Why Silexan Matters
Silexan is one specific lavender oil, steam-distilled from Lavandula angustifolia flowers to a fixed specification and sold in 80 mg softgels. About 71% of it is two monoterpenes, linalool (36.8%) and linalyl acetate (34.2%). The randomized trials behind oral lavender's reputation in anxiety disorders all used this extract (code name WS 1265), and the results belong to it alone. The sponsor-linked pooled analysis of the subthreshold trials says the results cannot be generalized to other lavender oil products.
Its legal status changes at the border, and that matters for how you recommend it. In Germany it is Lasea, a licensed medicine since 2009, sold in pharmacies without a prescription for restlessness with anxious mood. In the United States it is a dietary supplement, CalmAid from Nature's Way. Nothing about it is FDA-approved. The FDA does not review supplements for efficacy before they are sold, and the CalmAid box makes structure/function claims ("Reduces occasional nervousness"), not treatment claims. Other lavender capsules on the same shelf are not Silexan, and the trial data do not transfer to them.
The authors of the 2023 pooled analysis, who include sponsor employees, call it probably the best-researched herbal medicine for anxiety, which may well be true, and the evidence is still narrower than its reputation. There are six randomized trials in anxiety and one in depression. The manufacturer (Dr. Willmar Schwabe) sponsored every one, and all the anxiety trials were run in Germany. Pooled across the five placebo-controlled anxiety trials, 80 mg/day beat placebo by 2.9 points on the Hamilton Anxiety Rating Scale (HAMA), a standardized effect of 0.35. In the best-known effect-size analysis of GAD trials (Hidalgo 2007), SSRIs came in at 0.36, from different trials, populations and decades. The authors of the Silexan pooled analysis describe their 0.35 as in the range of SSRIs and SNRIs, which is a cross-trial comparison and not a head-to-head result. The 0.35 includes a dose-finding trial that failed. In the one GAD trial that tested 160 mg, that arm had the largest numerical effect, though it was not formally compared with 80 mg.
Tolerability is the real case for it. Lavender-flavored belching is the side effect patients notice, in roughly 7% to 20% of trial patients. It did not impair simulated driving in healthy volunteers (single doses up to 320 mg, 80 mg daily for a week), it rated like placebo in an abuse-liability study at 640 mg, abrupt stopping produced no withdrawal, and a 160 mg drug-cocktail study found no clinically relevant CYP effects.
Silexan is a reasonable choice for an adult with mild to moderate GAD or subthreshold anxiety who wants to avoid a prescription drug or could not tolerate one. Expect a modest effect at best, numerically similar to SSRIs' in cross-trial comparison, building over weeks, from an evidence base that one company produced. Tell the patient to buy CalmAid (or another box that names Silexan or WS 1265), consider 160 mg for full GAD because that dose has the stronger GAD result (practice), and measure the anxiety before and after.
Part 1: Indications
Regulatory Status
There is no FDA label, so there is no FDA-approved use and, strictly, no "off-label" either. Here is what each regulator says.
- United States: dietary supplement. CalmAid (Nature's Way) lists Silexan 80 mg per softgel. Its current label claims are "Feel more relaxed," "Reduces occasional nervousness," "Supports restful sleep when disrupted by stress" and "Helps manage stress & tension," with the FDA disclaimer. Lavela WS 1265 (Integrative Therapeutics) is an 80 mg softgel whose name carries the WS 1265 extract code, sold mainly through practitioner channels. Neither has been reviewed by the FDA for any indication.
- Germany: licensed medicine, no prescription. Lasea 80 mg, authorized June 2009 and renewed October 2021, for "states of restlessness with anxious mood" (Unruhezustände bei ängstlicher Verstimmung) in adults. One capsule daily, for no more than 3 months. Pharmacy-only, not reimbursed.
- Austria and Switzerland. Austria: "temporary anxious mood" in adults, without prescription. Switzerland: "anxiety, restlessness and associated sleep disturbance," non-prescription category D, alongside a second, reimbursed prescription product; the Swiss prescribing information does not cap total duration, though the non-prescription product's entry asks for a doctor visit if there is no improvement within 2 weeks.
- EU herbal monograph. The EMA's Committee on Herbal Medicinal Products adopted a monograph on lavender oil in 2012 for traditional use only: "relief of mild symptoms of mental stress and exhaustion and to aid sleep," 20 to 80 mg/day by mouth, not under age 12. It declined well-established use because too few patients had been studied in RCTs and the lorazepam trial had no placebo arm. A periodic review opened with a call for data in 2020; no revised monograph had been published as of November 2026.
