Insomnia is a prevalent and persistent sleep disorder associated with significant impairments in daytime functioning, quality of life, and long-term health. It frequently co-occurs with mental and somatic conditions, yet requires independent diagnosis and targeted treatment. This review synthesizes current evidence on non-pharmacological and pharmacological therapies for insomnia across various populations. Cognitive Behavioral Therapy for Insomnia (CBT-I) remains the first-line treatment, supported by robust evidence demonstrating improvements in sleep onset, sleep maintenance, and daytime functioning. Core components, including stimulus control, sleep restriction, cognitive restructuring, relaxation techniques, and sleep hygiene, address the behavioral and cognitive mechanisms that sustain insomnia. CBT-I is effective in adults, older adults, adolescents, and perinatal populations, and often outperforms pharmacological interventions in long-term outcomes, although more research is needed to quantify sustained benefits. Pharmacological options are typically reserved for short-term or adjunctive use. Benzodiazepine receptor agonists provide modest, transient benefits but carry substantial risks such as dependence, cognitive impairment, and falls, particularly in older adults. Safer alternatives include melatonin, melatonin receptor agonists such as ramelteon, and dual orexin receptor antagonists, which offer improved sleep quality with favourable safety profiles. Off-label agents, including low-dose doxepin, trazodone, hydroxyzine, and quetiapine, show variable efficacy and require cautious use due to limited long-term data and potential adverse effects. Special considerations apply to older adults, children, and pregnant or postpartum women, for whom behavioural interventions remain the preferred approach. Contemporary evidence supports a shift toward behavioural treatments, individualized care, and reasonable short-term medication use to reduce the clinical and societal burden of insomnia.
Marek et al. (Fri,) studied this question.