Advertisement

Effects of Lavender on Anxiety, Depression, and Physiological Parameters: Systematic Review and Meta-Analysis

Open AccessPublished:November 11, 2021DOI:https://doi.org/10.1016/j.anr.2021.11.001

      Summary

      Purpose

      The recent evidence suggested substantial anxiolytic efficacy of lavender. The aim of this study was to examine the efficacy of lavender for anxiety, depression, and physiological parameters and to elucidate the differential effects of lavender on anxiety and depression by study characteristics.

      Methods

      A systematic review and meta-analysis was performed following the PRISMA guidelines. We searched PubMed, Embase, Cochrane Library, Web of Science, and Cumulative Index of Nursing and Allied Health Literature databases for randomized controlled trials investigating the efficacy of lavender on anxiety, depression, or physiological parameters in humans. We assessed the risk of bias within studies with the revised Cochrane risk of bias tool for randomized trials. We used random effect model to estimate the average effect and computed bias-corrected standardized mean difference as effect size metric, Hedges’ ĝ for all outcomes.

      Results

      Lavender was superior to placebo or no treatment in reducing anxiety (Hedges' ĝ = −0.72, 95% confidence interval [CI] −0.90 to −0.55, p value <.001), depression (Hedges' ĝ = −0.43, 95% CI, −0.59 to −0.27, p value <.001), and systolic blood pressure (Hedges' ĝ = −0.23, 95% CI, −0.41to −0.05, p value = .01). The moderator analysis by meta-regression indicated that route of administration accounted 6.5% (p value = .187) for the heterogeneity in anxiolytic effects, sessions of treatment accounted 13.2% (p value = .055), and participants’ health state accounted 8.9% (p value = .131) for the variance in anxiolytic effects.

      Conclusion

      Lavender aromatherapy showed substantial effect in reducing anxiety and depression, and sessions of administration increased the anxiolytic effects. The effects on physiological parameters showed small with inconsistent significances and randomized controlled trials on the effect of lavender on depression were scarce. Future trials on depression and physiological parameters are recommended, and increasing the sessions of administration is recommended.

      Keywords

      Introduction

      Anxiety disorders are the most prevalent mental disorders around the world and are associated with significant comorbidity and morbidity [
      • Stein D.J.
      • Scott K.M.
      • de Jonge P.
      • Kessler R.C.
      Epidemiology of anxiety disorders: from surveys to nosology and back.
      ]. Anxiety is a characteristic feature of modern times, and the prevalence of anxiety disorders has increased in response to political, societal, economical, and environmental changes [
      • Bandelow B.
      • Michaelis S.
      Epidemiology of anxiety disorders in the 21st century.
      ]. The result of meta-regression adjusted for methodological difference indicated the global prevalence of anxiety disorders as 7.3% [
      • Baxter A.J.
      • Scott K.M.
      • Vos T.
      • Whiteford H.A.
      Global prevalence of anxiety disorders: a systematic review and meta-regression.
      ].
      The etiology of anxiety disorders includes an interaction of psychosocial factors, for example, childhood adversity, stress, or trauma, and a genetic vulnerability, which manifests in neurobiological and neuropsychological dysfunctions [
      • Bandelow B.
      • Michaelis S.
      • Wedekind D.
      Treatment of anxiety disorders.
      ]. Anxiety disorders are often comorbid with other anxiety disorders, major depression, or substance abuse [
      • Kessler R.C.
      • Chiu W.T.
      • Demler O.
      • Merikangas K.R.
      • Walters E.E.
      Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the national comorbidity survey replication.
      ]. Current anxiolytic treatment options have limited efficacy, such as delayed onset (e.g., selective serotonin reuptake inhibitors, serotonin–norepinephrine reuptake inhibitors, and buspirone) as well as the potential for habituation, tolerance, and abuse (e.g., benzodiazepines and pregabalin) [
      • Malcolm B.J.
      • Tallian K.
      Essential oil of lavender in anxiety disorders: ready for prime time?.
      ]. In addition, anxiolytic agents may cause side effects, such as sedation, impaired concentration, amnesia, depression, delirium, dependency, and, not least, withdrawal syndrome [
      • Kasper S.
      • Müller W.E.
      • Volz H.P.
      • Möller H.J.
      • Koch E.
      • Dienel A.
      Silexan in anxiety disorders: clinical data and pharmacological background.
      ]. Therefore, there is a demand for efficacious, safe, and acceptable anxiolytics that are also applicable in subthreshold conditions [
      • Rickels K.
      • Garcia-Espana F.
      • Mandos L.A.
      • Case G.W.
      Physician withdrawal checklist (PWC-20).
      ].
      Lavender oil administered by different routes has been recognized for centuries for promoting “well-being” and for reduction of distress [
      • Müller W.E.
      • Sillani G.
      • Schuwald A.
      • Friedland K.
      Pharmacological basis of the anxiolytic and antidepressant properties of Silexan®, an essential oil from the flowers of lavender.
      ]. Lavender, a plant from the Lamiaceae family, comes in many species with different chemical characteristics. The lavandula genus has approximately 30 species grown around the world that share similar major chemical constituents and properties [
      • Cavanagh H.M.
      • Wilkinson J.M.
      Biological activities of lavender essential oil.
      ]. Lavender oil is the essential oil extracted from flowers and stalks of the lavender plant by steam distillation. It is a colorless or pale-yellow liquid with a sweet, floral, herbaceous aroma [
      • Battaglia S.
      The complete guide to aromatherapy.
      ]. Lavender oil is a multi-ingredient mixture that contains more than 160 substances. The major components of lavender oil are linalool, linalyl acetate, 1,8-cineole, b-ocimene, terpinen-4-ol, and camphor [
      • Kasper S.
      • Gastpar M.
      • Müller W.E.
      • Volz H.P.
      • Möller H.J.
      • Dienel A.
      • et al.
      Efficacy and safety of silexan, a new, orally administered lavender oil preparation, in subthreshold anxiety disorder - evidence from clinical trials.
      ].
      Silexan, a proprietary essential oil from Lavandula angustifolia flowers, has been approved in Germany and several other countries for the oral treatment of anxiety [
      • Müller W.E.
      • Sillani G.
      • Schuwald A.
      • Friedland K.
      Pharmacological basis of the anxiolytic and antidepressant properties of Silexan®, an essential oil from the flowers of lavender.
      ]. Silexan® showed pronounced anxiolytic effects in patients with subthreshold anxiety disorders [
      • Kasper S.
      • Gastpar M.
      • Müller W.E.
      • Volz H.P.
      • Möller H.J.
      • Dienel A.
      • et al.
      Silexan, an orally administered Lavandula oil preparation, is effective in the treatment of 'subsyndromal' anxiety disorder: a randomized, double-blind, placebo controlled trial.
      ,
      • Kasper S.
      An orally administered lavandula oil preparation (Silexan) for anxiety disorder and related conditions: an evidence based review.
      ] at a daily oral dose of 80 mg (1 capsule) as well as in anxiety-related restlessness and agitation [
      • Kasper S.
      • Anghelescu I.
      • Dienel A.
      Efficacy of orally administered Silexan in patients with anxiety-related restlessness and disturbed sleep--A randomized, placebo-controlled trial.
      ] and generalized anxiety disorder for daily single doses of 80 mg and 160 mg [
      • Woelk H.
      • Schläfke S.
      A multi-center, double-blind, randomised study of the lavender oil preparation Silexan® in comparison to lorazepam for generalized anxitety disorder.
      ,
      • Kasper S.
      • Gastpar M.
      • Müller W.E.
      • Volz H.P.
      • Möller H.J.
      • Schläfke S.
      • et al.
      Lavender oil preparation Silexan is effective in generalized anxiety disorder--a randomized, double-blind comparison to placebo and paroxetine.
      ,
      • Kasper S.
      • Möller H.J.
      • Volz H.P.
      • Schläfke S.
      • Dienel A.
      Silexan in generalized anxiety disorder: investigation of the therapeutic dosage range in a pooled data set.
      ]. Moreover, evidence for antidepressant-like properties of Silexan has been observed in anxious patients suffering from comorbid depressive symptoms and in patients with mixed anxiety–depression disorder [
      • Kasper S.
      • Volz H.P.
      • Dienel A.
      • Schl¨afke S.
      Efficacy of Silexan in mixed anxiety-depression -A randomized, placebo-controlled trial.
      ], which may indicate intrinsic antidepressant-like properties independent of its anxiolytic activity [
      • Friedland K.
      • Silani G.
      • Schuwald A.
      • Stockburger C.
      • N¨oldner M.
      • Koch E.
      • et al.
      Neurotrophic properrties of Silexan®, an essential oil from the flowers of lavender - preclinical for antidepressant-like properties.
      ].
      The neuroendocrine response to the stressors involves the activation of the hypothalamic–pituitary–adrenocortical axis, resulting in the release of the glucocorticoid hormone cortisol from the adrenal cortex into blood, and the autonomic response is the activation of the sympathetic-adrenergic system, culminating in the release of adrenaline and noradrenaline from adrenal medulla into the blood circulation [
      • Micale V.
      • Drago F.
      Endocannabinoid system, stress and HPA axis.
      ]. Cortisol is produced in the adrenal cortex and is the main glucocorticoid hormone in humans. It is released in response to various psychosocial stimuli, such as anxiety and stress via hypothalamus–pituitary–adrenal axis. Endocrinological stress markers such as cortisol are useful for objectively evaluating psychosocial distress, including stress or anxiety. In addition to self-reporting anxiety measure, physiological parameters including blood pressures, heart rate, or salivary cortisol level are useful for evaluating anxiety objectively.
      The existing evidence has suggested anxiolytic and antidepressant properties of lavender based on clinical trials. However, the evidence based on systematic review and meta-analysis has concentrated on anxiolytic effects exclusively [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ,
      • Möller H.J.
      • Volz H.P.
      • Dienel A.
      • Schläfke S.
      • Kasper S.
      Efficacy of Silexan in subthreshold anxiety: meta-analysis of randomised, placebo-controlled trials.
      ,
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ]. And the reviews have presented the overall anxiolytic effects with substantial heterogeneity, but the potent source of variations in effects, such as study design, sample characteristics, or intervention characteristics, has not yet been identified adequately.
      The first aim of the present review is to identify the overall effects of lavender for anxiety and its physiological referents and depression. The second aim is to investigate moderating factors for substantial variations in effect on anxiety and depression. Specifically, we assumed that the effects of lavender might vary with the study characteristics comprising the routes of administration, sessions of intervention, and health conditions of populations. The results of this review could provide a scientific evidence for applying lavender for the amelioration of anxiety and depression levels.

