Diazepam
| Evidence Level: L4 | Predicted Indications: 10 |
Table of Contents
Diazepam: From Epilepsy and Anxiety to Insomnia
One-Sentence Summary
Diazepam is a benzodiazepine originally used for epilepsy, status epilepticus, alcohol withdrawal, anxiety and tetanus. The TxGNN model predicts it may be effective for Insomnia, with 0 registered clinical trials and about 18 publications (mostly reviews and indirect studies) currently supporting this direction. Evidence is limited, so this is best treated as a research question rather than an actionable repurposing candidate.
Quick Overview
| Item | Content |
|---|---|
| Original Indication | Epilepsy, status epilepticus, alcohol withdrawal (including delirium), Wernicke-Korsakoff syndrome, anxiety, tetanus (from the knowledge graph; approved indication text is not available for the Singapore licences) |
| Predicted New Indication | Insomnia |
| TxGNN Prediction Score | 99.99% |
| Evidence Level | L4 |
| Singapore Market Status | ✓ Marketed |
| Number of Registrations | 8 |
| Recommended Decision | Hold |
Why is This Prediction Reasonable?
Detailed mechanism-of-action data is not available in the data compiled for this report. Based on well-established pharmacology, diazepam is a positive allosteric modulator of the GABA-A receptor. It enhances inhibitory neurotransmission, which produces sedative, anxiolytic, anticonvulsant and muscle-relaxant effects.
Its original indications (anxiety, seizures, alcohol withdrawal) all involve excessive neuronal excitability. Insomnia, especially when linked to anxiety or hyperarousal, shares this problem. Benzodiazepines have long been used for short-term insomnia, so the model's prediction is mechanistically coherent.
However, the supporting records contain no clinical trials. The literature is mainly reviews, herbal and acupuncture studies, and papers on safety and pharmacokinetics. Long-term use is associated with dependence, tolerance and altered sleep architecture, so the risk-benefit balance is a real concern. Insomnia is also already served by other approved drug classes, so any new role for diazepam would be narrow.
Clinical Trial Evidence
Currently no related clinical trials registered.
Literature Evidence
| PMID | Year | Type | Journal | Key Findings |
|---|---|---|---|---|
| 6113175 | 1981 | Double-blind comparison | J Int Med Res | Lormetazepam 1 mg was significantly better than diazepam 5 mg at reducing time to fall asleep in 100 outpatients with insomnia over 7 days |
| 39581171 | 2024 | Review | Bioorg Chem | Reviews GABA-A receptor modulators, including diazepam, for epilepsy, anxiety, insomnia and depression; notes sedation and other side effects |
| 36692463 | 2023 | Meta-analysis | Acta Pharm | Assesses tranquilizers in elderly patients with chronic diseases, with attention to dose and adverse effects |
| 40583063 | 2025 | Cohort | Cell Mol Biol Lett | Long-term use of benzodiazepines and Z-drugs (including diazepam) is linked to increased breast cancer risk |
| 23330992 | 2013 | Review | Expert Opin Drug Metab Toxicol | Pharmacokinetics of anxiolytic drugs |
| 31770357 | 2019 | Review | Med Lett Drugs Ther | Review of cannabis and cannabinoids |
| 35196378 | 2022 | Non-randomised intervention | Fam Pract | Discontinuation of chronic benzodiazepine use in primary care |
| 37984023 | 2024 | Predictive model | Value Health Reg Issues | Trends and economic burden of benzodiazepine use; long-term use linked to memory loss, dependence and falls |
Related evidence from the broader "sleep disorder, initiating and maintaining sleep" prediction (rank 3, also L4) includes a randomised trial of masked tapering plus behavioural therapy to discontinue benzodiazepine receptor agonists (39374004, JAMA Intern Med, 2024) and a meta-analysis comparing newer hypnotics with benzodiazepines including diazepam (15252823, Hum Psychopharmacol, 2004). These largely address deprescribing and comparative efficacy, not new support for diazepam as a hypnotic.
