Saffron
Saffron is derived from the dried stigmas of Crocus sativus, a flowering plant cultivated primarily in Iran, Spain, and Kashmir. It has been used in culinary and traditional medical contexts for millennia, but its emergence as a subject of rigorous clinical research in mood disorders is more recent, accelerating from the early 2000s onwards. The bioactive constituents most relevant to its proposed psychiatric effects are crocins — water-soluble carotenoid glycosides that give saffron its characteristic colour — and safranal, a volatile compound responsible for much of its distinctive aroma. A third constituent, picrocrocin, contributes to flavour. Clinical trials have used standardised extracts of the stigma or, in some cases, the petal of the plant, typically at 30 mg per day in divided doses, rather than culinary quantities.
What it is
Crocin and its precursor crocetin are the primary focus of pharmacological investigation. Several mechanisms have been proposed to explain saffron's effects on mood, including inhibition of serotonin reuptake — broadly analogous in principle to the mechanism of SSRIs — as well as modulation of dopamine and noradrenaline systems, antioxidant and anti-inflammatory activity, and effects on BDNF expression. These are proposed mechanisms based on preclinical and in vitro evidence; their relative contribution in humans at supplemental doses has not been definitively established, and the mechanistic picture should be treated as hypothetical rather than confirmed.
The form and standardisation of saffron extract matters considerably for interpreting the clinical evidence. Most trials have used proprietary standardised preparations at specified concentrations of bioactive compounds — such as affron (standardised to 3.5% total bioactive compounds including safranal and crocin isomers) or similar products. Culinary saffron, used in cooking at a few strands per dish, delivers amounts far below the doses studied in trials. Claims about mood effects from dietary saffron intake cannot be extrapolated from the clinical evidence base, which is entirely based on concentrated supplemental extracts.
What the evidence shows
Saffron has one of the more developed human clinical evidence bases of any herbal supplement investigated for mood disorders, ahead of comparators such as ashwagandha or lion's mane in terms of trial number, meta-analytic synthesis, and head-to-head antidepressant comparisons. However, the evidence has important limitations that prevent a stronger rating.
For depression, the clearest signal comes from a meta-analysis by Shafiee et al. (2024, Nutrition Reviews) that pooled eight RCTs comparing saffron directly with SSRIs in adults with depression or anxiety. The pooled result showed no significant difference between saffron and SSRIs on depression scores (SMD 0.10, 95% CI −0.09 to 0.29), a pattern of non-inferiority rather than superiority. A parallel analysis of four studies reporting anxiety outcomes produced similarly non-significant differences between groups (SMD 0.04, 95% CI −0.22 to 0.29). Saffron was associated with fewer adverse events than SSRIs in this analysis. A broader meta-analysis by Han et al. (2024, Phytotherapy Research) synthesised 46 RCTs across multiple neuropsychiatric outcomes and found saffron superior to placebo on depression scores with an overall pooled effect size of −4.26 points on rating scales (95% CI −5.76 to −2.77) and on anxiety scores (−3.75, 95% CI −5.83 to −1.67). These are pooled estimates across heterogeneous trials and should be read with caution given the methodological concerns described below.
The most recent and largest independent placebo-controlled trial is Lopresti et al. (2025, Journal of Nutrition, n=202, ages 18–70), which examined 12 weeks of 28 mg daily affron extract in adults with subclinical depressive symptoms — a less severe population than earlier trials. Saffron was associated with a statistically significant reduction in DASS-21 depression score compared with placebo (beta −2.92 points, 95% CI −5.13 to −0.71, Cohen's d=0.39), with 72.3 percent of saffron participants achieving a clinically significant reduction versus 54.3 percent in the placebo group. Secondary outcomes including anxiety, stress, and sleep did not differ significantly between groups, with the exception of an exploratory sleep finding in participants with greater baseline sleep disturbance. The effect size of d=0.39 is modest by conventional standards and may fall below thresholds considered clinically meaningful by some scales; the subclinical population also limits direct applicability to clinical depression.