      Methods

      Study design

      A systematic review and meta-analysis was performed to examine the effects of lavender on anxiety, depression, and physiological parameters following the PRISMA guidelines.

      Eligibility criteria

      Study characteristics used as criteria for eligibility are as follows: (1) population: clinical trials with human subjects of any age, sex, and with or without diseases were included; (2) intervention: lavender administration with any route of administration, any type of preparation, and any species of lavender; (3) comparator: no intervention, standard or routine care, or placebo; (4) outcomes: primary outcomes were anxiety and depression measured by validated or standardized measures; secondary outcomes were physiological parameters of anxiety, that is, blood pressures, heart rate, or salivary cortisol; and (5) study design: randomized controlled trials (RCTs). We excluded RCT studies that compared different types of lavender preparations without a control group or used combined lavender treatments. Trials with missing essential data were excluded from qualitative and quantitative synthesis. Trials with animal subjects were excluded.
      Report characteristics used as eligibility criteria are all studies written in English and published from 2010 to 2019. Since recent systematic reviews and meta-analyses on assessing the effect of lavender in the treatment of anxiety screened up to November 2018 [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ,
      • Möller H.J.
      • Volz H.P.
      • Dienel A.
      • Schläfke S.
      • Kasper S.
      Efficacy of Silexan in subthreshold anxiety: meta-analysis of randomised, placebo-controlled trials.
      ,
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ], we limited publication year from 2010 to 2019 for up-to-date evidence and avoiding duplication of results. Trials regardless of publication status were all included except for those published in abstract form only.

      Information sources

      The title/abstract/keywords fields of Cochrane Library, MEDLINE and PubMed Central via PubMed, Embase, Cumulative Index of Nursing and Allied Health Literature, and Web of Science databases were systematically searched for eligible articles. We checked out references of articles retrieved from database searches to locate additional relevant articles.

      Search

      We searched databases and references from located articles from May 15, 2019, to June 15, 2019. We used Boolean operators to search for the following terms: (lavender OR lavandula OR silexan) AND (anxiety OR anxious OR anxiolytic OR stress OR depression OR depressive) and derivatives of those terms, including MeSH thesaurus terms. We set additional filters to publication years: “from 2010 to 2019,” article type: “randomized controlled trial,” language: “English,” and species: “humans” in the database searches. The full electronic search strategy for MEDLINE and PubMed Central was presented in supplementary material (Appendix E).

      Study selection and data collection process

      Two reviewers (E.N. and Y.L.) performed eligibility assessment individually. In case of disagreement, the items were discussed and resolved by consensus between the two reviewers. We then established a coding structure for data extraction and pilot-coded it on five randomly selected included trials and revised it accordingly. Two of the authors (M.K. and H.K.) independently coded the data, and the other two authors (E.N. and Y.L.) checked the coded data. Disagreements were resolved through discussion among all reviewers.

      Data items

      The authors extracted the following descriptive and numerical data from the included studies: (1) settings and characteristics of participants; (2) intervention (such as a method of application, dose, frequency, and duration of lavender aromatherapy); (3) measured outcomes of anxiety, depression, and physiological parameters of anxiety; (4) comparator interventions; (5) adverse effects of the intervention; (6) numerical data for meta-analysis: mean, standard deviation, randomized, and analyzed sample sizes of treatment groups, sessions and doses of interventions, and length of follow-up.

