In one line
Perinatal mental illness runs a spectrum from transient baby blues through perinatal depression and anxiety to puerperal psychosis, and the consultant task is to separate the self-limiting from the dangerous — recognising that suicide is a leading cause of maternal death and that puerperal psychosis is a psychiatric emergency demanding admission — while prescribing psychotropics whose benefit to a treated mother almost always outweighs the over-stated fetal risk.
Mechanism & pathophysiology
The puerperium is the steepest neuroendocrine cliff in human physiology. Within forty-eight hours of placental separation, oestradiol and progesterone fall from late-pregnancy levels — hundreds of times the non-pregnant concentration — to near-zero, and this abrupt withdrawal acts on a brain whose serotonergic, GABAergic and HPA-axis systems have spent nine months remodelling around a high-steroid environment. The baby blues — the tearful, labile, over-sensitive state peaking around days three to five and resolving by day ten in up to two-thirds of women — is best understood as the normal central nervous system response to that withdrawal, and its self-limiting course is the feature that distinguishes it from depression. The same steroid-withdrawal model underlies the recent interest in neuroactive steroids: allopregnanolone, a progesterone metabolite and positive allosteric modulator of the GABA-A receptor, falls precipitously after delivery, and the licensing of brexanolone and oral zuranolone (synthetic allopregnanolone analogues) as rapid-acting treatments for postpartum depression is the clinical vindication of the hypothesis that disordered GABAergic signalling, not simply "low mood", sits at the centre of the biology.
Depression and anxiety in the perinatal period are not a single disease with a hormonal trigger; they are the common final pathway of a genetic-temperamental vulnerability colliding with the steroid shift and a heavy load of psychosocial stress. The strongest single predictor of perinatal depression is a past history of depression or anxiety, and the second is the absence of social support; a previous perinatal episode, an unwanted or unsupported pregnancy, intimate-partner violence, and adverse life events stack the risk. In the South African setting these psychosocial determinants are not background noise but the dominant signal — poverty, food insecurity, HIV, the high prevalence of gender-based violence, and the particular vulnerability of the adolescent mother mean perinatal depression here is driven as much by the social world as by neuroendocrinology, which is why the SA evidence repeatedly shows prevalence figures far above the global average. The groundwork on screening, the social determinants and the first-line response is laid in the Intermediate chapter on GBV and mental health in pregnancy; the consultant-level task is the differential diagnosis, the psychotropic decisions, and the management of the emergency that the Intermediate chapter does not.
Puerperal (postpartum) psychosis is a biologically distinct entity and must not be filed under "severe depression". It is the most dramatic illness in obstetrics-adjacent medicine: an abrupt onset, usually within the first two weeks and often within the first few days, of a fluctuating, polymorphic psychotic state — grandiose or persecutory delusions, hallucinations, profound mood disturbance that swings between elation and despair, and a characteristic perplexity and confusion that gives it an almost delirium-like quality. It is, in the great majority of cases, a presentation of the bipolar spectrum: the single strongest risk factors are a personal history of bipolar disorder, a previous episode of puerperal psychosis, and a first-degree family history of either. A woman with bipolar disorder has roughly a one-in-five-to-one-in-two risk of a severe postpartum episode, and a woman with a previous puerperal psychosis carries a recurrence risk of the same order in a subsequent pregnancy — figures that turn the booking history into a screening question of consequence. The mechanism is incompletely understood but is thought to involve the interaction of the steroid-withdrawal and sleep-deprivation of the early puerperium with an underlying bipolar diathesis and, in some women, an immune or thyroid-autoimmune contribution. The clinical point that falls out of the biology is that this is an illness of rapid onset and rapid fluctuation, so a woman can look reassuring at one review and be floridly unwell hours later.