Other Predicted Indications (Summary)
| Rank | Predicted Indication | Score | Evidence Level | Assessment |
|---|---|---|---|---|
| 2 | Cauda equina syndrome | 99.99% | L5 | Model prediction only; a surgical emergency not plausibly modified by diazepam |
| 3 | Sleep disorder, initiating and maintaining sleep | 99.97% | L4 | Class use is established, but literature concerns tapering and chronic-use harms |
| 4 | ADHD, inattentive type | 99.97% | L5 | No evidence; sedation may worsen inattention |
| 5 | Barbiturate abuse | 99.95% | L4 | Indirect support via cross-tolerance in withdrawal management; no specific trials |
| 6 | Hallucinogen abuse | 99.95% | L4 | Symptomatic use in intoxication only; no efficacy evidence for the disorder |
| 7 | Antidepressant type abuse | 99.95% | L4 | Likely a disease-class artifact; diazepam is the dependence-causing agent in the literature |
| 8 | Specific developmental disorder | 99.95% | L5 | Preclinical papers only |
| 9 | ADHD | 99.93% | L5 | Paradoxical reactions reported in children; no supporting trial |
| 10 | Neurogenic bladder (obsolete entry) | 99.91% | L5 | Model prediction only; the disease entry is obsolete |
Singapore Market Information
| Authorization Number | Product Name | Dosage Form | Approved Indication |
|---|---|---|---|
| SIN08242P | Diazepam Lipuro Injection 10 mg/2 ml | Injection | Not available |
| SIN00044P | AA Pharma Diazepam Tablet 5 mg | Tablet | Not available |
| SIN02309P | Stesolid Enema 5 mg/2.5 ml | Enema | Not available |
| SIN00042P | AA Pharma Diazepam Tablet 2 mg | Tablet | Not available |
| SIN00421P | Diapine Injection 5 mg/ml | Injection | Not available |
All 8 licences are valid single-ingredient products (injection, tablet and enema forms). Three further licences (SIN00424P Diapine Tablet 2 mg, SIN00045P AA Pharma Diazepam Tablet 10 mg, SIN08909P Diapine Tablet 5 mg) are not shown above.
Safety Considerations
- Contraindications suggested by the knowledge graph (not taken from package insert text): liver failure and liver disease, respiratory failure and lung disease, sleep apnoea syndromes (including obstructive and congenital central hypoventilation), open-angle and angle-closure glaucoma, myasthenia gravis and other neuromuscular junction diseases, gallbladder disease, anxiety/neurotic/dysthymic disorders, and substance-related disorders or substance abuse/dependence.
- Drug interactions: 100 interactions were found in total. Notable examples:
- Major: morphine (additive respiratory and CNS depression).
- Moderate: bupropion, aprepitant, omeprazole, esomeprazole, cimetidine, clarithromycin, cisapride, clotrimazole, miconazole, dronabinol, metoclopramide.
- Minor: dexamethasone, betamethasone and several antacids (magnesium and aluminium salts, calcium carbonate).
- Literature-derived concerns: dependence and withdrawal, tolerance, altered sleep architecture with chronic use, cognitive decline and falls in older adults, paradoxical reactions in children, and a possible association with breast cancer risk with long-term use.
Please refer to the package insert for authoritative warnings and contraindications.
Conclusion and Next Steps
Decision: Hold
Rationale: The prediction for insomnia is mechanistically sensible and the drug is widely marketed in Singapore, but there are no clinical trials, and the literature is indirect and focused on the risks of long-term benzodiazepine use. Package insert safety data have also not yet been obtained, so the safety screening cannot proceed. The evidence level is L4, which supports treating this as a research question only.
To proceed, the following is needed:
- Package insert warnings and contraindications from the HSA website (a blocking gap)
- Mechanism of action data from DrugBank
- Diazepam-specific controlled trials in insomnia compared with current standard hypnotics
- A benefit-risk assessment addressing dependence, tolerance, and use in older adults
- Approved indication text for the Singapore licences, to confirm which uses are already on-label
Disclaimer
This content is for research purposes only and does not constitute medical advice. Clinical validation is required before any clinical application.