      Risk of bias in individual studies

      Bias assessment was performed by two independent assessors (M.K. and H.K.) based on the primary outcome (self-rated anxiety and depression) level using the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [
      • Sterne J.A.C.
      • Savović J.
      • Page M.J.
      • Elbers R.G.
      • Blencowe N.S.
      • Boutron I.
      • et al.
      RoB 2: a revised tool for assessing risk of bias in randomised trials.
      ]. Disagreements were resolved by discussion between reviewers (E.N. and Y.L.) until consensus was made. As we planned to estimate the effect of starting and adhering to lavender intervention, we assessed the risk of bias based on per-protocol analysis.
      The RoB 2.0 for individually randomized trials has five domains, including bias (1) from the randomization process, (2) due to deviations from intended interventions, (3) from missing outcome data, (4) in measurement of the outcome, and (5) in selection of the reported result. Each risk of bias domain has three response options, comprising low, some concerns, and high risk of bias. One of the key innovations of the RoB 2.0 is automatic judgement of overall risk of bias via algorithm by the risk of bias judgments of the individual domains in each study.

      Summary measures and synthesis of results

      In the meta-analysis of all outcomes, including anxiety, depression, and physiological parameters, the bias-corrected standardized mean difference (Hedges' ĝ) was calculated as the effect size metric. Standardized mean differences are upwardly biased when samples are small, especially less than 20 participants, and Hedges suggested a correction for small sample bias, known as Hedges' ĝ [
      • Littell J.H.
      • Corcoran J.
      • Pillai V.
      Systematic reviews and meta-analysis.
      ].
      We assessed the heterogeneity in effects using I2 statistics and Cochran's Q based on Chi-squared statistics. If an I2 value was greater than 50% and the p value of Chi-squared was below 0.1, we concluded that there was substantial heterogeneity.
      We used the inverse variance weighting for pooling the results of individual studies. Based on the assumption that the true effect might vary across samples and populations, depending on the health conditions of populations, type or sessions of interventions, and study design artifacts, we estimated the mean effects using the random effect model.
      Meta-analyses were calculated in R software version 4.0.2 [
      ] using packages meta and metafor. We also performed meta-analyses in Review Manager 5.4 (Version: 5.4.1) [
      The Cochrane Collaboration
      Review manager (RevMan) [Internet].
      ] using non-Cochrane mode.

      Risk of bias across studies

      Publication bias across studies was assessed with funnel plot followed by linear regression of intervention effect estimate against its standard error (Egger's regression) as tests for funnel plot asymmetry. Publication bias was assessed only when at least 10 studies were included in the meta-analysis because when there are fewer studies, the power of the tests is too low to distinguish chance from real asymmetry [
      ]. We assessed publication bias in each meta-analysis on the primary outcomes anxiety and depression.

      Additional analyses

      When substantial heterogeneity in effects in any meta-analysis was identified, moderator analysis by meta-analysis of variance (ANOVA) and/or meta-regression was performed to examine factors creating variations in effect sizes across studies.
      In each meta-analysis, subgroup analysis and/or moderator analysis was performed to see whether lavender intervention might have differential effects for different subgroups by study characteristics. Moderator analysis can be performed by two main statistical methods including meta-ANOVA and meta-regression; both approaches require at least 10 studies for every moderator in the analysis. In the meta-analysis on anxiety, we performed three moderator analyses by the route of administration, health state of participants, and sessions of treatment. Meta-regression was used to identify the amount of heterogeneity accounted for by each moderator.
      In the meta-analysis on depression, as the number of included studies was 10, we performed a moderator analysis by the route of administration of lavender.
      To identify the effect of the risk of bias assessments for the variation of mean effect, we performed a sensitivity analysis to examine whether inclusion of the studies at high overall bias influences the mean effect. And we performed a subgroup analysis by assessment of risk to examine the difference between studies at high, some concerns, and low risk.

      Results

      Study selection

      A total of 562 citations were retrieved through database searches, and additional 12 trials were identified by reviewing the references of the selected articles. After duplicates were excluded, 378 studies remained. Then we evaluated the titles and abstracts and excluded 298 articles. The remaining 80 full-text articles were screened for eligibility, and 42 articles were excluded. Finally, 38 articles were included in qualitative analysis, and 37 articles were included in quantitative synthesis. Details of the process of screening and selection of the studies were presented in Figure 1. The final included articles are listed in the supplementary material (Appendix A).
      Figure 1
      Figure 1Flow diagram of the study selection process.Note. RCT = randomized controlled trial; SD = standard deviation.

      Study characteristics

      Characteristics of all included studies were summarized in the supplementary materials (Appendix B). Methods and overview of the studies are as follows: 38 RCTs published in English from 2010 to 2019 were included in qualitative synthesis, and 37 of 38 studies were included in quantitative synthesis. Geographic origins of the studies are Iran (17 trials), Turkey (8), Germany (4), Greece (1), India (1), South Korea (2), Taiwan (3), the United States (1), and Thailand (1).
      Across all studies included in quantitative analyses, a total of 4316 participants were randomized to either lavender (2165) or control treatment (2151). In the meta-analysis on anxiety, a total of 3906 participants (lavender 1955 and control 1951) were randomized, and 3825 (lavender 1917 and control 1908) were analyzed. In depression, a total of 1312 participants (lavender 657 and control 655) were randomized, and 1282 (lavender 644 and control 638) participants were analyzed. Studies included in meta-analysis on cortisol a total of 206 participants were randomized to either lavender (102) or control treatment (104), and 180 (lavender 96, control 94) were analyzed. Pooled premature withdrawal rates were 1.94% for lavender and 2.2% for control group in analysis for anxiety, 1.98% for lavender and 2.60% for control group in analysis for depression, and 5.9% for lavender and 9.6% for control in analysis for salivary or serum cortisol.
      The populations of the studies included in the analysis for anxiety consisted of patients undergoing surgery or invasive procedure, critically ill patients with cardiac diseases or in intensive care units, healthy students under stressful conditions, pregnant or postpartum women, and patients in anxiety and/or depressive disorders. The participants of studies on depression were composed of patients undergoing hemodialysis, women in pregnancy, postpartum, or menopause, patients in anxiety and/or depression or dementia, or healthy students with premenstrual syndrome.
      The participants in the experimental group received one of four routes of administration of lavender: inhalation, massage, tea, or oral preparation (silexan). The participants in the control group received standard or routine care, placebo, or no treatment. The details of dose, duration, and sessions of experimental and control treatments are presented in supplementary material (Appendix B).
      The primary outcomes measured were anxiety and depression, and the secondary outcomes measured were physiological indicators of anxiety. Of all 38 included studies, self-rated anxiety was assessed in 30 studies, and depression was evaluated in 10 studies. Anxiety was measured by standardized measure (visual analog scale) or validated measures (Beck Anxiety Inventory, Depression Anxiety Stress Scale, Hospital Anxiety and Depression Scale, Hamilton Anxiety Scale, Modified Dental Anxiety Scale, State-Trait Anxiety Inventory, or Zung Self-Rating Scale). Depression was measured by validated measures including Beck Depression Inventory, Cornell Scale for Depression in Dementia–Chinese version, Edinburgh Postnatal Depression Scale, Premenstrual syndrome (depressive affect subscale), Hospital Anxiety and Depression Scale, Hamilton Rating Scale for Depression, and Montgomery Åsberg Depression Rating Scale.
      Blood pressures were assessed in seven studies and heart rate was assessed in six studies, salivary cortisol was assessed in two studies, and serum cortisol was assessed in one study.

