Therapeutic effects of Silexan on somatic symptoms and physical health in patients with anxiety disorders: A meta‐analysis

Abstract A meta‐analysis was performed to examine therapeutic effects of Silexan on somatic symptoms, including insomnia/fatigue, and physical health in patients with anxiety disorders. Five randomized, placebo‐controlled trials were included in this analysis: The efficacy of Silexan (80 mg/day) was investigated in patients with subthreshold anxiety disorders (three trials) and in patients with generalized anxiety disorder (two trials). Silexan was superior to placebo in terms of the mean change from baseline in the Hamilton Anxiety Rating Scale (HAMA) subscore somatic anxiety at week 10 with a standardized mean difference of −0.31 [95% Cl: −0.52 to −0.10, p = .004]. Treatment effects of silexan on somatic anxiety were independent of gender and age. Statistically significant differences were also shown for single HAMA items somatic muscular, cardiovascular, respiratory, and genitourinary symptoms, indicating clinical relevance with small to medium effects of Silexan. Similar clinically meaningful effects of Silexan on SF‐36 physical health, including reduced bodily pain and improved general health, and on insomnia complaints and fatigue, were demonstrated. In this meta‐analysis including all placebo‐controlled clinical trials in patients with anxiety disorders to date, statistically significant and clinically meaningful advantages of Silexan over placebo treatment were found in improving somatic symptoms and physical health.

impairment that do not meet the required number of symptoms of threshold definitions) are highly prevalent and impairing conditions among both adolescents and adults (Baldwin et al., 2011;Burstein et al., 2014;Haller et al., 2014;Siegel & Dickstein, 2011). In a primary healthcare setting, the prevalence of GAD in adult patients has been reported as 4.1%-6.0% among men, and 3.7%-7.1% among women (Munk-Jorgensen et al., 2006). Prevalence data from a community mental health survey reported that threshold and subthreshold anxiety affect similar percentages of adolescents aged 15 years or older (2.6% versus. 2.3%, respectively) (Gilmour, 2016).
GAD and subthreshold anxiety share the same common elements of disproportionate and debilitating worry characterizing all anxiety disorders (Burstein et al., 2014). According to the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5), somatic symptoms belong to the definition criteria of GAD, including restlessness, increased fatigability, muscle tension, and sleep disturbance (American Psychiatric Association (2013)). These symptoms may be accompanied by autonomic hyperarousal, that is, sweating, dizziness, and shortness of breath (Gelenberg, 2000). If untreated, anxiety symptoms persist and are associated with significant impairments in daily functioning, poor quality of life, and a huge economic burden owing to lost work productivity and high healthcare utilization costs (Greenberg et al., 1999;Hoffman et al., 2008).
Typically, patients with GAD present with mostly somatic complaints (Crawford et al., 1994) that vary between individuals. Somatic symptoms tend to be prolonged in patients with anxiety disorders.
However, many anti-anxiety drugs such as benzodiazepines, other gamma-aminobutyric acid (GABA)-facilitatory drugs, for example, clomethiazole, azapirones, for example, buspirone, and antidepressants are contraindicated for treatment because of side effects and/ or the risk of dependence associated with chronic use (Morgan & Tyrer, 1994).
The anxiolytic efficacy of the orally administered lavender oil preparation Silexan has been investigated in GAD and other anxiety disorders (Kasper et al., 2010(Kasper et al., , 2016Kasper, Gastpar, Muller, et al., ).
In a meta-analysis of three randomized, placebo-controlled trials (Kasper et al., 2010(Kasper et al., , 2016, Silexan demonstrated superiority to placebo in reducing somatic symptoms and improving physical health in patients with subthreshold anxiety (Möller et al., 2019).
Secondary to its anxiolytic effect, Silexan had a beneficial effect on sleep, often accompanied by fatigue, and it also improved patients' health-related quality of life (HRQoL) (Möller et al., 2019). Five placebo-controlled clinical trials investigating the efficacy of Silexan (80 mg/day) in patients with anxiety disorders have been performed to date (Kasper et al., 2010(Kasper et al., , 2016(Kasper et al., , 2017Kasper, Gastpar, Müller, et al., 2014). We conducted a meta-analysis of all five clinical trials to further elucidate the therapeutic effects of Silexan with respect to somatic complaints in patients with anxiety. Besides including two additional studies, the novelty of the current versus a previous meta-analysis (Möller et al., 2019) is a detailed investigation of individual anxiety-related somatic symptoms, including pain, insomnia complaints along with fatigue, and physical health (Eliasen et al., 2016).

| Clinical trials included
In the meta-analysis, placebo-controlled clinical trials investigating the efficacy of Silexan in patients with anxiety disorders were included. These trials are Kasper et al., 2010 (A), Kasper et al., 2015 (B), Kasper et al., 2017 (C), Kasper, Gastpar, Muller, et al., (D) and Kasper et al., 2016 (E). Since all trials included a group of patients receiving a dosage of 80 mg/day Silexan, and no other dosage was investigated in more than one trial, the efficacy of 80 mg/day Silexan was compared to placebo in the meta-analysis. Individual patient data are used for the meta-analysis. To identify any additional randomized placebo-controlled clinical trials conducted with 80 mg Silexan in patients with anxiety disorders, one author performed literature searches by using the PubMed database. Search terms included "Silexan" in combination with "anxiety disorder" before 30 April 2019 (no further restrictions applied). The searches did not reveal any additional randomized placebo-controlled clinical trial in patients with anxiety disorders.

