Substance and claimed effects

This entry covers depressive and anxiety disorders that arise (or persist) during pregnancy itself — antenatal depression and antenatal anxiety as clinical states, distinct from the postpartum-onset cluster, though often overlapping with it. The substance is the maternal mood/anxiety disorder during gestation, recognised, screened for, and treated. Claims under examination: that the condition is common (10–20% of pregnancies depending on definition); that it has been historically under-recognised because somatic symptoms overlap with normal pregnancy; that it produces real consequences for maternal sleep, appetite, and day-to-day functioning; that it modestly raises risks of preterm birth and low birth weight; that it predicts postpartum depression and offspring socioemotional / behavioural difficulty into adulthood; and that screening and treatment (psychotherapy and/or carefully selected antidepressants) materially improve maternal and child outcomes when delivered, while the choice to start or continue medication during pregnancy involves a real but moderate trade-off against untreated disease Howard et al. 2014 Stein et al. 2014.

Evidence by addressing question

mechanism

Three threads of mechanism are load-bearing. First, the symptom-overlap problem. Pregnancy itself produces fatigue, sleep disruption, appetite shifts, and somatic complaints that map onto seven of the nine DSM-5 major depression criteria — clinicians historically attributed the syndrome to gestation rather than to depression, which is why prevalence was systematically under-estimated until structured-interview studies replaced clinical impression Bennett et al. 2004 Howard et al. 2014. Second, the maternal-fetal stress axis. Maternal hypothalamic-pituitary-adrenal (HPA) axis activation in chronic anxiety or depression elevates cortisol; the placenta's 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) enzyme normally inactivates maternal cortisol before it reaches the fetus, but the enzyme is downregulated under chronic stress, increasing fetal cortisol exposure. This is the leading candidate mechanism for the observed associations with shortened gestation, reduced birth weight, and altered offspring HPA reactivity Glover 2014 Stein et al. 2014. Third, behavioural and downstream pathways. Antenatal depression is associated with reduced prenatal-care attendance, poorer nutrition, increased smoking and substance use, more interpersonal-violence exposure, and impaired postpartum maternal sensitivity and mother-infant interaction, each of which independently affects fetal and child outcomes. Untangling the direct biological pathway from these correlated behavioural pathways is one of the field's hardest measurement problems Stein et al. 2014 Howard et al. 2014.

evidence

Prevalence. Bennett's 2004 systematic review pooled structured-interview prevalence at 7.4% in the first trimester, 12.8% in the second, and 12.0% in the third Bennett et al. 2004. Self-report instrument prevalence runs higher. Antenatal anxiety meta-analyses pool point-prevalence around 15-21%, with generalised anxiety disorder around 4-8% across pregnancy Dennis et al. 2017. The pregnancy-or-postpartum question. Wisner's 2013 cohort of 10,000 postpartum-screen-positive women established that roughly one-third of "postpartum" depressive episodes had onset in pregnancy, and about a quarter had onset before conception, reframing the perinatal episode as a continuum rather than a postpartum event Wisner et al. 2013. Obstetric outcomes. Grote's 2010 meta-analysis of 29 studies found antenatal depression associated with preterm birth (pooled relative risk 1.39, 95% CI 1.10–1.78), low birth weight (RR 1.49), and intrauterine growth restriction (RR 1.45); effects were modest in absolute terms and varied by population and depression severity Grote et al. 2010. Grigoriadis 2013 reproduced the preterm-birth association at smaller magnitude (RR 1.13) after adjustment for confounders Grigoriadis et al. 2013. Ding's 2014 meta-analysis of prospective cohorts found maternal antenatal anxiety raised preterm birth risk (RR 1.50) and low birth weight (RR 1.76) Ding et al. 2014. Child outcomes. O'Connor's ALSPAC cohort showed antenatal anxiety in late pregnancy predicted children's behavioural and emotional problems at age 4 (odds ratio roughly 2.0) independent of postnatal depression O'Connor et al. 2002. Pearson's 2013 ALSPAC follow-up reported offspring of antenatally depressed mothers had a 1.28-fold increased risk of meeting depression criteria at age 18, even after adjusting for postnatal depression and a wide set of socioeconomic covariates Pearson et al. 2013. Madigan's 2018 meta-analysis of 71 studies confirmed small but consistent associations between prenatal depression/anxiety and child internalising and externalising problems (effect sizes d ≈ 0.10–0.20), with publication bias and unmeasured-confounder concerns flagged honestly Madigan et al. 2018. Maternal mortality. US Maternal Mortality Review Committee data across 36 states (2017–2019) found mental-health conditions — predominantly suicide and overdose — accounted for 22.7% of pregnancy-related deaths, the leading cause overall, ahead of haemorrhage and cardiovascular conditions Trost et al. 2022.