      Risk of bias within studies

      Figure 2A shows the risk of bias summary presenting the assessment in each domain and overall risk of bias. The overall risk of bias was evaluated as low in 16 trials, as some concern in 16 trials and as high risk of bias in five trials.
      Figure 2
      Figure 2A. Risk of bias summary according to the revised Cochrane risk-of-bias tool for randomized trials (R0B 2). D1: Randomization process; D2: Deviations from intended interventions; D3: Missing outcome data; D4: Measurement of the outcome; D5: Selection of the reported result. B. Risk of bias graph according to the revised Cochrane risk-of-bias tool for randomized trials (R0B 2).
      The risk of bias from the randomization process was rated as high in only one study, some concern in 10 studies, and low in the remaining 26 studies. Bias in measurement of the outcome was assessed as high in four trials, some concern in 15, and low in 18 trials. Bias due to deviations from intended interventions was rated as some concern risk in only two trials and low risk in the remaining 35 trials. Figure 2B presents the risk of bias graph of included studies according to the revised Cochrane risk-of-bias tool for randomized trials (ROB 2).

      Results of individual studies and synthesis of results

      We first conducted basic meta-analyses for each outcome: anxiety levels, physiological parameters (systolic blood pressure, diastolic blood pressure, heart rate, and cortisol levels), and depression levels.
      As treatment effects are inconsistent across study characteristics of populations, sessions, durations, and types of intervention, comparison conditions, and methodological features, we hypothesized that the route of administration, sessions of treatment, and health state of population might moderate variations in effect sizes across studies. Subgroup and moderator analysis may be used to explore possible sources of variability in combined effects. According to Cochrane Handbook for Systematic Reviews of Interventions version 6.1 [
      ], subgroup and moderator analysis require at least 10 studies for each moderator. The reason is the statistical power of moderator analysis is affected by the number of included studies and moderators. As our meta-analysis on anxiety included 30 studies, we conducted three subgroup and moderator analyses by route of administration, health state of population, and sessions of treatment using meta-analysis of variance or meta-regression whether the moderator variable is continuous or categorical. Also, the analysis on depression included 10 studies; only one moderator analysis by the route of administration was conducted.
      Finally, we performed a sensitivity analysis to determine the effect of study quality on the mean effect. Specifically to examine the inclusion of studies at high risk of bias affects the overall effect on anxiety levels, sensitivity analysis deleting each study was done.
      The meta-analysis for self-rated anxiety included 30 studies to evaluate the overall effects of lavender intervention (Figure 3A). Lavender was significantly superior to comparators (standard care, placebo, or no treatment). The mean effect (Hedges’ ĝ) was −0.72 (95% confidence interval [CI] −0.90 to −0.55), and the direction of the mean effect favored lavender. The analysis also showed that lavender intervention was significantly superior to comparator in 21 of 30 trials, with the largest effect of −2.56 (95% CI −3.25 to −1.86). The heterogeneity statistics were I2 = 84%, p > .001, indicating substantial heterogeneity.
      Figure 3
      Figure 3A. Forest plot on the efficacy of lavender on self-rated anxiety levels. B. The effect of lavender on physiological parameters C. The effect of lavender on depression levels. Note. CI = confidence interval; SMD = standardized mean difference; df = degrees of freedom; IV = inverse variance; SD = standard deviation; SBP = systolic blood pressure; DBP = diastolic blood pressure; HR = heart rate.
      Figure 3
      Figure 3A. Forest plot on the efficacy of lavender on self-rated anxiety levels. B. The effect of lavender on physiological parameters C. The effect of lavender on depression levels. Note. CI = confidence interval; SMD = standardized mean difference; df = degrees of freedom; IV = inverse variance; SD = standard deviation; SBP = systolic blood pressure; DBP = diastolic blood pressure; HR = heart rate.
      Figure 3
      Figure 3A. Forest plot on the efficacy of lavender on self-rated anxiety levels. B. The effect of lavender on physiological parameters C. The effect of lavender on depression levels. Note. CI = confidence interval; SMD = standardized mean difference; df = degrees of freedom; IV = inverse variance; SD = standard deviation; SBP = systolic blood pressure; DBP = diastolic blood pressure; HR = heart rate.
      The results for the physiological parameters are presented in Figure 3B. The efficacy on the self-rated anxiety was not supported by the physiological parameters except for systolic blood pressure (SBP). Each meta-analysis for SBP and diastolic blood pressure (DBP) included seven studies, and six studies for heart rate, and three studies were included in the analysis for cortisol. The effect size on the SBP was −0.23 (95% CI −0.41 to −0.05). The effect of lavender on DBP was −0.15, the effect on heart rate was −0.2, and the effect on salivary/serum cortisol was −1.4. However, the effects of DBP, heart rate, and cortisol showed no significance.
      The meta-analysis on the antidepressive effect included 10 studies. The meta-analysis showed that lavender was superior to placebo or no treatment comparators with the mean effect of −0.43 (95% CI −0.59 to −0.27; Figure 3C). The meta-analysis showed that lavender was superior to control treatment significantly in 7 of 10 RCTs, with the treatment effects ranging −0.18 to −1.2. The statistics of heterogeneity showed a significant medium size heterogeneity (I2 = 47%, p = .05).

      Risk of bias across studies

      To evaluate potential publication bias, funnel plots were drawn, and then Egger's regression test was performed for self-rated anxiety and depression (see Funnel plots in Supplementary material, Appendices C and D). The funnel plot for standard error and effect sizes on anxiety seemed somewhat asymmetrical, seemingly empty in the lower-right area. However, Egger's regression showed no evidence of significant publication bias (t = −1.04, df = 28, p = .308). Egger's regression on depression also showed no evidence of publication bias (t = −1.61, df = 8, p = .146).
      Because the publication biases for both anxiety and depression showed no significant evidence of biases, the results of our meta-analyses on anxiety and depression could be regarded as representative of the population of all published studies.

