of concerns about infant exposure. Clinicians should have a higher threshold for prescribing psychotropic medications during pregnancy and breastfeeding. Fetal and newborn exposure, however small a risk statistically, nonetheless changes the fundamental risk-benefit equation because of the potential for long-term impact on the fetus or newborn. With increasing severity of depression, the equation might tip towards pharmacotherapy. In women who have had severe episodes of major depression, the slight risk to the fetus or baby must be weighed against the risk posed by depression in the mother, of self- or infant-neglect, or suicidal behaviour.

Updated information about potential harms to breastfeeding infants from antidepressants and other pharmaceuticals can be found at various resources. TOXNET: LactMed external link opens in a new window 

For more detailed information, please see our separate topic on Postnatal depression.

Perimenopausal women

Although symptoms of oncoming menopause can complicate the experience of depression in women, risk for depression in this population is more closely associated with prior depression than with hormonal status, and treatment is the same as for other patients.[211]

Seasonal affective disorder

Seasonal affective disorder (SAD) is a subtype of major depression, occurring with seasonal change. SAD occurs more commonly in northern latitudes and responds to bright-light or blue-light therapy, preferably combined with CBT,[212] as well as adjuvant therapy with antidepressants.

For more detailed information, please see our separate topic on Seasonal affective disorder.


Antidepressants may be effective in reducing depression and alcohol consumption in patients with comorbid depression and alcohol dependence.[213] Antidepressant use in depressed patients who are on opioid agonist therapy is not well supported.[214] Available evidence on the use of antidepressants with depression comorbid with dementia is poor, suggesting their potential value may be outweighed in many cases by the potential for adverse effects.[215] Evidence is also low quality, but more favourable, for antidepressants in patients with depression and HIV infection.[216] Support for antidepressants for depression comorbid with cancer is mixed.[217][218]

Use of this content is subject to our disclaimer