The danger of the spectrum is not evenly distributed. Perinatal depression harms through chronicity — impaired mother-infant bonding, disrupted attachment, poorer infant cognitive and emotional development, and, at its severe end, suicide. Puerperal psychosis harms through acuity and through its two catastrophic outcomes: suicide and infanticide. Suicide in the perinatal period is characteristically violent and determined — hanging, jumping, drowning — rather than the impulsive overdose more typical outside pregnancy, which is why a perinatal suicide is so often completed, and why the soft reassurance that "she only has fleeting thoughts" is so dangerous. Maternal suicide is a leading cause of death in the year after birth in every well-conducted confidential enquiry, and the deaths cluster in the late postnatal period rather than in pregnancy itself.
Assessment
The first discipline is to ask. Perinatal mental illness is massively under-detected because women conceal symptoms out of shame and fear of having the baby removed, and because clinicians who are comfortable asking about bleeding and blood pressure go quiet around mood. Routine, structured screening at booking and across the perinatal contacts is what closes that gap.
- Screening instruments. The two-item Whooley questions ("During the past month, have you often been bothered by feeling down, depressed or hopeless? … by having little interest or pleasure in doing things?") are the recommended case-finding opener at first contact — quick, validated, and a positive answer to either triggers fuller assessment. The Edinburgh Postnatal Depression Scale (EPDS) is the workhorse self-report tool: ten items, completed in about five minutes, validated specifically in the perinatal population so that it deliberately omits the somatic items (fatigue, appetite, sleep) that are confounded by normal new-motherhood. A score around ≥13 flags probable depression, lower thresholds raise sensitivity, and — the item that is too often skipped — question 10 asks directly about thoughts of self-harm and any positive answer demands immediate risk assessment regardless of the total. The EPDS also picks up anxiety. Both tools are validated in the urban South African setting, but the SA validation work is explicit that a screening score is a flag, not a diagnosis: it must be followed by a clinical interview, because cultural idioms of distress, language, and the somatic presentation common in this population mean the numbers alone misclassify.
- History. The booking history is a risk-stratification exercise: past or current psychiatric illness and its treatment, and specifically any history of bipolar disorder, previous puerperal psychosis, or a first-degree relative with either — these are the questions that identify the woman at high risk of the emergency, and they change the antenatal plan. Then the present state: mood, anhedonia, sleep (distinguishing the can't-sleep of depression and the doesn't-need-sleep of incipient mania from normal infant-driven sleep loss), appetite, anxiety and intrusive thoughts, the quality of the bond with the baby, and the social scaffold — partner, support, housing, food security, immigration or documentation status, substance use, and a direct enquiry about intimate-partner violence.
- The risk assessment is the examination. Ask explicitly about thoughts of death, of self-harm, and — separately and without flinching — about thoughts of harming the baby. Distinguish the egodystonic intrusive thoughts of perinatal OCD (the mother is horrified by an unwanted thought of harm, recognises it as wrong, and takes steps to avoid the baby; these carry a low risk of acting) from the ego-syntonic, delusionally-driven risk of psychosis or severe depression (the harm is congruent with a delusional belief — that the baby is evil, or that mother and baby must die together to be spared some imagined horror; this is high-risk and an emergency). Altruistic or "extended" suicide, in which a severely depressed or psychotic mother kills the infant believing she is protecting it, is the mechanism behind maternal infanticide and is the scenario the assessment exists to catch.
- Examination and investigations are about exclusion. New-onset psychosis or confusion in the puerperium is puerperal psychosis until proven otherwise, but the differential includes organic causes that are dangerous to miss: sepsis and delirium, eclampsia, intracranial events, thyroid disease (postpartum thyroiditis), Wernicke's encephalopathy after hyperemesis, drug intoxication or withdrawal, and HIV-associated neurocognitive or opportunistic CNS disease in the SA setting. A first psychotic episode therefore warrants observations, bloods (FBC, U&E, calcium, glucose, thyroid function, CRP), and a low threshold for imaging and lumbar puncture if any feature points organic.