Depression is common—the prevalence of major depressive disorder (MDD) varies between 2% and 21%, depending on the country and measure used (Gutiérrez-Rojas et al., 2020), and prevalence is increasing over time (Moreno-Agostino et al., 2021). MDD is defined according to the DSM-V as the presence of five or more out of nine symptoms experienced frequently within the same two-week period, including depressed mood and loss of pleasure, in addition to other symptoms such as changes in appetite, insomnia, and fatigue. Symptoms must also cause distress or impairment, with the severity of depression determined by the level of impairment, and the condition cannot be better explained by another cause (Uher et al., 2014).

In the United Kingdom (UK), 11.3% report mild depressive symptoms, 4.2% report moderate depressive symptoms, and 3.3% report severe depressive symptoms (Arias De La Torre et al., 2021). MDD is associated with multiple interacting pathogenic factors, including genetics, stressful life events, and chronic diseases (Cui et al., 2024). Underlying pathophysiological theories for depression include 1) deficiencies in monoamine neurotransmitters (e.g., serotonin and dopamine), 2) an increase in hypothalamic–pituitary–adrenal (HPA) axis activity, increasing production of glucocorticoids, which provide negative feedback to the limbic system, hypothalamus, and pituitary, and 3) low-grade inflammation that is self-sustaining, with the presence of reactive oxygen species further triggering inflammatory cytokines, which compromise neuroendocrine activities (Cui et al., 2024). There may also be impaired neurogenesis due to reduced brain-derived neurotrophic factor (BDNF) (Dobrek and Głowacka, 2023) or the effects from the gut microbiota on the function of the HPA axis (Dobrek and Głowacka, 2023).

Depression has a strong impact on individuals, reducing the quality of life and increasing the risk of physical health conditions such as cardiovascular disease and diabetes (Graham et al., 2020; Harshfield et al., 2020). There is also a widespread societal impact as depression increases the risk of sickness absence from employment (Amiri and Behnezhad, 2021). Globally, 16% of disability-adjusted life years can be attributed to depression, reflecting an economic cost of USD$ 4.7 trillion (Arias et al., 2022). There is, therefore, a strong need to find effective methods to prevent, manage, and reduce depressive symptoms in adults.

In the UK, the National Institute for Health and Care Excellence (NICE) guidelines recommend discussions of psychological treatments (e.g., group and individual psychological therapies and guided self-help), exercise, and pharmacological treatment options (e.g., antidepressants), with a decision made jointly with the patient as to which is the most appropriate, considering the least intrusive and least resource-intensive options first (National Institute for Health and Care Excellence, 2022). However, it is estimated that in Great Britain, only 13.4% of individuals reporting depressive symptoms receive some form of treatment (Smits and Huijts, 2015), with particular challenges in accessing psychological therapies due to long waiting times (Baker and Kirk-Wade, 2024). Although antidepressants are more easily accessible, they require help-seeking with a medical practitioner in order to receive them, and views and experiences of their perceived effectiveness, desirability, and side effects are mixed (Crowe et al., 2023).

Consequently, it is unsurprising that prior to, or whilst receiving treatment, people often find ways to self-manage depressive symptoms, most commonly using herbal medical products (HMPs) and vitamins and minerals rather than practitioner-directed modalities (Solomon and Adams, 2015). Homoeopathic preparations are also a possible approach, with one of the most common reasons for using over-the-counter (OTC) homoeopathic preparations in the UK being psychological problems (Reid, 2002). Although there are no specific OTC medicines licensed for depressive symptoms, it is possible that OTC medicines are also used more frequently to mitigate depressive symptoms or symptoms linked to depression (e.g., insomnia).