- Guidelines. The 2022 WFSBP and CANMAT taskforce on nutraceuticals and phytoceuticals rated lavender "provisionally recommended" for anxiety disorders and "weakly recommended" for unipolar depression. One of the Silexan trialists was a co-author. Germany's own 2021 S3 guideline on anxiety disorders went the other way: it rated the GAD evidence level 1b but made no recommendation, because it set aside the lorazepam trial, was left with one positive placebo-controlled GAD trial, and found no long-term or relapse-prevention studies.
Uses Supported by Trials
- Generalized anxiety disorder (three RCTs: one positive against placebo, one negative, one against lorazepam with no placebo arm; plus pooled analyses)
- Subthreshold anxiety: anxiety disorder not otherwise specified, and restlessness with anxiety-related insomnia (two RCTs)
- Mixed anxiety and depressive disorder (one RCT)
- Mild to moderate major depressive disorder (one RCT, versus sertraline 50 mg and placebo)
- Anxiety-related sleep complaints, as a secondary outcome of the anxiety trials
Not supported by any controlled trial: panic disorder, social anxiety disorder, OCD, PTSD (one open-label study; a placebo-controlled trial is under way), treatment-resistant anxiety or depression, bipolar disorder, and anyone under 18.
The Evidence by Indication
| Trial | Population | Arms (n analyzed) | Weeks | Result |
|---|---|---|---|---|
| Kasper 2010 | Anxiety disorder NOS (subthreshold) with disturbed sleep; general and psychiatric practices | 80 mg (104) vs placebo (108) | 10 | HAMA fell 16.0 vs 9.5 points; response 76.9% vs 49.1% |
| Kasper 2015 | Restlessness with anxiety and disturbed sleep | 80 mg (86) vs placebo (84) | 10 | HAMA fell 12.0 vs 9.3 (p=0.03); response 48.8% vs 33.3% |
| Kasper 2016 | Mixed anxiety and depressive disorder | 80 mg (159) vs placebo (156) | 10 | HAMA fell 10.8 vs 8.4; MADRS fell 9.2 vs 6.1 |
| Woelk 2010 | GAD | 80 mg (40) vs lorazepam 0.5 mg/day (37); no placebo | 6 | HAMA fell 11.3 vs 11.6 |
| Kasper 2014 | GAD | 160 mg (121), 80 mg (135), paroxetine 20 mg (132), placebo (135) | 10 | HAMA fell 14.1, 12.8, 11.3 and 9.5; both Silexan doses beat placebo; paroxetine did not in the full analysis set (p=0.10) |
| Dose-finding trial (pooled report, Kasper 2017) | GAD | 10, 40 and 80 mg vs placebo (80 mg 103, placebo 102) | 10 | Negative: 80 mg was 0.5 points better than placebo (p=0.16) |
| Kasper 2024 | Mild to moderate MDD (MADRS 19 to 34) | 80 mg (170), sertraline 50 mg (171), placebo (157) | 8 | MADRS 2.17 points better than placebo for Silexan, 2.59 for sertraline |
The 2010, 2014 and 2016 trials tested their primary outcome one-sided. All three reported p under 0.01, so the results would also hold on the usual two-sided test.
Generalized anxiety disorder. The largest trial (Kasper et al., Int J Neuropsychopharmacol 2014; 539 adults, 57 German practices) gives the best-case numbers. At 160 mg, response (HAMA fall of at least 50%) was 60.3% vs 37.8% on placebo and remission (HAMA under 10) 46.3% vs 29.6%. Entry was limited to ages 18 to 65, and no other psychotropic drug was allowed during the trial or for 30 days before it. Paroxetine 20 mg missed separation from placebo in the main analysis (p=0.10) but did separate in the observed-cases analysis (p<0.01). Silexan and paroxetine were compared only descriptively, with no prespecified test, so the trial does not show that Silexan beat paroxetine, and it does not show the opposite either. The second placebo-controlled GAD trial, a dose-finding study of 10, 40 and 80 mg, was negative, which the sponsor attributes to a placebo response as large as the active arms in other trials, and it appears only inside a pooled report. That pooled analysis of 925 patients concluded that 160 mg is effective and that 80 mg may be the lower end of the therapeutic range. The lorazepam comparison (Woelk and Schläfke, Phytomedicine 2010) enrolled 77 patients, had no placebo arm and no power calculation (the EMA assessors' point), and gave lorazepam 0.5 mg once a day, a starting dose of a short-acting drug. HAMA fell equally in both arms and response favored Silexan (52.5% vs 40.5%, a secondary outcome), but a trial built this way proves little either way.