protocol

Screening. The USPSTF 2023 statement recommends universal screening for depression and suicide risk in adults including pregnant and postpartum women (B recommendation) USPSTF 2023. The 2019 USPSTF statement on preventive interventions recommends clinicians refer pregnant and postpartum persons at increased risk to counseling interventions (B recommendation; pooled relative risk reduction for new perinatal depression episodes 0.61 across 50 trials of CBT or IPT-based prevention) USPSTF 2019. ACOG's 2023 Clinical Practice Guideline No. 4 recommends screening at the initial prenatal visit, again later in pregnancy, and at postpartum follow-up — and, critically, that any positive screen be paired with a clinical pathway for diagnosis and treatment (screening without a referral pathway shows no outcome benefit) ACOG 2023a. The Edinburgh Postnatal Depression Scale (EPDS; 10 items, threshold typically ≥13 for major depression in pregnancy) is the most-validated antenatal instrument because its items strip out somatic symptoms that overlap with pregnancy Cox et al. 1987. PHQ-9 (depression; threshold ≥10) and GAD-7 (generalised anxiety; threshold ≥10) are also widely used, accepting that PHQ-9 somatic items inflate scores in pregnancy Kroenke et al. 2001 Spitzer et al. 2006. Treatment — first-line. ACOG 2023b and the Yonkers/APA-ACOG consensus support stepped care: mild-to-moderate antenatal depression or anxiety, first-line psychotherapy (CBT or IPT); moderate-to-severe, combined or medication; prior severe disease on maintenance medication, continued treatment with careful agent selection ACOG 2023b Yonkers et al. 2009. Psychotherapy evidence. Sockol 2015 pooled 40 CBT trials in perinatal women; the treatment standardised mean difference was d = 0.65 against control, with similar effects for prevention and treatment Sockol 2015. Sockol 2018 pooled 14 IPT trials in perinatal women, effect size d = 0.57 against control Sockol 2018. Medication selection. Sertraline is the most commonly recommended first-line SSRI in pregnancy and lactation; fluoxetine and citalopram are reasonable alternatives; paroxetine is generally avoided in first trimester due to historical cardiac-malformation signals; benzodiazepines are reserved for short-term use given dependence and neonatal sedation/withdrawal risks ACOG 2023b Yonkers et al. 2009.

contraindications

Few absolute contraindications to treatment exist; the harder calls are about specific medication agents. Paroxetine in the first trimester is associated with a small absolute increase in cardiac malformations (pooled odds ratio around 1.4 in older data, attenuated in confounder-adjusted analyses) and is generally avoided when an alternative SSRI is acceptable Huybrechts et al. 2014. Mood stabilisers (lithium, valproate, carbamazepine) carry teratogenic signals and are managed by perinatal psychiatry, not primary care. Benzodiazepines near term raise neonatal-withdrawal and floppy-infant-syndrome risk. ECT remains an option in severe, treatment-resistant, or psychotic depression and is considered safe in pregnancy with anaesthesia/obstetric coordination — not contraindicated by pregnancy per se ACOG 2023b.