      Additional analysis

      Although the overall effect of the lavender treatment on self-rated anxiety levels showed significant medium to large size (ĝ = −0.72, 95% CI: −0.90 to −0.55), the effect sizes of individual studies around the mean effect showed substantial variability. The heterogeneity statistics showed I2 = 84%, Chi-squared = 104.75, df = 29, and p > 0.001. Intervention effects are often heterogenous across study characteristics, including populations, interventions, comparisons, measures of outcomes, and methodological features. These study factors can moderate the effects of interventions and be sources of heterogeneity in combined effects.
      Subgroup and moderator analyses can be done for the purpose of investigating heterogeneous results or to explore specific questions about particular patient populations, methods of intervention, or quality of study [
      ]. In subgroup analysis, the test of significance indicates whether effects were significant within subgroups not whether differences in effects were significant between subgroups. Moderator analysis provides tests of the differences in effects between subgroups and influences of moderators on the overall effect. The two statistical methods for moderator analysis are meta-analysis of variance (ANOVA) and meta-regression, both approaches require at least 10 studies for each moderator to ensure statistical power of the analysis.
      We hypothesized in the protocol that the effects of lavender could vary with study characteristics, including routes and sessions of administration of lavender, health conditions of populations, and methodological quality, but the number of included studies in the meta-analysis for anxiety levels was 30; we conducted three moderator analyses, including (1) the routes and (2) sessions of administration and (3) health conditions of populations. The results of subgroup/moderator analysis for the anxiolytic effect are presented in Table 1.
      Table 1Subgroup/Moderator Analyses for the Impact of Study Characteristics on Anxiolytic Effect of Lavender.
      Moderators/subgroupskHedges' ĝ95% CIQb (df)pb
      Route of administration
       Massage5−0.60−1.02, −0.183.35(2).187
       Inhalation21−0.83−1.03, −0.62
       Silexan 80 mg4−0.41−0.84, 0.03
      Meta-regressionQM(df = 2) = 3.35, p = .187
      R2 (amount of heterogeneity accounted for): 6.5%
      Health condition of populations
       Patients undergoing surgery or invasive13−0.79−1.05, −0.535.63(3).131
       Patients with coronary disease or in ICU7−1.00−1.35, −0.64
       Healthy population6−0.53−0.90, −0.15
       Anxiety or depression4−0.41−0.83, 0.02
      Meta-regressionQM (df = 3) = 5.63, p = .131
      R2 (amount of heterogeneity accounted for): 8.9%
      Overall risk of within-study bias
       Some concern14−0.81−1.08, −0.530.68(2).713
       Low11−0.64−0.93, −0.34
       High5−0.74−1.20, −0.28
      Meta-regressionQM (df = 2) = 0.68, p = .713
      R2 (amount of heterogeneity accounted for): 0.0%
      Sessions of intervention
      Meta-regressionQM(df = 1) = 3.68, p = .055
      R2 (amount of heterogeneity accounted for): 13.2%
      Overall effectkHedges' ĝ95% CIQ (df)pI2
      30−0.72−0.90, −0.55184.75(29)<.00184.3%
      Note. CI = confidence interval; df = degree of freedom; ICU = intensive care unit; Qb = Q between groups; QM = Q moderator; pb = p between groups.
      The subgroup analysis by the route of administration showed that the application of lavender using massage, inhalation, and oral administration (silexan) was significantly superior to standard care, placebo, or no treatment comparators. The mean effect of lavender inhalation was −0.83, and the effect of lavender using massage was −0.60, and silexan showed the smallest effect of −0.41.
      The moderator analysis using meta-ANOVA to test statistical significance for different mean effects between routes of lavender indicated no statistical significance (Qb(df) = 3.35(2), pb = .187). To examine the possible impact of the route of administration for the heterogeneity of the overall effect of lavender on anxiety, we conducted additional moderator analysis using meta-regression. Meta-regression can be used to assess the potential impact of one or more continuous or categorical moderators. Categorical variable of the route of administration was expressed as a set of dummy variables with one omitted category in the meta-regression. Therefore, our meta-regression showed that routes of administration accounted for 6.5% for the heterogeneity of the mean effect; the result showed no significance just the same as the statistical result of meta-ANOVA.
      As the statistical power of moderator analysis is affected by the number of studies and the number of moderators, we should be cautious to interpret the results. The moderator analyses displayed meaningful results that meta-ANOVA showed that treatment effects according to the routes of administration were clearly different, and the meta-regression indicated the route of administration accounted 6.5% for the variance of mean effect. But both moderator analyses showed no statistical significance; we can consider the possibility of the lack of statistical power by insufficient number of studies in the analyses.
      The subgroup analysis for the effect of lavender on anxiety by health conditions of populations showed that the mean effect was −0.79 (95%CI: −1.05 to −0.53) for populations undergoing surgery or invasive treatment, −1.00 (−1.35, −0.64) for the populations with coronary diseases or patients in intensive care unit (ICU) group, −0.53 (−0.90, −0.15) for healthy population and −0.41 (−0.83, 0.02) for populations in anxiety and/or depression conditions. Lavender aromatherapy showed the largest efficacy in population with coronary diseases and/or patients in ICU and the efficacy for population with anxiety and/or depression showed the smallest and no statistical significance.
      We performed both meta-ANOVA and meta-regression to explore possible influence of health conditions of populations on the variation of the mean effect. The meta-ANOVA indicated that different mean effects between health conditions of population were not statistically significant (Qb(df) = 5.63(3), pb = .131). The mean effects of subgroups of health condition of population were considerably different and the meta-regression showed that health conditions of population accounted 8.9% for the heterogeneity in mean effect (QM(df = 3) = 5.63, p = .131), but both moderator analyses showed no statistical significance. This might be ascribable to possible lack of power because of insufficient number of studies.
      The meta-regression to test moderating effect of sessions of treatment showed that sessions of intervention accounted 13.18% for variations of mean effects significantly (QM = 3.68, df = 1, p = .05). Regression coefficient of sessions was b = 0.0057, and the regression equation can be suggested as Y = 0.0057∗sessions – 0.82. This result means that the more sessions of treatment make the larger effect sizes on anxiety.
      Consequently, we could explain the moderators such as previously mentioned route of administration of lavender and health states of population and sessions of administration can reasonably influence the variability of the mean effect; only the test of statistical significance failed because of possible insufficient statistical power. Random effect model easily concludes nonsignificance.
      We performed a subgroup and moderator analysis to test the different means of antidepressive effects by routes of administration (Table 2). The subgroup analysis showed the mean effect of inhalation on depression was −0.43 (95% CI −0.77 to −0.08), the mean effect of massage was −0.63 (95% CI −1.08 to −0.17), the mean effect of silexan was −0.42 (95% CI −0.72 to −0.12). Lavender tea showed no significant effect, and effect size was the smallest (−0.32).
      Table 2Subgroup/Moderator Analysis on the Impact of the Route of Administration on Antidepressive Effects of Lavender.
      Moderator/subgroupkHedges' ĝ95% CIQb (df)pb
      Route of administration
       Inhalation3−0.42−0.77, −0.080.95(3).814
       Tea2−0.32−0.76, 0.12
       Massage2−0.63−1.08, −0.17
       Silexan 80 mg3−0.42−0.72, −0.12
      Meta-regressionQM(df = 3) = 0.95, p = .814
      R2 (amount of heterogeneity accounted for): 0.0%
      Overall effectkHedges' ĝ95% CIQ (df)pI2
      10−0.43−0.59, −0.2716.88 (9).05146.7%
      Note. CI = confidence interval; df = degree of freedom; Qb = Q between groups; QM = Q moderator; pb =p between groups.
      The moderator analysis by meta-ANOVA indicated that the different mean effects between inhalation, massage, tea, and silexan was not significant (Q = 0.95, df = 3, pb = .814). And also, the meta-regression indicated that the route of administration did not account for heterogeneity in mean effect (0%).
      Finally, we performed a sensitivity analysis to examine whether the risk of within-study bias influence the mean effect of lavender on anxiety (Figure 4). Specifically, to examine the inclusion of studies at high overall risk impact the mean effect, we estimated the mean effect after deleting each study at high overall risk.
      Figure 4
      Figure 4Sensitivity analysis: the mean effect after dropping each study at high overall bias for anxiety levels; studies enclosed by square dotted line are at high overall bias. Note. CI = confidence interval; SMD = standardized mean difference.
      According to sensitivity analysis, the mean effect changed to −0.68 from −0.66 after deleting the study of Bekhradi [
      • Bekhradi R.
      • Vakilian K.
      The effect of lavender aromatherapy on test anxiety in female students.
      ], −0.66 after deleting Hosseini [
      • Hosseini S.
      • Heydari A.
      • Vakili M.
      • Moghadam S.
      • Tazyky S.
      Effect of lavender essence inhalation on the level of anxiety and blood cortisol in candidates for openheart surgery.
      ], −0.65 after deleting Senturk [
      • Şentürk A.
      • Tekinsoy Kartın P.
      The effect of lavender oil application via inhalation pathway on hemodialysis patients’ anxiety level and sleep quality.
      ], −0.68 after deleting Yayla [
      • Yayla E.M.
      • Ozdemir L.
      Effect of inhalation aromatherapy on procedural pain and anxiety after needle insertion into an implantable central venous port catheter: a quasi-randomized controlled pilot study.
      ], and −0.62 after deleting Zabirunnisa [
      • Zabirunnisa M.
      • Gadagi J.S.
      • Gadde P.
      • Myla N.
      • Koneru J.
      • Thatimatla C.
      Dental patient anxiety: possible deal with Lavender fragrance.
      ]. Consequently, deleting each study at high overall risk did not change the mean effect significantly. So the mean anxiolytic effect of lavender demonstrated relatively robust and does not seem to be sensitive to the inclusion of the studies at high risk.