All five placebo-controlled anxiety trials at 80 mg. An individual-patient meta-analysis (Dold et al., Eur Arch Psychiatry Clin Neurosci 2023; 1,213 randomized) found a 2.9-point HAMA advantage (SMD 0.35, 95% CI 0.13 to 0.56). Response was 51.8% vs 38.8% (NNT 8) and remission (HAMA under 7) 28.4% vs 22.2%. Four of five trials were positive, heterogeneity was substantial (I² 72.7%), and participants were 18 to 65. A later patient-level analysis of the same trials found the effect grew with baseline severity, within the range the trials enrolled. The subthreshold trials alone (Möller et al. 2019, 697 patients) gave a 3.83-point advantage.
Depression. One 8-week trial (Kasper et al., Eur Arch Psychiatry Clin Neurosci 2024; Germany and Poland) randomized 498 adults with a mild or moderate episode. Response was 53.5% on Silexan 80 mg, 54.0% on sertraline 50 mg and 41.5% on placebo; remission was 44.4%, 45.2% and 32.6%, with an NNT of 9 for Silexan and 8 for sertraline. The response and remission differences from placebo were descriptive, not part of the primary analysis. The trial excluded anyone reporting suicidal thoughts and patients who had not responded to an antidepressant in the current episode (or to sertraline 50 mg or more for 6 weeks in an earlier one), used the European 80 mg dose, held sertraline at 50 mg, was not designed to compare the two drugs, and has not been replicated. Treat this as one positive trial in mild illness.
Sleep. Sleep scores improved in the anxiety trials, but a mediation analysis of the subthreshold trial attributed 98.4% of the sleep effect to reduced anxiety and found no direct effect. It is not a hypnotic. A German database study found fewer repeat visits for insomnia after Silexan than after Z-drugs (15.6% vs 28.6%), an observational result with obvious confounding.
PTSD and other conditions. An open-label 6-week study of 50 patients with neurasthenia, PTSD or somatization disorder reported improvement, with no control group. An Australian placebo-controlled trial of adjunctive 160 mg in PTSD (STOP, NCT06412757, target 278 patients, funded by the US Department of Defense) published its protocol in 2026; results are not yet available.
Every efficacy trial and the safety studies come from one sponsor and an overlapping group of investigators, and a company statistician ran the patient-level pooled analyses. Independent meta-analyses exist, but they pool the same sponsor trials, and one independent review rated the trials at low risk of bias in most domains apart from that sponsorship. The 2026 network meta-analysis that ranked it well among standard anxiolytics was supported by a Schwabe grant. The authors of the Australian PTSD trial describe it as the first Silexan trial run by a team independent of the manufacturer. Read the numbers as an upper estimate until an independent result appears, and do not stretch them to patients the trials excluded, such as people over 65 and those with other significant psychiatric disorders.
Who Is a Good Candidate?
- An adult with mild to moderate GAD or subthreshold anxiety who declines an SSRI or SNRI, or who stopped one for side effects
- A patient for whom sedation, sexual side effects, weight gain or dependence would be a real problem, including someone with a substance use history
- Anxiety with low mood below the threshold for a major depressive episode
- A patient on several medicines, where you want something with clean drug-interaction data
- Anxiety-driven poor sleep when you would rather not add a hypnotic
Who Is a Poor Candidate?
- Severe, disabling GAD as the only treatment; trial patients averaged HAMA scores in the mid-20s, and the largest trial required a psychic anxiety subscore of 21 or less (of a possible 28) at entry
- Panic disorder, social anxiety disorder, OCD or PTSD as the only treatment; there are no controlled data in these
- Moderate to severe or treatment-resistant depression, or depression with suicidal thoughts; the one depression trial enrolled mild to moderate episodes and excluded prior nonresponders and anyone with suicidal thoughts
- A patient who needs relief this week; separation from placebo took 2 to 6 weeks
- Anyone under 18, pregnant or breastfeeding, or with liver disease (Lasea contraindications or cautions, below)
- A patient with bipolar disorder in whom you need a proven agent; it has not been studied there
Part 2: Mechanism
How Silexan works is not settled. The measured findings, most of them preclinical:
- Voltage-gated calcium channels: in cell systems (as summarized in a sponsor-linked review) it reduced N-type, T-type and P/Q-type calcium currents by about 20% each. Pregabalin inhibited P/Q channels about twice as much in the same work, and Silexan does not bind the α2δ subunit that pregabalin and gabapentin target.
- Serotonin: after at least 8 weeks of 160 mg, PET in 17 healthy men showed lower 5-HT1A receptor binding potential in the temporal and fusiform gyri, hippocampus, insula and anterior cingulate. In rats it raised extracellular serotonin, dopamine and norepinephrine in prefrontal cortex without blocking the serotonin transporter.