It is suggested that natural products, such as HMPs or nutrients, may help manage depression through various mechanisms, including inhibiting inflammation, ameliorating oxidative stress, changing the microbiota–gut–brain axis, suppressing hyperactivity in the hypothalamic–pituitary–adrenal axis, and regulating neurotransmitters (Wu et al., 2022). These products are generally derived from different medical traditions and are now an important part of self-care in many countries, either regulated as medicines or supplements/botanicals. Qualitative studies suggest that products such as St John’s Wort can be viewed as more desirable due to perceptions of safety, naturalness, and greater ability to control usage (Pirotta et al., 2014), and they may be used as alternatives to, or in addition to, prescribed antidepressants. There is little mention in the NICE guidelines regarding OTC products for depression, apart from a recommendation not to advise St John’s Wort due to uncertainty about doses and the potential for interactions (National Institute for Health and Care Excellence, 2022).

There has been an abundance of evidence evaluating OTC products for depression over the last few decades, particularly products such as St John’s Wort and omega-3 supplements. These often have relatively conclusive separate systematic reviews and meta-analyses [e.g., St John’s Wort (Ng et al., 2017), saffron (Tóth et al., 2019), and omega-3s (Liao et al., 2019)]. Although this can provide a rigorous assessment of the strength and level of evidence available for each product and a single estimate of effect, it tends to focus the evidence base on a few key products and neglects the wider scope of other promising products that may offer future avenues for prevention and self-care of depressive symptoms.

Where reviews have previously attempted to summarise the wider evidence base beyond single products, they have often relied on narrative summaries (Sarris, 2018) or included only a limited number of studies (Kamat et al., 2023). It is, therefore, valuable to scope the evidence base as a whole, including ongoing trials, to understand where the literature is concentrated and where gaps exist. This allows future research directions to be clearly outlined, broadening the potential field for clinical studies in products for depressive symptoms beyond those with an already high level of evaluation. We, therefore, aimed to

1. Summarise the size and nature of the evidence base that assesses the effectiveness of OTC products for depression in adults aged 18–60.

Following Joanna Briggs Institute (JBI) guidance (Peters et al., 2020) and PRISMA guidance for scoping reviews (Tricco et al., 2018), we carried out a scoping review. The review formed part of a larger project summarising the available trial evidence for OTC products for depression, anxiety, and insomnia. It was prospectively registered on the Open Science Framework (https://osf.io/rkm57/). Due to the large volume of results, we synthesised the findings separately for each condition and grouped them according to whether the samples were adults (aged 18–60) or older adults to explore whether an age bias existed in these trials. Two public contributors were involved in developing the review questions, designing the review, and selecting which products to include.

We searched five databases (MEDLINE, Embase, PsycINFO, AMED, and CENTRAL) from inception to December 2022. The CENTRAL searches included trial registry entries. Search terms were grouped into OTC product terms, mental health terms, and randomised controlled trial (RCT) filters (where applicable) and combined using Boolean operators (see Supplementary Material 1). Given the multitude of possible products, product categories rather than individual names were used in searches. The research team piloted searches in MEDLINE and Embase and refined these as needed.

Search results were imported into Rayyan (Ouzzani et al., 2016) after deduplication. Due to the large volume of results, dual title and abstract screening was undertaken for 10% of titles and abstracts by pairs of reviewers (RF, SM, SU, VT, and AS). Reviewers were provided with explicit instructions to maximise consistency, and each pair conducted three batches of screening; discussions and decisions were agreed upon by all five reviewers, and inclusion criteria refined as needed. Only when agreement was 85%–90% between all pairs of reviewers (after the third batch) were the remaining studies screened by a single reviewer. Studies deemed unclear were moved to the full-text stage. The same process was carried out for full-text screening, with decisions documented in MS Excel, and unclear studies at this stage were screened by a second reviewer. As part of screening, studies were labelled according to whether they focussed on anxiety, depression, insomnia, or a combination and whether their sample included people aged 18–60 or 60+. In this article, studies relating to depression in adults aged 18–60 are synthesised.

We included randomised controlled trials (including parallel, factorial, and crossover trials) evaluating products that met the following criteria:

• were likely to be available OTC in a number of countries globally.