misconceptions

Five widely repeated errors. (1) "Pregnancy is protective against depression." The hormonal-bath-of-pregnancy view dominated obstetrics through the 1990s; the structured-interview evidence shows pregnancy is not protective and depression rates in pregnancy are comparable to non-pregnant women of reproductive age Bennett et al. 2004 Howard et al. 2014. (2) "It's just hormones / baby blues." Baby blues is a transient mood lability in the first two weeks postpartum that resolves; it has nothing to do with antenatal-onset depression or anxiety, which require detection and treatment. (3) "SSRIs cause autism." The earlier observational signal collapses under sibling-controlled and confounder-adjusted designs. Brown 2017 (Ontario, n ≈ 36,000) found the unadjusted association vanished after adjustment for maternal psychiatric history; Sujan 2017 (Swedish national, n ≈ 1.6 million) found no association in within-sibling comparisons for ASD or ADHD, identifying confounding by indication as the main driver of the earlier signal Brown et al. 2017 Sujan et al. 2017. (4) "Not treating is the safe default." Untreated antenatal depression carries quantified risks to gestation length, birth weight, postpartum continuation, and offspring outcomes — and constitutes the leading cause of pregnancy-related death in the US Grote et al. 2010 Trost et al. 2022. Doing nothing is a choice with consequences, not a neutral baseline. (5) "Therapy is too slow for pregnancy." CBT and IPT show effects within 8–12 weeks; pregnancy is long enough that even mid-pregnancy initiation reaches benefit before delivery and into the postpartum period Sockol 2015 Sockol 2018.

audience

The entry's audience is pregnant people — by anatomical fact, women — across the reproductive age band. Higher-risk subgroups who warrant earlier and more frequent screening: prior history of major depression, bipolar disorder, or anxiety disorder (highest single predictor); prior perinatal mood episode; current or recent SSRI/SNRI use; first-trimester pregnancy loss or fertility-treatment history; intimate-partner violence exposure; adolescent pregnancy; low socioeconomic status; immigrant status without local family support; non-English-language; lack of social support Howard et al. 2014 ACOG 2023a. Preconception planning is the underused window — women on maintenance antidepressants who become pregnant unplanned face the worst choice (continue / discontinue / switch) under time pressure; pre-pregnancy psychiatric review optimises that decision before conception.

alternatives

The major alternatives are watchful waiting (appropriate only for mild, brief, first-episode self-limiting cases — and the field's experience is that this category is over-applied), structured psychotherapy as monotherapy (CBT, IPT — robustly evidenced first-line), antidepressant medication (sertraline first-line, fluoxetine and citalopram acceptable), combination therapy (most evidenced for moderate-to-severe), and somatic treatments for severe / treatment-resistant cases: ECT, transcranial magnetic stimulation (TMS — limited but growing pregnancy data), bright light therapy (modest evidence for antenatal depression), exercise (modest effect sizes, generally safe). Brexanolone and zuranolone, the neurosteroid GABA-A modulators, are FDA-approved for postpartum depression specifically — brexanolone as IV infusion (Meltzer-Brody 2018; Hamilton depression score reduction -19.5 vs placebo -12.8) and zuranolone as a 14-day oral course (Deligiannidis 2023; HAMD-17 day-15 difference -4.0 vs placebo). Neither is approved for antenatal use and pregnancy safety is not established Meltzer-Brody et al. 2018 Deligiannidis et al. 2023.