      Discussion

      Summary of evidence

      The use of lavender essential oil has become popular in aromatherapy, and its therapeutic efficacy has been assessed in a large number of clinical trials. Aromatherapy with lavender essential oil was found to be effective in decreasing anxiety and its comorbid depression in various settings. Physiological parameters did not demonstrate consistent effects among parameters. Lavender showed significant decrease in systolic blood pressure but did not affect DBP, heart rate, and salivary or serum cortisol significantly.
      Our meta-analysis demonstrated that lavender is superior to controls, including standard care, placebo, or no treatment in decreasing self-rated anxiety in diverse populations. The overall risk of bias in the primary studies assessed with revised ROB tool displayed that 5 of 37 studies were rated as high risk. But judging from the sensitivity analysis, deleting each high-risk study did not change the mean effect distinctly, implying the effect sizes of high-risk studies might not be overestimated. Consequently, the overall effects demonstrated relatively robust and do not seem to be influenced by the study quality of the included studies.
      Our meta-analysis confirmed the results of Kang et al [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ] and Donelli et al [
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ] in the efficacy for a significant decrease in anxiety, although the magnitude of the effect varied slightly. The mean effect on anxiety levels of −0.72 can be interpreted as medium to large [
      • Cohen J.
      Statistical power analysis for the behavioral sciences.
      ]. This effect is larger than the evidence (−0.65) of the review of Kang et al [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ], which synthesized 19 studies published 2000 to 2019. As the present review included 30 studies published from 2010 to 2019, the change of effect sizes between the reviews is noteworthy.
      As our meta-analysis included studies administering routes of inhalation, massage, and oral silexan and also included studies comprising participants in diverse health states, the analysis demonstrated substantial heterogeneity in effects.
      According to subgroup analysis by the route of administration, the effect size of inhalation was the largest other than massage and silexan. Our meta-analysis confirmed the results of Kang et al [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ] and Donelli et al [
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ] in the efficacy of inhalation for a significant decrease in anxiety, although the magnitude of the effect varied slightly. The effect estimate of inhalation (−0.83) is slightly larger than other evidence −0.71 [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ] and −0.73 [
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ]). The result of our review and previous reviews suggest that inhalation of lavender oil is effective in decreasing anxiety levels in high anxiety inducing situations considerably. The inhalation of lavender essential oil can be recommended as efficacious intervention to decreasing anxiety in people in diverse situations of anxiety.
      The massage with lavender oil showed medium to large anxiolytic effect (−0.60). This effect estimate is similar to other evidence (−0.61 [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ], −0.66 [
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ]). Therefore, combined with the previous evidence, the massage with lavender oil can be interpreted to have substantial effect of relieving anxiety for populations in anxiety conditions.
      The oral lavender silexan also confirmed a significant anxiolytic effect. The included studies in meta-analysis for silexan showed high study quality of all low risk in overall risk of bias. The result confirms the evidence of Kang et al [
      • Kang H.J.
      • Nam E.S.
      • Lee Y.
      • Kim M.
      How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
      ], Donelli et al [
      • Donelli D.
      • Antonelli M.
      • Bellinazzi C.
      • Gensini G.F.
      • Firenzuoli F.
      Effects of lavender on anxiety: a systematic review and meta-analysis.
      ], and Möller et al [
      • Möller H.J.
      • Volz H.P.
      • Dienel A.
      • Schläfke S.
      • Kasper S.
      Efficacy of Silexan in subthreshold anxiety: meta-analysis of randomised, placebo-controlled trials.
      ], although the magnitude of the effect is different slightly and the effect measures are different (Hedges’ ĝ vs. weighted mean difference).
      The analysis for publication bias by funnel plot and Egger's regression showed no evidence of publication bias, which signifies our sample of meta-analysis may be representative of the population of published studies on this topic. Consequently, the evidence of the anxiolytic effect of lavender can be interpreted as fairly robust considering the quality of research designs, no evidence of publication bias, and CIs that do not cross the line of no effects.
      Vital signs and salivary cortisol are recognized as important physiological measures that indirectly indicate anxiety. Our results indicated that lavender has a decreasing effect on SBP. The effect size is small, but it can be interpreted as meaningful change because systolic blood pressure may be difficult to change. The risk of bias of included studies on physiological measures was low, as the effects of these outcomes would not be affected by participants’ awareness of intervention. Therefore, study quality might not influence treatment effects. In conclusion, the effect of lavender on blood pressure is small but not weak based on consistent effect sizes, significant effect, and strong study quality.
      The efficacy of lavender on diastolic pressure and heart rate showed small effect sizes of −0.15 and −0.20 and no significant effects. The effect of cortisol was −1.4 with no statistical significance. We can interpret that there is no evidence that the mean effect is statistically different from no effect. However, no evidence of an effect is not the same as evidence of no effect, that is to say, no significant effects do not prove that there is no effect. An alternative explanation may be because of too small sample size or too much heterogeneity. The results of DBP, heart rate, and cortisol can be attributed to insufficient statistical power due to overly few studies because statistical power is affected by the number of studies in the meta-analysis. Therefore, we recommend future studies of RCTs investigating the anxiolytic efficacy of lavender on physiological or endocrinological stress markers such as vital signs or cortisol.
      Depression has been recognized as a major comorbidity symptom of anxiety. Our results demonstrate that lavender has a favorable relieving effect on depression levels. The mean effect was medium effect size according to Cohen's standard [
      • Cohen J.
      Statistical power analysis for the behavioral sciences.
      ]. Subgroup analysis indicated that route of lavender application tea, massage, and silexan showed significant antidepressive effects, and massage with lavender demonstrated the largest effect size. Only inhalation showed no significant effect. The evaluation of the risk of bias across studies indicated that there is no evidence of publication bias. This evidence on antidepressive effect of lavender is not able to be compared with other evidence because published evidence on this topic could not be located.
      In conclusion, lavender aromatherapy by means of massage, silexan, or tea significantly decreases depression in people with various health conditions. There is some evidence of the efficacy of lavender on depression levels on the grounds that there was no evidence of publication bias, and the quality of studies showed no evidence of impact on the observed effect.

