Benzodiazepines

Benzodiazepines work — rapidly and reliably — for acute anxiety, panic, alcohol withdrawal, seizures and procedural sedation, which is precisely why they became some of the most prescribed drugs in history. The catch has been documented since the 1980s: tolerance and physical dependence develop at ordinary therapeutic doses within weeks, which is why UK regulators limited recommended use to 2-4 weeks as early as 1988 and most guidelines worldwide followed. The defining feature of the benzodiazepine story is not scientific uncertainty but a four-decade gap between what guidance says and what prescribing does: long-term use has persisted continuously, concentrated in precisely the populations — the elderly, the chronically anxious — where risk is highest.

Benzodiazepines depress respiration modestly alone and dangerously in combination: co-involvement in prescription-opioid overdose deaths runs around a third, and the pharmacology is straightforward enough that the FDA applied boxed warnings to both drug classes in 2016 for concurrent prescribing, then broadened benzodiazepine boxed warnings in 2020 to cover abuse, addiction, dependence and withdrawal explicitly. Co-prescription has declined but not disappeared. Alongside falls, fractures and road accidents in older adults — the reason every geriatric prescribing standard lists benzodiazepines as potentially inappropriate — this is the least contested part of the field.

Stopping after long-term use can produce rebound anxiety, insomnia, perceptual disturbances, and — after abrupt cessation at high doses — seizures; management is slow tapering, classically via diazepam substitution, over weeks to many months. A minority experience symptoms lasting far longer, a phenomenon patients organised around for decades (the Ashton Manual, written from a Newcastle withdrawal clinic, circulated worldwide through support communities long before mainstream adoption) and which research has recently begun formalising under terms like benzodiazepine-induced neurological dysfunction. As with antidepressant withdrawal, the historical pattern is uncomfortable: the affected community described the syndrome accurately years before medicine named it.

Early case-control studies associated long-term benzodiazepine use with substantially increased dementia risk, generating alarming headlines; later, better-controlled cohorts found weaker or null associations, and the confounding problem is severe — anxiety and insomnia are themselves prodromal symptoms of dementia years before diagnosis, so the drugs may mark early disease rather than cause it. Reverse causation, protopathic bias and inconsistent dose-response leave this genuinely unresolved. The honest statement: a causal dementia effect is neither established nor excluded, and the settled harms in the elderly — falls, fractures, cognitive slowing while on the drug — already justify restraint without it.

Zolpidem and its relatives were marketed as safer non-benzodiazepine hypnotics; they act on the same receptor complex and have accumulated the same problem list — dependence, next-day impairment, falls, complex sleep behaviours carrying their own boxed warning — a pharmacological rebranding more than an advance. Both drug families persist because they treat, instantly, the two commonest complaints in primary care — anxiety and insomnia — in systems where the evidence-based alternatives (CBT, CBT-I) have waiting lists measured in months. The prescription is not usually ignorance; it is the only tool available in a ten-minute consultation, which makes this as much a services problem as a pharmacology one.

Short-term, situational use — days to a few weeks, for genuine crisis — is legitimate and effective. Long-term daily use should be a rare, deliberate, reviewed decision, not an accumulated accident of repeat prescriptions. Never stop abruptly after months or years of use: seizure risk is real, and tapering should be slow, individualised and, when needed, guided by established protocols. Never combine with opioids or alcohol without explicit medical oversight. And for chronic insomnia and anxiety, the treatments with durable effect after discontinuation are the psychological ones — the drugs work while taken; CBT keeps working after it ends.