failure-modes

Under-detection from symptom overlap. The single most common failure: depressive fatigue, sleep disruption, and appetite change are dismissed by clinician and patient as normal pregnancy. The EPDS deliberately excludes somatic items to mitigate this; clinics that screen with PHQ-9 alone over-detect (somatic-driven false positives) and under-act (clinicians discount positive screens as "just pregnancy") simultaneously Cox et al. 1987. Screening without a treatment pathway. Programs that screen but lack a referral and follow-up infrastructure show no improvement in outcomes; ACOG 2023a is explicit that screening alone is not the intervention ACOG 2023a. Abrupt antidepressant discontinuation on positive pregnancy test. Common, often unguided, and the highest-risk move: discontinuation increases relapse risk roughly five-fold in women with prior recurrent depression (the classic Cohen 2006 study), and the relapse risk itself produces the SSRI-comparable obstetric outcomes the discontinuation was meant to avoid Wisner et al. 2009. Confounding-by-indication in the SSRI literature. Observational studies that compare SSRI-exposed pregnancies to unexposed pregnancies (rather than to untreated-depression pregnancies) systematically over-state SSRI risk by attributing the depression's effect to the medication. Sibling-controlled and active-comparator designs (Sujan 2017, Brown 2017, Wisner 2009) consistently show much weaker or absent associations Sujan et al. 2017 Brown et al. 2017 Wisner et al. 2009. Stigma-driven non-disclosure. Pregnant women under-report depressive and anxious symptoms because they fear social-service involvement, partner reaction, or being seen as a "bad mother before becoming one." Clinics that ask directly, screen routinely, and normalise disclosure get higher detection rates.

practicalities

Screening in the US is integrated into prenatal care: most obstetric practices now administer EPDS or PHQ-9 at the new-OB visit, around 28 weeks, and at the 6-week postpartum visit. Insurance covers screening as preventive care. Therapy access remains the bottleneck — perinatal mental health is an undersupplied subspecialty; waitlists are long; telehealth has substantially widened access since 2020. The Postpartum Support International network and state-level perinatal-psychiatry consultation lines (e.g., MCPAP for Moms in Massachusetts, the National Maternal Mental Health Hotline 1-833-852-6262 in the US) provide clinician-to-clinician backup that lets a general OB or family-medicine clinician manage moderate cases without subspecialty referral. Medication monitoring in pregnancy typically increases through gestation (volume of distribution changes; some SSRIs require dose increases in the third trimester); a perinatal psychiatrist or pharmacist-led service is the standard of care for women on medication.

stakes

Untreated antenatal depression and anxiety carry a stack of consequences that compound across gestation, delivery, and the child's first years. Maternal: persistence into postpartum (~50% of antenatal-onset cases continue), poorer adherence to prenatal care, increased smoking and substance use, increased risk of pre-eclampsia in some studies, and — at the severe end — suicide and overdose, the leading category of pregnancy-related death in the US Trost et al. 2022. Obstetric: modest absolute increases in preterm birth and low birth weight, with their downstream NICU and developmental implications Grote et al. 2010 Grigoriadis et al. 2013. Bonding: impaired postpartum maternal sensitivity and reduced infant-directed positive affect, both of which are mediators rather than just outcomes Stein et al. 2014. Child: doubled risk of meeting depression criteria at age 18 (Pearson 2013 ALSPAC) and small-but-consistent elevations in childhood internalising and externalising problems (Madigan 2018) — effect sizes that look modest at the individual level but matter at the population level Pearson et al. 2013 Madigan et al. 2018.

payoff

Treatment evidence comes from a mixture of perinatal-specific trials (psychotherapy) and extrapolation from general adult depression treatment (medication). The psychotherapy meta-analyses produce moderate effect sizes within pregnancy itself (d ≈ 0.57–0.65) Sockol 2015 Sockol 2018. The USPSTF preventive-counseling recommendation rests on a 50-trial pooled relative risk reduction of about 39% for new perinatal depression episodes among at-risk women referred to CBT- or IPT-based prevention USPSTF 2019. Medication continuation in women with prior severe depression substantially reduces relapse rates in pregnancy. Successful treatment normalises maternal sleep and appetite; reduces postpartum-depression incidence; restores maternal sensitivity and bonding behaviours; and — by removing the maternal substrate — likely attenuates the offspring developmental signal, though the latter remains harder to demonstrate directly because treatment trials in pregnancy rarely follow offspring to school age.