We included studies where the sample mean ages fell between 18 and 60 years and where participants had depressive symptoms established at baseline—either by meeting a threshold on an established depression questionnaire or with a diagnosis of minor or major depressive disorder. Studies that included people with depression and a comorbid condition were eligible, but those focussing on other mental health conditions (e.g., bipolar disorder, dementia, and substance abuse) were excluded. Studies needed to include depressive symptoms or depression remission as an outcome measure. We included trials from any time period and in any language, using Google Lens or a multilingual colleague where available for non-English texts. Subgroup analyses and grey literature presenting result data (e.g., theses) were excluded.

For all identified trial registry entries, published protocols, and conference abstracts, during the period of April–June 2024 we endeavoured to locate the main trial publication through trial number searches, title searches, and “cited by” functions (published protocols only) and by reviewing named author publication lists. Where full texts could not be located, we summarised ongoing trials and published protocols (see Supplementary Material 2) but not conference abstracts. Full texts located through these methods were screened by one author (RF) and were included and extracted if eligible. Reference list screening of relevant reviews was undertaken by SM for 10% (n = 200) reviews; as few new studies were detected and resources were limited, no further reference lists were screened after this point.

Data were extracted using a data extraction form designed according to the JBI data extraction template, including study details such as country, setting, sample size, inclusion criteria, participant characteristics, product characteristics, comparators, effectiveness outcomes for depression, and safety outcomes. Data were extracted by a reviewer (AZ), who also conducted a second eligibility check of included studies. Codings/groupings were applied by RF, who checked any uncertainties by consulting the original papers if needed and confirmed the exclusion of any ineligible studies. RF also extracted data from studies identified through trial registry and conference abstract follow-up. For plants, both common and Latin names were extracted where reported; however, where these were not reported, we avoided imposing potentially incorrect names and instead used the name(s) listed in the study.

To synthesise the data, we grouped products by overall and specific product type. We summarised the products evaluated, comparators used, comorbid conditions, and whether the products were used as adjuncts or evaluated alone. Due to the large volume of studies, we summarised effectiveness and safety data using vote counting. We classified effectiveness findings according to whether they showed significant effects at ≥ 1 timepoint on any depression trial outcome compared to each type of control treatment, and we classified safety findings by whether there were differences between groups in adverse events (see Table 1 for categorisations applied by RF). Only comparisons relevant to the review were included (e.g., prescribed antidepressant vs. placebo comparisons in a three-arm trial are not presented). We did not carry out a quality assessment as this was not the aim of scoping reviews, which provide a descriptive overview of the evidence base (Peters et al., 2020). We aimed to highlight areas with larger volumes of evidence and identify gaps within the evidence base needing further research.

Figure 1 shows the flow of studies through this review. Out of 15,339 records screened, 13,972 were excluded on the basis of title and abstract, and 1,367 full texts were screened. One hundred and ninety papers relating to depression in adults aged 18–60 were included, with a further 16 identified from following up registered protocols and conference abstracts, providing a total of 209 included trials. Of these, 196 evaluated products for depression alone, 10 for depression and anxiety, 2 for depression and insomnia, and 1 for all three conditions. Tables 2–5 summarise all included studies with references.

This review found a total of 209 trials assessing OTC products in people with depressive symptoms or a depression diagnosis. Most products evaluated were dietary supplements or herbal medical products. Only one single chemical medication was found (aspirin), and no homeopathic products were found. We identified notable gaps in the evidence:

• Although a wide range of products have been evaluated, for many of these, we found only a small number of trials, and further replication of results is needed. Ongoing studies currently follow this trend and either evaluate products already with a substantive evidence base or completely new products.

Questionnaires recording the use of products for depression have found that in the US, Serbia, Mexico, and Iran, the most commonly reported HMPs were borage, chamomile, lavender, St John’s Wort, ginseng, ginkgo, orange blossom, valerian, tilia, lemon balm, and peppermint (Roy-Byrne et al., 2005; Ashraf et al., 2021; Alonso-Castro et al., 2021; Stojanović et al., 2017; Niv et al., 2010). There remain notable gaps in the evidence base regarding ginseng, ginkgo, tilia, orange blossom, and peppermint, and there are relatively few studies for products other than St John’s Wort. Borage was commonly used in only one survey in Iran; although the species name was not specified, it is likely to refer to Iranian borage (Echium amoenum Fisch. & C.A.Mey), which was evaluated in two studies included in this review, rather than Borago officinalis L. Echium showed mixed results for its effects on depression and warrants further research. Products containing mixes of herbs and/or dietary supplements were rarely evaluated. This may arise from the fact that standardisation and ensuring supply chain quality are more challenging when multiple extracts are present (Ekor, 2014). Future research should, therefore, focus on individual extracts and evaluating potential additive effects of effective herbal products using methods such as factorial trials, which were rarely used in this review. No homoeopathic products were found that were evaluated for depressive symptoms in this review; this suggests that this is not a good avenue for further research at present.

