What are the reversible dementias? How would you investigate and treat them?

Definition and scale. Dementia is grouped into reversible or partially reversible and non-reversible types. Reversible dementia accounts for probably less than 4% of cases, but concurrent medical conditions causing mild cognitive disturbance are much commoner and are potentially treatable; identifying them avoids unnecessary disability and misdiagnosis. Crucially, finding a reversible cause does not guarantee that the dementia will improve once the cause is treated.

The causes. Drugs (anticholinergics and antihistamines, analgesics including narcotics and NSAIDs, anticonvulsants, antihypertensives, antiparkinsonian agents, digoxin, antiarrhythmics, H2 blockers, antianxiety drugs, antidepressants, antipsychotics, sedative-hypnotics, muscle relaxants, steroids, alcohol); psychiatric — depression (“depressive pseudodementia”); metabolic and endocrine (hypothyroidism, hyperparathyroidism, hyponatraemia, hypercalcaemia, hepatic and renal failure); nutritional (vitamin B12, folate, thiamine — Wernicke–Korsakoff, pellagra); neoplasms; trauma — chronic subdural haematoma; toxins (alcohol, heavy metals, organic poisons); infections (HIV and opportunistic infection, neurosyphilis, tuberculous meningitis); autoimmune (CNS vasculitis, temporal arteritis, SLE, multiple sclerosis); and other — normal-pressure hydrocephalus.

Clinical evaluation. These disorders are detected by careful history, physical examination and selected laboratory studies. Take a complete drug history including over-the-counter agents, maintain a high index of suspicion for excessive alcohol intake (easily missed, and a cause of dementia, delirium, depression and falls), and screen for depression. Pointers to depressive pseudodementia are prominent complaints of memory loss, patchy and inconsistent cognitive deficits, and frequent “don’t know” answers.

Laboratory investigation. Complete blood count, glucose, blood urea nitrogen, electrolytes (sodium and potassium in mEq/L), calcium and phosphorus (mg/dL), liver function tests, thyroid-stimulating hormone (µIU/mL), vitamin B12 (pg/mL), serological test for syphilis if suspected, and HIV antibodies if suspected. These blood tests are directed at finding common comorbidities rather than at identifying the cause of the underlying dementia.

Imaging and neuropsychology. CT or MRI of the head should be done in dementia of recent onset with no other clinical finding to explain it, and in those with focal neurological signs or symptoms, to exclude subdural haematoma, tumour and normal-pressure hydrocephalus. CT and MRI are roughly equivalent for detecting most remediable structural lesions; MRI shows more lesions in multi-infarct dementia but also white-matter changes of uncertain significance. Neuropsychological testing is particularly useful for differentiating depression from dementia and where mental-status score and functional status are discordant.

Treatment. Discontinue whenever feasible every drug known to impair cognition, especially strong anticholinergics. Treat depression energetically — an SSRI such as citalopram or sertraline, or the NaSSA mirtazapine, all with minimal cognitive side effects; where doubt persists, a careful trial of an antidepressant is justified both diagnostically and therapeutically. Correct B12 deficiency, hypothyroidism, hyponatraemia and hypercalcaemia; refer normal-pressure hydrocephalus for shunt assessment and chronic subdural haematoma for neurosurgery; treat neurosyphilis, HIV and tuberculous meningitis specifically; and control vascular risk factors, especially hypertension, which may favourably influence the course of both degenerative and vascular cognitive impairment. Note that thyroid hormone and vitamin B12 replacement may be important for general health but have not been shown to reverse cognitive deficits, and that older patients with reversible cognitive impairment during depression are at high risk of later dementia and need close follow-up.

Shunting for idiopathic normal-pressure hydrocephalus now has randomised, placebo-controlled proof. The PENS trial (NEJM, 2025) randomised 99 patients across 17 centres to open versus placebo (closed) shunt valve settings: gait velocity improved by 0.21 m/s more with an open shunt (95% CI 0.12–0.31; p<0.001), with 80% versus 24% achieving meaningful improvement, better balance and fewer falls — but no significant improvement in cognition or continence at 3 months. Alongside this, antibody-mediated autoimmune encephalitis (LGI1, CASPR2, NMDAR, GAD65) is now an established, immunotherapy-responsive cause of subacute cognitive decline, to be suspected with rapid progression over weeks to months, seizures, hyponatraemia, movement disorder or psychiatric prominence.