history

Antenatal depression was largely invisible in the obstetric literature through the 1980s. The Edinburgh Postnatal Depression Scale was published in 1987 as a postpartum tool but was rapidly validated in pregnancy Cox et al. 1987. The structured-interview prevalence studies of the 1990s and 2000s overturned the "pregnancy is protective" view. The 2014 Lancet series on perinatal mental disorders (Howard, Stein, Glover, and colleagues) consolidated the field's case for treating perinatal mental health as a core component of maternity care rather than a postpartum specialty add-on Howard et al. 2014 Stein et al. 2014. The USPSTF 2016 (depression screening including pregnancy) and 2019 (preventive counseling) statements anchored the policy shift; ACOG's 2018 committee opinion and 2023 clinical practice guidelines made it operational in US obstetric practice USPSTF 2019 ACOG 2023a ACOG 2023b.

out-of-scope

Postpartum-onset depression as a distinct clinical entity (and brexanolone/zuranolone, which are postpartum-only); postpartum psychosis (a psychiatric emergency, separate pathway); bipolar disorder in pregnancy (mood-stabiliser teratogenicity makes it a different conversation); breastfeeding-while-medicated decisions; preconception medication optimisation; perinatal OCD and PTSD as distinct entities.

The credibility range

The optimist case

Antenatal depression and anxiety are common, detectable with a 10-item questionnaire, treatable with effect sizes comparable to non-perinatal depression, and treatment improves a stack of outcomes for two people. The mechanism is well-articulated (HPA axis, placental enzyme regulation, behavioural pathways) Glover 2014. The obstetric and offspring signals replicate across cohorts and continents Grote et al. 2010 Madigan et al. 2018. Psychotherapy evidence is RCT-grade for both treatment and prevention Sockol 2015 USPSTF 2019. The medication-safety literature has, in the past decade, repeatedly shown that earlier observational alarms (autism, cardiac defects, PPHN) attenuate to small or null under sibling-controlled designs Sujan et al. 2017 Brown et al. 2017 — meaning the historical "SSRIs are dangerous in pregnancy" framing has been substantially weakened, and untreated depression now looks worse than treated depression on most measured outcomes. USPSTF and ACOG have aligned around universal screening with a referral pathway USPSTF 2023 ACOG 2023a. This is a tractable problem with established tools.

The skeptic case

The observed associations with obstetric and child outcomes are modest in effect size and heavily confounded — Madigan 2018 effect sizes of d ≈ 0.10–0.20 are at the boundary of clinical meaningfulness, and the residual confounding from genetics, family environment, and postnatal maternal mood may explain a large share of what looks like an antenatal causal signal Madigan et al. 2018. RCTs of antenatal depression treatment measuring offspring outcomes are scarce; most evidence is observational and cannot definitively show that treating antenatal mood improves child outcomes (as opposed to merely improving maternal outcomes, which is real but a smaller win). SSRI safety in pregnancy remains genuinely contested in some sub-areas (paroxetine, persistent pulmonary hypertension of the newborn — the Huybrechts 2015 study attenuated but did not eliminate the signal) Huybrechts et al. 2015. Screening programs implemented without a treatment pathway have shown no outcome benefit and consume clinician time. Many "preventive" psychotherapy trials in the USPSTF meta-analysis had high attrition and high risk of bias USPSTF 2019. The narrative "untreated depression is worse than treated depression in pregnancy" is true in expectation but rests partly on assumed counterfactuals — the woman who chooses not to medicate is not necessarily exposed to the full untreated-depression risk profile.