Products also rarely focused on teas (a small number of lavender and chamomile products), as opposed to capsules, even though in some studies, teas were the most common way of consuming HMPs for depression (Stojanović et al., 2017). Lemon balm (75.9%) and chamomile (60.3%) were the two most commonly used teas in a Portuguese consumer panel (Rocha et al., 2020). Herbal tea consumers closely associated herbal teas with emotions (e.g., calm and relaxation), sensory experiences (e.g., taste and appearance), and perceived effects on health and wellbeing (Rocha et al., 2020). Herbal teas are generally under-researched; a scoping review of the general health benefits of herbal teas found only 16 clinical studies and five observational studies (Poswal et al., 2019). Given their wide availability, low cost, and ability to become part of an everyday lifestyle, herbal teas warrant further clinical and observational research for their role in managing depression.

With regards to dietary supplements, there was a particular gap between the most commonly used products and the evidence base available. One study from New Zealand found that the most common dietary supplement among people with depression was a multivitamin, followed by vitamin B complex (Silvers et al., 2006), while a US study found that people with psychiatric symptoms were more likely to take melatonin compared to those without psychiatric symptoms (Niv et al., 2010). Only a small number of trials evaluated these products for depression, with few positive effects. Rates of omega-3 consumption were not significantly different in a large US survey in those with and without psychiatric symptoms (Niv et al., 2010) despite its substantive evidence base. We found few trials for multivitamins, and these had variable results. Multivitamins in particular present a challenge to evaluate due to the potential for variation between the number, type, and dose of vitamins and minerals included. Melatonin (more commonly used for insomnia) showed mixed effects on depressive symptoms and may warrant further research as a potential treatment for depression.

The use of OTC single-chemical products for depression has been evaluated in relatively few studies. This is not particularly surprising as most single-chemical medications targeting depression are prescription-only. One Swedish study found that adolescents with depressive symptoms were more likely to use OTC analgesics than their non-depressed peers (Hena et al., 2019); however, few studies have explored this relationship in adults. A single trial of aspirin (hypothesised to exert effects through anti-inflammatory mechanisms) found no effects on depressive symptoms. Few randomised trials have assessed aspirin for depression; however, a review of prospective cohort studies found that aspirin use was associated with an increased risk of depression (Kim et al., 2020). Although this may arise from confounding factors such as underlying pain or cardiovascular disease, the existing evidence suggests that aspirin is unlikely to be a promising candidate for future depression research.

Whilst a substantive evidence base was found for a small number of products, existing meta-analyses provide relatively recent conclusions regarding the level and quality of evidence available for these products. A 2019 meta-analysis of nine saffron studies found it to be effective against placebo and not significantly different from antidepressants, with studies mostly having a low or unclear risk of bias (Tóth et al., 2019). A 2017 meta-analysis of 27 studies on St John’s Wort showed similar efficacy and lower dropout rates compared to antidepressants, supported by a moderate-quality evidence base (Ng et al., 2017). Omega-3s showed significant effects in reducing depressive symptoms compared to placebo across 26 studies, with greater effects in EPA-enriched or EPA-only products (Liao et al., 2019). The evidence base for probiotics (19 trials) was generally of high quality and showed that probiotics reduce depressive versus placebo, with greater effects observed in people with major depressive disorder (Goh et al., 2019). Vitamin D (nine trials) showed limited evidence of efficacy to date (Gowda et al., 2015). There are some differences between these meta-analyses and the findings of our review. As included studies had only a small sample size (median 62), it is possible that positive but non-significant effects became significant when pooled, in contrast to our use of vote counting, which has well-established limitations. It may also reflect differences in search strategies and inclusion criteria. Our searches identified more trials than some of these reviews (Tóth et al., 2019; Liao et al., 2019; Gowda et al., 2015) despite similar inclusion criteria, suggesting that future meta-analyses should adopt a more thorough search approach. Other reviews found a large number of studies but employed wider inclusion criteria, including healthy samples (Goh et al., 2019). Generally, products with a more substantive evidence base had more trials showing some differences in mild adverse events; however, this may represent more rigorous detection methods or criteria rather than actual safety differences to products with emerging evidence. Only 10 studies in our review evaluated different doses of the same product; dose-response relationships require further study for all products.