- Not benzodiazepine-like: rats trained to recognize diazepam treated Silexan as saline, which fits the human abuse-liability and driving results.
- Neuroplasticity: neurite outgrowth and synapse formation in cell cultures, and antidepressant-like behavior in rodents. These are preclinical findings.
Pharmacokinetics. The product information tracks linalool, the main constituent. It peaks about an hour after a capsule, plasma levels fall within 5 hours, and the terminal half-life is about 6 to 7 hours. In severe renal impairment, exposure (AUC) was 1.96 times normal, judged clinically irrelevant. In liver cirrhosis, peak levels and exposure rose, the basis for the German liver contraindication. The trials still dosed once daily, and the clinical effect builds over weeks.
The short half-life does not make it a rescue medicine. No study tested it as needed, and it separated from placebo only after 2 to 6 weeks of daily use. Prescribe it the way you would an SSRI: every day, judged over weeks (practice).
Part 3: Before You Start: Workup and Candidacy
No labs are required. The workup is about the diagnosis, the contraindications and the product.
| Assessment | Why |
|---|---|
| Diagnosis and baseline severity (GAD-7, HAMA, or a 0 to 10 rating) | The effect is modest and gradual; without a baseline number you will not know if it worked (practice) |
| Depression and suicide risk | Moderate or severe depression needs a proven antidepressant; Silexan was tested only in mild to moderate episodes without suicidal thoughts |
| History of mania or hypomania | No data in bipolar disorder |
| Liver disease (history; LFTs only if suspected) | Lasea: contraindicated in liver dysfunction; cirrhosis raised exposure |
| Severe kidney disease or dialysis | Lasea: not recommended in dialysis (no data) |
| Pregnancy, plans for pregnancy, breastfeeding | No human data; European and US labels advise against use |
| Age | Lasea: contraindicated under 18; CalmAid: not under 12 |
| Allergy to lavender or soy | Serious hypersensitivity reactions are listed; CalmAid contains soy |
| Alcohol, benzodiazepines, other sedatives; all supplements | Lasea advises against combining with other anxiolytics, sedatives or alcohol, as a precaution |
| Breast or other hormone-sensitive cancer | No data; discuss with the oncologist (practice) |
| Planned surgery | One case report of hard intravenous induction of anesthesia; tell the anesthesiologist (practice) |
Ask what the patient has already tried. Many have taken lavender tea, sprays, oils or a different lavender capsule, and none of those tells you how Silexan would do.
Part 4: How to Start and Dose
Dosing at a glance
| Situation | Product information or trial dose | Notes |
|---|---|---|
| FDA-labeled dose | None; Silexan is not an FDA-approved drug | US CalmAid label: 1 softgel (80 mg) once or twice daily |
| European label (Lasea, Germany) | 80 mg once daily, adults | No longer than 3 months; see a doctor if no better after 1 month |
| Subthreshold anxiety, mixed anxiety and depression, MDD | Trial dose 80 mg once daily | 10-week anxiety trials (the lorazepam comparison ran 6 weeks); 8-week MDD trial |
| GAD | Trial doses 80 or 160 mg once daily | 160 mg had the larger numerical effect in the one trial that tested it; not formally compared with 80 mg |
| Usual start | 80 mg once daily | Practice: for full GAD, move to 160 mg (two softgels together) after 1 to 2 weeks if tolerated |
| Highest dose studied | 160 mg/day for weeks; single doses to 640 mg in healthy volunteers | No efficacy data above 160 mg/day |
| How to take it | Same time each day, swallowed whole with a full glass of water, upright | With food if it causes belching (CalmAid label) |
| Liver disease | Contraindicated (Lasea) | Cirrhosis raised exposure |
| Severe renal impairment | No adjustment stated (Lasea) | Exposure about doubled; not recommended in dialysis |
| Age over 65 | No separate dose | Limited data; anxiety trials enrolled ages 18 to 65 |
| Under 18 | Contraindicated (Lasea); not under 12 (CalmAid) | No pediatric trials |
Which Product to Buy
Tell the patient the brand, not "lavender." In the US that is CalmAid (Nature's Way), one 80 mg Silexan softgel, sold in 30- and 90-count packs online and in stores. Lavela WS 1265 (Integrative Therapeutics) is the same extract in 80 mg softgels. In Germany, Austria and Switzerland it is Lasea 80 mg, and in Australia it is distributed as Seremind. Any other lavender oil capsule is a different oil at an unknown dose. CalmAid names Silexan on its Supplement Facts panel, and Lavela carries the code WS 1265 in its name; a box that shows neither is not the studied product. Both US brands belong to the sponsor: Schwabe has owned Nature's Way outright since 2001 and added Integrative Therapeutics in 2008. Patients buy it over the counter and pay out of pocket, since insurance generally does not cover supplements, and 160 mg/day takes two softgels a day, so it costs twice as much as 80 mg.