Several limitations run through the broader evidence base. Placebo-controlled evidence, while generally positive, is largely derived from small early trials and may overestimate the true effect size due to small-study bias and likely publication bias. The concentration of trials within a single research network and geography raises the possibility of systematic bias and limits confidence in generalisability to other populations. Many early trials were small (n=20 to 40), used active antidepressant comparators rather than placebo, and were of short duration (6 to 8 weeks). Most trials are short-term, and durability of effect beyond 8 to 12 weeks is not established. Active-comparator designs showing non-inferiority to SSRIs do not independently establish efficacy — a non-inferior result against an active treatment only confirms efficacy if the active treatment itself is assumed to be effective in the studied population. Furthermore, SSRIs show only modest benefit over placebo in mild depression and may not significantly outperform placebo in some mild cases, which means non-inferiority to SSRIs does not imply a large absolute treatment effect in that population.
For PMS and PMDD, a 2025 meta-analysis pooled seven RCTs (n=612) and found saffron superior to placebo on PMS symptom scores (SMD −0.64, 95% CI −0.84 to −0.44). One small RCT (n=120) demonstrated non-inferiority of saffron to fluoxetine in PMDD. The PMS and PMDD evidence is directionally consistent but shares the same limitations as the depression literature: it is primarily sourced from the same Iranian research groups, trials are individually small, and independent replication outside that network is limited. It should be interpreted as less robust than the depression evidence due to smaller and less independently replicated datasets.
For sleep, meta-analyses exist but the evidence is less consistent across outcome measures and populations than for mood.
Five questions
Does low status cause harm? Saffron is a culinary spice rather than an essential nutrient, so there is no deficiency state in the conventional sense. There is no evidence that low dietary saffron intake causes mood disorders or any health outcome. The clinical evidence is about supplemental extracts, not nutritional adequacy.
Does supplementation prevent disease? No human evidence exists to support a disease-prevention claim for saffron. Trials have investigated symptomatic improvement in existing mild-to-moderate depression, not prevention of depressive episodes in at-risk populations. The question of whether saffron might reduce incidence, recurrence, or progression of mood disorders has not been studied.
Does it affect biomarkers? Some trials have reported changes in inflammatory markers and cortisol, but these have not been the focus of systematic investigation and are not consistently reported across the literature. Saffron is not associated with a validated biomarker profile. Claims about serotonin modulation rest on mechanistic inference rather than direct human measurement.
Does it help clinical populations? Most trials have enrolled adults with mild-to-moderate depression diagnosed by validated scales. In this population, saffron has shown consistent superiority to placebo and non-inferiority to SSRIs across multiple trials and meta-analyses. However, the placebo-controlled trials are predominantly small and geographically concentrated, and non-inferiority to SSRIs in mild depression is a weaker standard than it appears, given that SSRIs themselves show only modest benefit over placebo at that severity level. This constitutes the primary clinical population with evidence. Evidence for severe depression, treatment-resistant depression, bipolar depression, and major depressive disorder without mild-to-moderate specification is either absent or specifically excluded in most trials. Evidence in PMDD represents a meaningful secondary clinical population where trials support efficacy.
Does it benefit healthy individuals? The evidence in healthy individuals without diagnosed mood disorders is limited. Some trials have examined adults with subclinical symptoms rather than clinical diagnoses, and the 2025 Lopresti trial used a subclinical population specifically, finding a modest benefit. Whether saffron meaningfully improves mood or wellbeing in individuals with no baseline mood symptoms is not well supported. For sleep in otherwise healthy individuals, findings are inconsistent.
Individual variation
Baseline depression severity appears relevant. The signal is clearest in mild-to-moderate clinical depression; the largest trial in a subclinical population found smaller effects. Whether saffron produces meaningful benefit in people with no mood disturbance at all is not established.
Sex-specific evidence is meaningful here. Women of reproductive age have a distinct evidence base for PMS and PMDD, where the serotonergic mechanism most plausibly overlaps with luteal phase serotonin dysregulation. The depression trials include mixed-sex populations; most have not reported sex-stratified outcomes. Evidence during perimenopause and menopause specifically — where serotonin function and mood are often affected by declining oestrogen — has not been systematically investigated in saffron trials, representing a gap given the biological plausibility.
The form and standardisation of the extract used is likely a significant source of variability in response. Trials that have used standardised preparations with defined crocin and safranal content are more likely to produce reproducible effects than uncharacterised preparations. Generic saffron supplements with no standardisation specification cannot be assumed equivalent to the extracts used in trials.