Ongoing trials also focussed on omega-3s, probiotics, and vitamin D. Rather than conducting further trials of these products, future attention needs to be directed towards promising products with fewer trials. Products with fewer studies but with more positive than null results include folic acid, lavender, rhodiola, tryptophan, lemon balm, bitter orange, N. menthoides Boiss. & Buhse, chamomile, and zinc. Lavender, chamomile, and lemon balm are traditionally indicated and used in Western herbal medicine for depression, particularly with comorbid anxiety or stomach complaints (Bell Hunter, 2007; Bone, 2003). Bitter orange (Citrus x aurantium L.) and N. menthoides Boiss. & Buhse are traditionally used for psychological disorders, including depression, particularly in Iran (Eslami-Farouji and Jalili, 2024; Memariani et al., 2019). Rhodiola rosea L. is traditionally used across Europe and Asia for stress, fatigue, and depression (Tao et al., 2019). Most of these products show some evidence of influencing the pathophysiological mechanisms of depression. Lavender oil, rhodiola, lemon balm, N. menthoides Boisse. & Buhse, chamomile, folic acid, tryptophan, and zinc show anti-inflammatory activity in in vitro and in vivo studies (Dobrek and Głowacka, 2023; Zam et al., 2022; Dai et al., 2022; Liwinski and Lang, 2023; Wang et al., 2018; Nayak et al., 2019). Lavender, rhodiola, folic acid, and lemon balm modulate the activity of various neurotransmitters, whilst tryptophan is a precursor molecule for serotonin (Dobrek and Głowacka, 2023; Liwinski and Lang, 2023; Nayak et al., 2019). Bitter orange, N. menthoides Boiss. & Buhse, lemon balm, chamomile, tryptophan, and zinc show antioxidant effects (Dobrek and Głowacka, 2023; Dai et al., 2022; Wang et al., 2018; Nayak et al., 2019; Maksoud et al., 2021). There is also evidence that zinc, lemon balm, and rhodiola moderate HPA axis activity, whilst lavender, chamomile, and zinc showed neuroprotective activity by increasing or restoring BDNF (Dobrek and Głowacka, 2023; Dai et al., 2022; Wang et al., 2018). These potential mechanisms of effect, combined with traditional use and promising clinical evidence, suggest that further, larger trials should concentrate on these products for the treatment of depression. Despite this, few of these HMPs had registered ongoing trials (n = 1 lavender).

With 89 trials evaluating products as adjuncts alongside conventional treatment (mostly antidepressants), this represents a more realistic approach to evaluation (Solomon and Adams, 2015) and is welcomed. However, very few trials assessed the additional value of these therapies to psychological therapies. As these are considered potential first-line treatments in the UK (National Institute for Health and Care Excellence, 2022), this needs to be evaluated further. The substantive nature of the evidence base for some products also suggests that NICE guidelines should formally review these in future updates and provide clarity as to whether to recommend the use of these products for depression. Likewise, 54 trials included people with comorbidities. As depression is more common in people with long-term conditions (Gutiérrez-Rojas et al., 2020), this is heartening. Thirteen studies were also carried out in pregnant or postpartum women. Preliminary evidence indicated most products were safe, whether administered alone or alongside antidepressants. This will provide confidence to users of OTC products for depression, particularly those already taking antidepressants. However, 64 studies did not report adverse events, and reporting did not improve over time. Authors are strongly recommended to clearly report adverse events (or lack thereof) within OTC product trials, particularly as these products are most frequently used without clinical supervision. Further comprehensive analysis of adverse event data for each product is also recommended. Only one economic evaluation was found; further work on whether OTC products could provide savings in healthcare resource use for depression is needed.