Starting and Titrating
- Subthreshold anxiety or mild symptoms: 80 mg once daily, the dose every subthreshold trial used.
- Full GAD: start 80 mg and go to 160 mg once daily after a week or two (practice). That is above the one-capsule Lasea dose but within the CalmAid label's two softgels a day. In the 2014 trial the 160 mg arm had the larger numerical effect, and adverse events were no more frequent than on 80 mg or placebo. Starting at 160 mg is also defensible (practice); that arm separated from placebo by week 4.
- Added to an antidepressant: no completed trial has tested the combination. It is reasonable on interaction grounds (Part 8), but you are on your own for efficacy.
- Timing: once daily at about the same time (Lasea); neither label specifies morning or evening. For belching, take it with a meal and a full glass of water; some clinicians also move it to the evening (practice, not tested).
How Long Before Judging It
Separation from placebo appeared at about 2 weeks in the subthreshold trial, 4 weeks at 160 mg in GAD, 4 weeks in mixed anxiety and depression, and 6 weeks at 80 mg in GAD; in the mixed anxiety and depression trial the gap kept growing to week 10. Give it 6 to 8 weeks at 160 mg before calling it a failure in GAD (practice). The German label caps treatment at 3 months because no controlled trial ran longer. Swiss and US labeling set no limit. In a Swiss prescription database, 27.6% of users were treated for more than 3 months, and German and Swiss records showed no new safety signal with longer use (a Schwabe-funded analysis).
Missed dose
- One dose: neither label gives a rule. Take it later that day or skip it, and never double up (practice, not label).
- Several days: it clears within a day or two, and abrupt stopping caused no withdrawal. Whether a gap costs benefit is unstudied.
- Restarting after a gap: resume the usual dose (practice, not label).
What to tell the patient at the first prescription
- What to expect and when: if it helps, worry eases over 2 to 6 weeks; we judge it at 6 to 8. Studies showed a modest benefit, and the maker paid for all of them.
- Early side effects: lavender-tasting burps in up to 1 in 5; take it with a meal and water. It was not sedating in studies.
- Call me about: swelling, rash or trouble breathing (stop it), worse mood or suicidal thoughts, or a pregnancy.
- How to take it: every day, swallowed whole, not just on bad days; buy CalmAid or a box naming Silexan or WS 1265.
- Next visit: 2 to 4 weeks, with your 0 to 10 score.
Part 5: Monitoring
There are no labs to follow and no drug levels. Monitoring is clinical, and most of it is about whether it is working.
| Parameter | Baseline | Follow-up |
|---|---|---|
| Anxiety rating (GAD-7 or 0 to 10) | ✓ | At 2 to 4 weeks, at 6 to 8 weeks, then each visit (practice) |
| Mood and suicidal thoughts | ✓ | Each visit; switch to a proven antidepressant if depression deepens |
| Belching, nausea, other GI symptoms | Not needed | First weeks |
| Rash, swelling, breathing symptoms | Allergy history | Stop at once if they occur |
| Liver tests | Only if liver disease is suspected | Not routine |
| Product actually taken, refills, other supplements | ✓ | Each visit |
| Pregnancy plans | ✓ | Each visit in people who could become pregnant |
Because it is bought off the shelf, patients switch brands, run out and forget to reorder. Ask which box they are using.
If it isn't working
- First, check: the box (CalmAid, Silexan or WS 1265), daily rather than as-needed use, 160 mg in full GAD, and 6 to 8 weeks on it.
- At 6 to 8 weeks: compare the score with baseline. Trial response meant a 50% fall in HAMA; a result well short of that is a reason to move on (practice).
- Do not: go above 160 mg/day (no efficacy data) or wait while depression deepens.
- Then: an SSRI or SNRI first, or pregabalin or buspirone (Part 11). Silexan can stop the day the new drug starts.
Part 6: Side Effects and How to Manage Them
In the pooled placebo-controlled trials, adverse events (OR 1.16), serious adverse events and dropouts for adverse events did not differ from placebo. The rates below come from the individual trials.
Belching and Other GI Effects
Eructation, often with a lavender taste or breath odor, is the one consistent drug effect. It was 7.0% vs 0% in the restlessness trial, 20.0% vs 0% in mixed anxiety and depression, 7.5% in the lorazepam comparison and 16.5% in the depression trial. The Lasea label lists it as common. Nausea was 10.0% vs 2.7% against lorazepam and 3.8% vs 1.3% in mixed anxiety and depression. In the 2014 GAD trial, the excess risk of any GI event over placebo was 4.4% at 160 mg and 4.5% at 80 mg, against 8.0% for paroxetine; the Silexan confidence intervals included zero.