The author's call

The condition is real, common, treatable, and substantially under-detected; the case for routine screening with an attached treatment pathway is settled at guideline level (USPSTF and ACOG agree, B-grade evidence) USPSTF 2023 ACOG 2023a. The case for first-line psychotherapy is strong and uncontested. The case for antidepressant medication when warranted is, on net, in favour of treating — the historical alarm signals have substantially weakened under better study designs, and untreated severe depression is materially worse than treated severe depression on every measured maternal outcome. The case for an antenatal-treatment → improved-child-outcome causal chain is the area where the optimist case overreaches the evidence: effect sizes on child outcomes are small, residual confounding is real, and treatment-arm RCTs with long-term offspring follow-up barely exist. The entry should treat the maternal indication as primary and the offspring indication as secondary-and-honest: a real but modest expected benefit, not the central reason to act. Meta scores accordingly: mood dominant; evidence high (4) on the screening + psychotherapy + medication-safety packages, lower on the child-outcome causal chain; controversy moderate (3) because SSRI use in pregnancy still divides clinicians and the public.

Stakeholder + incentive map

Guideline bodies (USPSTF, ACOG, NICE, RCOG, RANZCOG): aligned in favour of screening + stepped treatment; their incentive is professional consensus and litigation-defensibility. Perinatal psychiatry subspecialty: pushes for greater recognition, more training, and broader access; modest commercial incentive (small subspecialty); strong professional incentive. Obstetric primary practice: implementation gap — recognises the guideline, often lacks the referral infrastructure to act on positive screens; incentive to under-screen if a positive screen creates more work than support. Pharmaceutical industry: limited direct commercial incentive in the SSRI space (mostly generic); active in the newer neurosteroid space (brexanolone/zuranolone for postpartum) where pricing is high. Patient-advocacy organisations (Postpartum Support International, 2020 Mom): push for screening and access. Cultural pushback: persistent "pregnancy is a happy time" expectation in popular media, and a long-running suspicion of medicating pregnant women on intuitive moral grounds; both produce under-treatment. Reproductive-rights and child-welfare tension: positive screens in some US states risk Child Protective Services involvement for substance-use disclosure, which structurally suppresses honest reporting and creates a real disincentive to disclose.

Population variability

Strongest single risk factor is prior history of depression, anxiety, or any perinatal mood episode; one prior major depressive episode roughly doubles risk; prior perinatal episode triples it. Socioeconomic gradient is steep — antenatal depression prevalence runs roughly twice as high in low-income samples. Adolescent pregnancy: substantially elevated rates. Intimate-partner violence exposure: among the most consistent risk factors, present in perhaps a fifth of antenatally depressed women in some samples Howard et al. 2014. Ethnic disparity in detection: US data show Black, Hispanic, and immigrant women are screened at lower rates and offered treatment at lower rates than white women with comparable presentations; the maternal-mortality data shows the mortality gap (Trost 2022 — Black women have 2-3× the pregnancy-related death rate) is partly attributable to mental-health-cause deaths Trost et al. 2022. Baseline status matters: women on maintenance SSRIs who become pregnant face a different decision than women experiencing first-onset depression in pregnancy. First-trimester tends to be the lowest-prevalence trimester for depression; anxiety often peaks in the first trimester and in the late third Bennett et al. 2004 Dennis et al. 2017.

Knowledge gaps

Five real holes. (1) Causal mediation of offspring outcomes. Whether antenatal mood causally affects offspring development independently of postnatal mood and shared genetics is hard to establish; designs that could (sibling-comparison, RCT-with-offspring-follow-up, Mendelian-randomisation-adjacent) are sparse Madigan et al. 2018. (2) Active-comparator SSRI safety RCTs are ethically constrained. Most pregnancy-SSRI safety data will remain observational; better confounder adjustment is the realistic ceiling. (3) Anxiety-specific treatment trials in pregnancy are thinner than depression trials. Anxiety meta-analyses pool fewer perinatal-specific studies and lean more on general-adult evidence. (4) Newer neurosteroid agents (brexanolone, zuranolone) have no antenatal data and will likely remain off-label in pregnancy for some years; whether they generalise to antenatal indication is an open question Meltzer-Brody et al. 2018 Deligiannidis et al. 2023. (5) Implementation evidence. Screening rates and pathway completion vary by clinic; what features of a referral pathway translate screening into outcome benefit is under-studied. The system-level intervention is at least as important as the choice of psychotherapy or medication.