Strengths of our review include the fact that we located a much larger number of trials than a previous review—focussing on products for all ages and for anxiety and insomnia (n = 76) (Kamat et al., 2023)—and more than some of the individual product meta-analyses discussed above, demonstrating the value of a scoping review methodology and a thorough approach to searching and follow up. We also summarised ongoing trials. We brought together multiple available products rather than focussing on a single product to provide an overall map of the evidence. We focussed on samples where participants had baseline depression symptoms or a diagnosis rather than those that examined depressive symptoms in non-depressed samples.

However, there were some key limitations to our review. Despite attempts to search exhaustively, some trials did not use product class keywords that were used in our searches, and we were unable to search each potential product name. It is, therefore, likely that some relevant studies were missed—for example, a meta-analysis of St John’s Wort (Ng et al., 2017) located more comparisons with SSRIs than those found in our review, and two potentially eligible trials of aspirin were found in another review (Dominiak et al., 2022). Future studies need to report product class, descriptors and names. Searches were also only carried out up to 2022. Follow-up of trial protocols and conference abstracts detected 16 further studies, some of which were published after the search dates. However, trial registry entries were not always up to date, and sometimes, a lack of detail limited further searching. Likewise, multiple publications from the same studies were not always clearly labelled. Only 10% of studies were screened by two authors, which may have introduced bias. In our protocol, we originally planned to have two reviewers screening all titles, abstracts, and full texts and to perform citation tracking and reference list screening of included studies; however, the large volume of studies identified and our limited resources precluded this. Future reviews should use a higher proportion of dual screening or have all studies dual-screened. Although we included non-English literature, we did not search using multi-language terms, and some studies were translated using Google Lens, which may not provide fully accurate details for data extraction. Future research should also consider including traditional Chinese medicine approaches to provide a more global perspective.

As the review was intended to be descriptive, we used simple vote counting to summarise the study’s effectiveness. This could be difficult to classify when different results occurred in multiple measures of depression or at different timepoints, and we did not account for sample size and, therefore, study power. We also did not explore the quality of statistical analyses or the use of intention-to-treat vs. per-protocol analysis, which were often poorly described and inconsistently labelled across studies. Results are, therefore, less precise than a meta-analysis or narrative review with consideration of study quality, and as such, they should be viewed as preliminary and not definitive. Safety data were variably reported and were not always tested for significance between adverse event incidences between groups; therefore, we took a cautious approach, and imbalances of more than a few between groups were classed as differences. As almost no serious adverse events were reported in trials; this is unlikely to affect review safety conclusions substantially. Including such a large number of products and trials necessarily resulted in some loss of nuance regarding dosage and preparation type.

The largest volume of evidence exists for omega-3s, St John’s Wort, saffron, probiotics, and vitamin D, all of which are relatively established products. However, a multitude of other products are promising, including folic acid, lavender, zinc, tryptophan, rhodiola, and lemon balm. Likewise, several products have mixed evidence, requiring further study such as melatonin, magnesium, curcumin, cinnamon, Echium, vitamin C, and vitamin D plus calcium. Among these, chamomile, lavender, lemon balm, and Echium represent commonly used products and should be prioritised for further research into their safety and effectiveness.

Most studies reporting on safety presented no safety issues, with a small number showing differences in mild adverse effects. One study on an omega-3 product reported a higher incidence of serious adverse events in the product group. Safety data were consistent whether the products were used alone or as adjuncts to antidepressants. However, trial authors should ensure that such information is reported completely and clearly in papers. There is a need for further evaluation of HMPs as adjuncts to antidepressants and for exploring their potential benefits when used adjunctively with psychological therapies to reflect likely clinical usage.