- Management: swallow it whole (do not chew), with a full glass of water and food, upright. It is mild and rarely a reason to stop.
Allergic Reactions
The Lasea label lists serious hypersensitivity reactions with swelling, circulatory or breathing symptoms, and allergic skin reactions, frequency not known. Stop it and seek care if they appear.
Sedation, Driving and Cognition
In a simulator study in healthy volunteers, 80 mg daily for a week was equivalent to placebo on lane-keeping, and single doses of 160 and 320 mg were clearly better than lorazepam 1 mg. The label adds that no driving data exist past age 58 and that anyone who feels impaired should not drive. In the lorazepam trial, fatigue was 0% on Silexan vs 16.2% on lorazepam.
Weight, Sexual Function and Sleep
A manufacturer-linked review states that Silexan has not been shown to cause weight gain, sexual dysfunction or sleep disturbance. The trials were short and were not designed to measure these closely, so tell patients they were not reported, with that caveat.
Part 7: Toxicity, Overdose, and Warnings
Silexan has no FDA label and so no boxed warning. The German Lasea product information lists three contraindications: liver dysfunction, hypersensitivity to lavender oil or an ingredient, and age under 18. It is not recommended in pregnancy or dialysis, breastfeeding women should not take it, and it should not be combined with other anxiolytics, sedatives or alcohol, as a precaution, because clinical interaction data are lacking. The US CalmAid label says not to use it if pregnant, nursing or under 12, and to consult a clinician if taking any medication.
Lavender, Tea Tree Oil and Breast Growth in Children
In 2007, Henley and colleagues (NEJM) described three otherwise healthy boys, aged 4 to 10, with prepubertal gynecomastia that began with topical products containing lavender and tea tree oils (a balm, a hair gel, soap) and resolved after the products were stopped. Both oils were weakly estrogenic and antiandrogenic in cell lines. By 2023 the case series totaled eight boys and four girls, all exposed through fragrances or skin products, and a review in that year concluded that causality has not been shown. Guideline-standard testing of linalool and linalyl acetate in cell assays, and of linalool in animals, found no relevant estrogenic or androgenic activity (the testers were employees of a company that makes both compounds). A US parent survey of 556 children found no significant difference in endocrine disorders between exposed and unexposed children, but it identified so few cases that the risk ratio of 2.8 had a 95% confidence interval of 0.35 to 22, far too wide to settle the question.
None of these reports involved oral Silexan or adults. In 24 women on a combined pill, 160 mg/day did not change ethinyl estradiol or levonorgestrel levels, and estradiol, progesterone, follicle size and an ovarian-activity score were similar to placebo. The reports do not establish a hormonal risk for adults taking Silexan. They are one more reason to keep it out of children, and the German label already contraindicates it under 18.
Hormone-Sensitive Cancers
There are no data. Given the in vitro signal, however weakly it has held up, it is reasonable to discuss it with the oncologist before using it in someone with an estrogen-receptor-positive cancer (practice judgment).
Dependence, Abuse and Withdrawal
- Abuse liability: in 34 evaluable recreational sedative users, single doses of 80 and 640 mg rated like placebo on drug liking and differed clearly from lorazepam 2 and 4 mg (p<0.001).
- Withdrawal: after 10 weeks at 80 or 160 mg in the 2014 GAD trial, Silexan was stopped abruptly, and the Physician Withdrawal Checklist showed no withdrawal over the next week.
Overdose
The product information reports no cases of overdose. Single 640 mg doses in healthy volunteers were well tolerated; the only events more frequent than with the comparison treatments were mild belching and taste change. Treat supportively and look for co-ingested drugs, which are the likelier danger. Poison Help (1-800-222-1222) can advise.
Part 8: Drug Interactions
The two formal interaction studies, both at 160 mg/day, found no clinically relevant interaction (one omeprazole measure slightly exceeded its prespecified limit). The gaps are pharmacodynamic: nobody has studied it with sedatives, and none of the efficacy trials tested it as an add-on.