      Limitations

      To ensure study quality, we synthesized only RCTs on the effect of lavender on anxiety and depression; however, the risk of bias assessment showed that of all 37 studies included in quantitative analysis, 16 were rated as some concern of risk, and five studies were evaluated as high risk of overall risk of bias. Only 16 of 37 studies were at low risk of bias. Outcomes such as self-rated anxiety or depression can be influenced by outcome assessor's knowledge of intervention received. Therefore, our results on anxiety and depression could have been influenced by participants' knowledge of the intervention received, for example, inhalation of lavender or massage using lavender oil.
      Although the evaluation of publication bias in our quantitative analysis showed no evidence of risk of bias across studies, in our review process, we included only studies written in English, published reports, and accessible reports based on preset inclusion criteria. These limits in the locating and screening process might have introduced sampling bias.

      Conclusions

      Our meta-analysis confirmed the results of existing reviews on the effect of inhalation and massage applicating lavender essential oil for a significant decrease in anxiety levels, although the magnitude of the effect varied slightly. The effect of silexan also confirmed a significant anxiolytic effect of previous evidence.
      The effects on physiological parameters, including DBP, heart rate, and salivary or serum cortisol, showed small in effect sizes and no evidence of significant effects. Only systolic blood pressure displayed significant small effect size. The statistical power of the analyses on physiological parameters might be weak because of overly small samples, and the magnitude of effects was small. Therefore, more and larger randomized trials testing the effect of lavender aromatherapy for anxiety measured with physiological measures including vital signs or cortisol are recommended.
      Our analysis on the effect of the application of lavender for the treatment of depression demonstrated a beneficial effect on decreasing depression. The effect size on depression cannot be compared with the literature because published data on this topic could hardly be located. Our review included any type of participant, method of intervention, or outcome measure in primary studies investigating the efficacy of lavender aromatherapy. We recommend future reviews focusing on populations in specific health conditions and routes of application of lavender.

      Conflict of interest

      No potential or any existing conflict of interest relevant to this article was reported.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article:

      References

        • Stein D.J.
        • Scott K.M.
        • de Jonge P.
        • Kessler R.C.
        Epidemiology of anxiety disorders: from surveys to nosology and back.
        Dialogues Clin Neurosci. 2017; 19: 127-136https://doi.org/10.31887/DCNS.2017.19.2/dstein
        • Bandelow B.
        • Michaelis S.
        Epidemiology of anxiety disorders in the 21st century.
        Dialogues Clin Neurosci. 2015; 17: 327-335https://doi.org/10.31887/DCNS.2015.17.3/bbandelow
        • Baxter A.J.
        • Scott K.M.
        • Vos T.
        • Whiteford H.A.
        Global prevalence of anxiety disorders: a systematic review and meta-regression.
        Psychol Med. 2013; 43: 897-910https://doi.org/10.1017/S003329171200147X
        • Bandelow B.
        • Michaelis S.
        • Wedekind D.
        Treatment of anxiety disorders.
        Dialogues Clin Neurosci. 2017; 19: 93-107https://doi.org/10.31887/DCNS.2017.19.2/bbandelow
        • Kessler R.C.
        • Chiu W.T.
        • Demler O.
        • Merikangas K.R.
        • Walters E.E.
        Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the national comorbidity survey replication.
        Arch Gen Psychiatr. 2005; 62: 617-627https://doi.org/10.1001/archpsyc.62.6.617
        • Malcolm B.J.
        • Tallian K.
        Essential oil of lavender in anxiety disorders: ready for prime time?.
        Ment Health Clin. 2017; 7: 147-155https://doi.org/10.9740/mhc.2017.07.147
        • Kasper S.
        • Müller W.E.
        • Volz H.P.
        • Möller H.J.
        • Koch E.
        • Dienel A.
        Silexan in anxiety disorders: clinical data and pharmacological background.
        World J Biol Psychiatr. 2018; 19: 412-420https://doi.org/10.1080/15622975.2017.1331046
        • Rickels K.
        • Garcia-Espana F.
        • Mandos L.A.
        • Case G.W.
        Physician withdrawal checklist (PWC-20).
        J Clin Psychopharmacol. 2008; 28: 447-451https://doi.org/10.1097/JCP.0b013e31817efbac
        • Müller W.E.
        • Sillani G.
        • Schuwald A.
        • Friedland K.
        Pharmacological basis of the anxiolytic and antidepressant properties of Silexan®, an essential oil from the flowers of lavender.
        Neurochem Int. 2020; 143: 104899https://doi.org/10.1016/j.neuint.2020.104899
        • Cavanagh H.M.
        • Wilkinson J.M.
        Biological activities of lavender essential oil.
        Phytother Res. 2002; 16: 301-308https://doi.org/10.1002/ptr.1103
        • Battaglia S.
        The complete guide to aromatherapy.
        3rd ed. Black Pepper Creative, Brisbane2018: 360
        • Kasper S.
        • Gastpar M.
        • Müller W.E.
        • Volz H.P.
        • Möller H.J.
        • Dienel A.
        • et al.
        Efficacy and safety of silexan, a new, orally administered lavender oil preparation, in subthreshold anxiety disorder - evidence from clinical trials.
        Wien Med Wochenschr. 2010; 160: 547-556https://doi.org/10.1007/s10354-010-0845-7
        • Kasper S.
        • Gastpar M.
        • Müller W.E.
        • Volz H.P.
        • Möller H.J.
        • Dienel A.
        • et al.
        Silexan, an orally administered Lavandula oil preparation, is effective in the treatment of 'subsyndromal' anxiety disorder: a randomized, double-blind, placebo controlled trial.
        Int Clin Psychopharmacol. 2010; 25: 277-287https://doi.org/10.1097/YIC.0b013e32833b3242
        • Kasper S.
        An orally administered lavandula oil preparation (Silexan) for anxiety disorder and related conditions: an evidence based review.
        Int J Psychiatr Clin Pract. 2013; 17: 15-22https://doi.org/10.3109/13651501.2013.813555
        • Kasper S.
        • Anghelescu I.
        • Dienel A.
        Efficacy of orally administered Silexan in patients with anxiety-related restlessness and disturbed sleep--A randomized, placebo-controlled trial.
        Eur Neuropsychopharmacol. 2015; 25: 1960-1967https://doi.org/10.1016/j.euroneuro.2015.07.024
        • Woelk H.
        • Schläfke S.
        A multi-center, double-blind, randomised study of the lavender oil preparation Silexan® in comparison to lorazepam for generalized anxitety disorder.
        Phytomedicine. 2010; 17: 94-99https://doi.org/10.1016/j.phymed.2009.10.006
        • Kasper S.
        • Gastpar M.
        • Müller W.E.
        • Volz H.P.
        • Möller H.J.
        • Schläfke S.
        • et al.
        Lavender oil preparation Silexan is effective in generalized anxiety disorder--a randomized, double-blind comparison to placebo and paroxetine.
        Int J Neuropsychopharmacol. 2014; 17: 859-869https://doi.org/10.1017/S1461145714000017
        • Kasper S.
        • Möller H.J.
        • Volz H.P.
        • Schläfke S.
        • Dienel A.
        Silexan in generalized anxiety disorder: investigation of the therapeutic dosage range in a pooled data set.
        Int Clin Psychopharmacol. 2017; 32: 195-204https://doi.org/10.1097/YIC.0000000000000176
        • Kasper S.
        • Volz H.P.
        • Dienel A.
        • Schl¨afke S.
        Efficacy of Silexan in mixed anxiety-depression -A randomized, placebo-controlled trial.
        Eur Neuropsychopharmacol. 2016; 26: 331-340https://doi.org/10.1016/j.euroneuro.2015.12.002
        • Friedland K.
        • Silani G.
        • Schuwald A.
        • Stockburger C.
        • N¨oldner M.
        • Koch E.
        • et al.
        Neurotrophic properrties of Silexan®, an essential oil from the flowers of lavender - preclinical for antidepressant-like properties.
        Pharmacopsychiatry. 2021; 54: 37-46https://doi.org/10.1055/a-1293-8585
        • Micale V.
        • Drago F.
        Endocannabinoid system, stress and HPA axis.
        Eur J Pharmacol. 2018; 834: 230-239https://doi.org/10.1016/j.ejphar.2018.07.039
        • Kang H.J.
        • Nam E.S.
        • Lee Y.
        • Kim M.
        How strong is the evidence for the anxiolytic efficacy of lavender ? : systematic review and meta-analysis of randomized controlled trials.
        Asian Nurs Res (Korean Soc Nurs Sci). 2019; 13: 295-305https://doi.org/10.1016/j.anr.2019.11.003
        • Möller H.J.
        • Volz H.P.
        • Dienel A.
        • Schläfke S.
        • Kasper S.
        Efficacy of Silexan in subthreshold anxiety: meta-analysis of randomised, placebo-controlled trials.
        Eur Arch Psychiatr Clin Neurosci. 2019; 269: 183-193https://doi.org/10.1007/s00406-017-0852-4
        • Donelli D.
        • Antonelli M.
        • Bellinazzi C.
        • Gensini G.F.
        • Firenzuoli F.
        Effects of lavender on anxiety: a systematic review and meta-analysis.
        Phytomedicine. 2019; 65: 153099https://doi.org/10.1016/j.phymed.2019.153099
        • Sterne J.A.C.
        • Savović J.
        • Page M.J.
        • Elbers R.G.
        • Blencowe N.S.
        • Boutron I.
        • et al.
        RoB 2: a revised tool for assessing risk of bias in randomised trials.
        BMJ. 2019; 366: l4898https://doi.org/10.1136/bmj.l4898
        • Littell J.H.
        • Corcoran J.
        • Pillai V.
        Systematic reviews and meta-analysis.
        ([Internet]) Oxford University Press, Oxford2009 ([cited 2021 Feb 2]. Available from:)
      1. R Core Team. R: A language and environment for statistical computing [Internet]. R Foundation for Statistical Computing, Vienna, Austria2020 ([cited 2021 Feb 2]. Available from:)
        • The Cochrane Collaboration
        Review manager (RevMan) [Internet].
        (Version 5.4) Cochrane, London2020 ([cited 2021 Feb 2]. Available from:)
      2. Higgins J.P.T. Thomas J. Chandler J. Cumpston M. Li T. Page M.J. Welch V.A. Cochrane handbook for systematic reviews of interventions version 6.2 [Internet]. Cochrane, London2020 ([cited 2021 Jan 20]. Available from:)
        • Bekhradi R.
        • Vakilian K.
        The effect of lavender aromatherapy on test anxiety in female students.
        Curr Wom Health Rev. 2016; 12: 137-140https://doi.org/10.2174/1573404812666161021114923
        • Hosseini S.
        • Heydari A.
        • Vakili M.
        • Moghadam S.
        • Tazyky S.
        Effect of lavender essence inhalation on the level of anxiety and blood cortisol in candidates for openheart surgery.
        Iran J Nurs Midwifery Res. 2016; 21: 397-401https://doi.org/10.4103/1735-9066.185582
        • Şentürk A.
        • Tekinsoy Kartın P.
        The effect of lavender oil application via inhalation pathway on hemodialysis patients’ anxiety level and sleep quality.
        Holist Nurs Pract. 2018; 32: 324-335https://doi.org/10.1097/HNP.0000000000000292
        • Yayla E.M.
        • Ozdemir L.
        Effect of inhalation aromatherapy on procedural pain and anxiety after needle insertion into an implantable central venous port catheter: a quasi-randomized controlled pilot study.
        Cancer Nurs. 2019; 42: 35-41https://doi.org/10.1097/NCC.0000000000000551
        • Zabirunnisa M.
        • Gadagi J.S.
        • Gadde P.
        • Myla N.
        • Koneru J.
        • Thatimatla C.
        Dental patient anxiety: possible deal with Lavender fragrance.
        J Res Pharm Pract. 2014; 3: 100-103https://doi.org/10.4103/2279-042X.141116
        • Cohen J.
        Statistical power analysis for the behavioral sciences.
        2nd ed. Lawrence Erlbaum Associates, New York1988: 24-27