Concurrent use of serotonergic medications is relevant both to safety and to the likely magnitude of any additive effect. This is discussed further in the safety section.
Testing and status assessment
There are no relevant clinical tests for saffron status or response prediction. Mood outcomes in trials are measured using validated scales such as the Hamilton Depression Rating Scale, Beck Depression Inventory, or DASS-21; these are not routine monitoring tools in the supplementation context. There is no biomarker that usefully predicts who will respond to saffron supplementation.
Safety
At supplement doses studied in clinical trials — typically 30 mg per day of standardised stigma extract — saffron has been well tolerated with no serious adverse events reported across the published evidence base. Adverse event rates in saffron groups have generally been comparable to or lower than those in SSRI comparison groups, and similar to placebo in placebo-controlled trials. Common mild adverse events include nausea, headache, and dry mouth at therapeutic doses. The risks described below vary considerably in their evidentiary basis: pregnancy risk is well supported by mechanistic, animal, and historical human data; anticoagulant interaction is supported by one case report and in vitro evidence; serotonin syndrome and bipolar risk are pharmacologically plausible but not confirmed in clinical data.
Pregnancy is a clear contraindication. At doses above those found in food, saffron has uterotonic properties and has been associated with increased miscarriage risk in animal models and case-level human data. Doses of 5 g or more are explicitly associated with uterine stimulation and abortifacient effects. Even at lower supplemental doses, saffron should be avoided during pregnancy. Breastfeeding safety data are absent.
The antiplatelet and anticoagulant interaction warrants attention. Saffron has been shown to inhibit platelet aggregation in vitro, and at least one case report documents significant bleeding complications in a patient combining saffron with rivaroxaban. Interaction with warfarin and other anticoagulants is pharmacologically plausible and should be considered a potentially clinically relevant risk, though it has not been quantified in controlled studies. People taking anticoagulants or antiplatelet therapy should discuss saffron use with a clinician before proceeding.
The serotonin syndrome risk with SSRIs and SNRIs is theoretical. Saffron's proposed serotonin reuptake inhibiting activity creates a pharmacological basis for concern, but no confirmed clinical cases of serotonin syndrome attributable to the saffron-plus-SSRI combination have been documented in the trial literature. The overlap remains a plausible interaction rather than an established one. Nonetheless, saffron is sometimes considered as a natural adjunct to prescribed antidepressants — this combination carries an uncharacterised interaction risk and is not supported by trial evidence of safety in combination.
Bipolar disorder is a specific contraindication. Saffron's mood-altering properties may trigger hypomanic or manic episodes in individuals with bipolar spectrum diagnoses, and its use in this population is not supported by clinical trial evidence.
High-dose toxicity is documented. Doses of 5 g or more have been associated with serious adverse effects including vomiting, bleeding, and in historical reports used as abortifacients, potentially fatal outcomes. These doses are far above any supplement formulation but the dose-safety relationship is relevant context.
Saffron should be discontinued at least two weeks before any planned surgical procedure due to its effects on platelet function and possible interaction with anaesthetic agents.
What can reasonably be concluded
Saffron has one of the more consistent human clinical evidence bases among herbal mood supplements, anchored to mild-to-moderate depression as its primary indication. Multiple meta-analyses support a modest antidepressant effect over placebo, and non-inferiority to SSRIs has been demonstrated in direct comparison trials. The PMS evidence adds a separate, directionally consistent area of support. The safety profile at studied doses is acceptable in otherwise healthy adults, with important exceptions.
The significant qualifications are: most placebo-controlled trials are small and concentrated within a single research network and geography, which raises the possibility of systematic bias and limits generalisability; effect sizes in older meta-analyses are likely inflated by small-study bias; the largest independent trial found a modest effect size of d=0.39 in a subclinical population, which may fall below thresholds of clinical meaningfulness on some scales; there is limited evidence that these changes translate into meaningful improvements in functional outcomes or long-term prognosis; non-inferiority to SSRIs does not imply a large treatment effect given SSRIs themselves show only modest benefit over placebo in mild depression; and the long-term evidence base is absent. A Moderate rating reflects an evidence base that is more developed than most supplements in this category, but narrowly scoped and still limited relative to what would be required for confident clinical recommendations.
Where evidence is limited or outcomes are uncertain, conclusions should be treated as provisional and subject to revision as the evidence base develops.