| Interaction | Effect | Risk | Action |
|---|---|---|---|
| Combined oral contraceptive (ethinyl estradiol 0.03 mg / levonorgestrel 0.15 mg) | 160 mg/day for two cycles: AUC ratios 0.97 (ethinyl estradiol) and 0.94 (levonorgestrel); no sign of lost contraceptive effect in 24 women | LOW | No change. Other hormonal methods were not studied |
| CYP1A2, 2C9, 2D6 and 3A4 substrates | No relevant change in caffeine, tolbutamide, dextromethorphan or midazolam after 11 days of 160 mg (16 volunteers) | LOW | No adjustment |
| CYP2C19 substrates | Omeprazole AUC ratio's 90% CI upper bound slightly above the 1.43 limit, attributed to omeprazole's variability; other CYP2C19 measures normal | LOW | No adjustment |
| Alcohol, benzodiazepines, other sedatives and anxiolytics | No clinical interaction data; Silexan was not sedating on its own | LOW | Lasea label advises against combining, as a precaution. Counsel and watch |
| SSRIs, SNRIs and other antidepressants | No completed controlled trial of the combination; no CYP interaction expected | LOW | Reasonable to combine (practice); efficacy as an add-on is unproven and pharmacodynamic interactions were not studied |
| General anesthesia | One case report: intravenous propofol and thiopental failed to induce anesthesia in an 18-year-old on long-term Silexan; inhaled sevoflurane worked | LOW | Tell the anesthesiologist (practice); causality unproven |
The clean interaction data are a practical reason to choose it in a young woman on the pill or an older patient on many drugs. "Clean" applies to the drugs studied. For anything else, the CalmAid label's own advice, check with a clinician before combining, is the honest position.
Part 9: Special Populations
Pregnancy
There are no human data. Animal studies showed no reproductive toxicity, according to the Lasea product information, which still advises against use in pregnancy. The EMA monograph also advises against use in pregnancy for lavender oil in general and notes that its reproductive toxicity has not been adequately tested, and the CalmAid label says not to use it if pregnant. Choose an agent with pregnancy data, and if a patient conceives while taking it, stopping is simple (Part 10). For patient information, MotherToBaby (mothertobaby.org) is a reasonable resource.
Lactation
It is not known whether linalool or other constituents pass into milk. The Lasea label says breastfeeding women should not take it, the Swiss information lists breastfeeding as a contraindication, and the CalmAid label says not to use it while nursing.
Elderly
- The anxiety trials enrolled adults up to 65, and the label notes limited data beyond that. Sponsor pooled analyses found effects in patients 60 and older at least as large as in younger adults, but those patients were few and none was over 65. The depression trial enrolled patients up to their early 80s.
- The lack of sedation, driving impairment and CYP effects is attractive in older patients on several drugs, though no study measured falls.
Renal Impairment
No dose change is stated. In 12 patients with severe renal impairment, linalool exposure was 1.96 times normal, which the product information calls clinically irrelevant. Use in dialysis is not recommended because there are no data.
Hepatic Impairment
Contraindicated in liver dysfunction in Germany and in liver insufficiency in Switzerland. In a single-dose study across Child-Pugh A to C cirrhosis, peak levels and exposure were higher than in healthy volunteers.
Pediatric / Adolescent
There are no pediatric trials. The German label contraindicates it under 18, the Swiss label sets 18 as the minimum, and the US CalmAid label says not to use it under 12. A German prescription database recorded 483 adolescents aged 12 to 17 who were prescribed it, with no rise in diagnoses suggesting adverse effects, but that is observational and small. Avoiding it under 18 is the cautious course, given the German contraindication and the topical lavender reports in Part 7 (practice judgment).
Bipolar Disorder
No trial enrolled patients with bipolar disorder, so its effect on mood switching is unknown. It did show antidepressant activity in MDD, so stabilize the mood disorder first and watch for switching as you would with an antidepressant (practice).
Part 10: Discontinuation and Taper
No taper is needed. After 10 weeks at 80 or 160 mg, abrupt stopping produced no withdrawal symptoms on the Physician Withdrawal Checklist. Linalool's half-life is 6 to 7 hours, so it is gone within a day or two.
- Relapse: there is no relapse-prevention trial, so nobody knows how long to continue. If it worked, reassess at 3 to 6 months and consider a trial off it with a rating scale in hand (practice).
- Fading benefit: some clinicians see many patients do well for the first month or two and then lose some of the effect. This has not been studied; recheck the product, the dose and the diagnosis before writing it off (practice observation, not trial data).
- Switching: because it has no withdrawal or known interactions with antidepressants, it can be stopped the day another treatment starts, or overlapped (practice).
- From an SSRI or SNRI onto Silexan: taper the antidepressant on its own schedule; Silexan was never studied for discontinuation symptoms. It can start during the taper and be judged on its own 6 to 8 week clock (practice).
- During a benzodiazepine taper: no data, and the German label advises against combining it with other anxiolytics. It is not diazepam-like in animals, so do not expect it to cover withdrawal; keep the taper schedule unchanged (practice).
Part 11: Silexan vs the Alternatives
Cross-trial effect sizes below are from the best-known GAD effect-size analysis (Hidalgo et al. 2007, 21 placebo-controlled trials: pregabalin 0.50, hydroxyzine 0.45, venlafaxine 0.42, benzodiazepines 0.38, SSRIs 0.36, buspirone 0.17) and, for Silexan, from the 2023 pooled analysis (0.35 at 80 mg). They come from different trials, populations and eras, so read the differences as a rough ordering and not as proof that one drug is stronger. The Silexan figure includes subthreshold patients.
| Comparison | Silexan Advantage | Alternative Advantage |
|---|---|---|
| vs SSRIs / SNRIs (GAD, subthreshold anxiety) | No sexual dysfunction, weight gain or withdrawal reported in the trials (they were not designed to measure these closely); no CYP effects in the one cocktail study; some patients will take a supplement who will not take an antidepressant. In the paroxetine trial Silexan beat placebo and paroxetine did not in the main analysis, but the two were not formally compared | FDA-approved; decades of data in GAD, panic, social anxiety, PTSD, OCD and MDD; similar pooled effect sizes in cross-trial comparison; in mild to moderate MDD, sertraline 50 mg and Silexan 80 mg had similar results in the only trial with both, which was not designed to compare them |
| vs Buspirone | Higher point estimate in cross-trial comparison (0.35 vs 0.17); once daily | FDA-approved for anxiety, generic and covered by insurance; no head-to-head trial |
| vs Pregabalin | No abuse signal or withdrawal in the studies done; no sedation in the driving study; not a controlled substance | Faster onset is reported (see the pregabalin guide); higher point estimate in cross-trial comparison (0.50, from GAD trials only); no head-to-head trial |
| vs Hydroxyzine | No sedation in the driving study; meant for daily use over weeks (practice) | Usable as needed; higher point estimate in cross-trial comparison (0.45); generic and cheap; no head-to-head trial |
| vs Benzodiazepines | No abuse signal, withdrawal or driving impairment in the studies done (better than lorazepam 1 mg in the simulator) | Relief within hours; the one comparison, against lorazepam 0.5 mg/day, showed similar 6-week HAMA change but was small, placebo-free and used a low lorazepam dose |
| vs Other lavender products (aromatherapy, other capsules) | Trial data exist for this extract and dose | Inhaled lavender is inexpensive and showed some short-term effect, but most of the trials were at high risk of bias |
No Silexan trial selected patients who had failed an SSRI or SNRI, and the earlier trials required other psychoactive drugs to be stopped, so using it in refractory GAD or as an add-on is extrapolation.
Silexan's niche is the adult with mild to moderate GAD or subthreshold anxiety who will not take, or did not tolerate, a prescription anxiolytic. It wins on side effects, dependence and interactions. It loses on breadth of evidence, independence of that evidence, speed and insurance coverage, and it is untested in panic, social anxiety, PTSD, OCD and moderate to severe depression.
The Bedside Cheat Sheet
What it is
- One standardized lavender oil (WS 1265): linalool 36.8%, linalyl acetate 34.2%
- US: dietary supplement (CalmAid), not FDA-approved. Germany: licensed, non-prescription medicine (Lasea) for restlessness with anxious mood
- RCTs in GAD, subthreshold anxiety, mixed anxiety and depression, mild to moderate MDD; all sponsor-run
- Pooled effect at 80 mg: SMD 0.35, numerically near SSRIs' in cross-trial comparison
Starting and dosing
- 80 mg softgel, once daily, swallowed whole with water, upright
- No FDA dose. Europe: 80 mg/day. Trials: 80 or 160 mg/day
- Full GAD: 160 mg once daily after 1 to 2 weeks (practice)
- Judge at 6 to 8 weeks; separation took 2 to 6 weeks
- Buy CalmAid (or a box naming Silexan or WS 1265); over the counter, out of pocket
Side effects that matter
- Lavender belching (7% to 20%): take with food and water
- Mild nausea; rare allergic reactions (stop it)
- Not sedating; no driving impairment, abuse signal or withdrawal in the studies done
- No weight or sexual side effects reported in the trials (not closely measured)
Don't forget
- Lasea contraindications: liver dysfunction, allergy, under 18
- Not in pregnancy or breastfeeding; not in dialysis
- No CYP or pill interaction at 160 mg; label advises against alcohol and sedatives
- Not for severe anxiety, panic, PTSD, OCD or moderate to severe depression on its own
Have the patient write down a 0 to 10 anxiety score before the first capsule. With an effect this size and a product bought off the shelf, that number is how both of you will know, eight weeks later, whether it earned its place.