| Bias assessment
Bias assessment on the study level was performed using the Cochrane Collaboration's tool for assessing risk of bias (Higgins et al., 2011).
Assessments were based upon the applicable publications, the patient raw data, and on the original protocols, and the full integrated study reports made available to the authors and to the assessor.

| Outcome measures
The analysis of treatment effects was based on the full analysis sets  (Brown et al., 2011).
These subscales have been widely used in GAD research, and the symptoms within these subscale domains have previously been shown to be differentially sensitive to various treatment effects (Feighner & Cohn, 1989;Meoni et al., 2001;Pollack et al., 2001;Rickels et al., 1993;Rickelset al., 1982). In all studies, assessments were performed prior to treatment (baseline) and at week 10.
Treatment effects related to somatic symptoms by gender and age (<60 years and ≥ 60 years) were also analyzed.

| Statistical methods
The meta-analysis was planned after all trials had been conducted, and the methods including the choice of the outcomes were specified before the meta-analysis was performed. Descriptive statistics (sample size, means and standard deviation (SD) for continuous data) were Treatment effects on somatic anxiety quantified by the change of the HAMA somatic anxiety subscore after 10 weeks of treatment were investigated separately for women and men. Treatment effects were compared within the subgroups following the strategy described above. The efficacy of Silexan 80 mg/day compared to placebo was investigated using mixed effects models assuming random effects within subgroups and a fixed effect across subgroups. The variance between trials was estimated separately for each subgroup.
The p-value of the Q test for heterogeneity was calculated (method 3) (Borenstein M et al., 2009) to compare results between the subgroups.
Results of the meta-analyses are presented using forest plots (R, package meta; version 4.3-2, function forest). The p-values of twosided tests are presented, a p-value < 0.05 is considered statistically significant. included. From trial D, which compared three Silexan doses (10 mg, 40 mg, 80 mg) to placebo, and from the reference and placebo-controlled trial E, only data of patients randomized to receive 80 mg/ day Silexan or placebo were used in this meta-analysis. In all five trials, the patients were treated for 10 weeks. Characteristics of the trial populations are shown in Table 1. Across all 5 studies, a total of 1,213 patients were randomized to receive either 80 mg/day Silexan or placebo. Of these, 1,172 patients could be analyzed for efficacy (FAS). Approximately 70% of the patients were women, and the mean age was approximately 46 years (Table 1).

| Baseline values of analyzed outcomes
Baseline values of the HAMA subscore somatic anxiety, HAMA item 7-13 describing somatic symptoms, HAMA item 4 (insomnia), and the SF-36 subscore physical health, and the SF-36 components related to physical health and vitality (as a measure of fatigue) are shown in the appendix (Table A1 and Table A2).

| Treatment effects with respect to somatic complaints
In 4 of the 5 trials analyzed, somatic anxiety improved more in patients treated with Silexan compared to the placebo groups; in 1 trial (trial D), there was no difference between treatment groups (Figure 1a).
In the meta-analyses, statistically significant and clinically meaningful differences between Silexan 80 mg/day and placebo were shown for the HAMA subscore somatic anxiety (p = .004) ( Figure 1a) and for the single HAMA items 7 (somatic muscular) (Figure 1b

| Treatment effects with respect to physical health
In trials A, C, D, and E, physical HRQoL was investigated using the SF-36 questionnaire. In all 4 single trials, physical health improved

| Treatment effects with respect to insomnia complaints and fatigue
In all 5 single trials, insomnia (HAMA item 4) improved more in patients treated with Silexan than in the placebo groups. The metaanalysis revealed a statistically significant and clinically meaningful result in favor of Silexan with an effect size of 0.30 (Figure 3a). A similar clinically meaningful difference between Silexan and placebo with a small to medium effect was observed for improved fatigue ( Figure 3b).

| Treatment effects related to somatic symptoms by gender
In 4 clinical trials (A, B, C, E), treatment effects of Silexan were more pronounced in women as well as in men treated with Silexan compared to women and men of the placebo groups (data not shown). In the bigger subgroup of women, the treatment effect with respect to changes of the HAMA subscore somatic anxiety was statistically meaningful while the difference observed for the smaller subgroup of men showed a similar extent but was not statistically meaningful.
This reflects limited statistical power in the smaller subgroup of male patients included, as treatment effects were not significantly different between men and women (p = .87) (data not shown).

| Treatment effects related to somatic symptoms by age
In total, 998 patients aged < 60 years (502 treated with Silexan of Silexan were more pronounced in patients of both age groups treated with Silexan compared to both age groups in the placebo groups (subgroup difference p = .4764; data not shown). The treatment effect with respect to changes of the HAMA subscore somatic anxiety was statistically meaningful for patients in both age groups (effect size 0.32; p < .0001).

| D ISCUSS I ON
The therapeutic effects of Silexan treatment on somatic symptoms, including insomnia complaints and fatigue, and on reduced physical health in patients with anxiety disorders were investigated.
Our results demonstrate that Silexan is superior to placebo in reducing somatic symptoms in patients suffering from at least mild-to-moderate subthreshold anxiety disorders or GAD. A statistically significant and clinically meaningful difference in favor of Silexan was shown for the HAMA subscore somatic anxiety.
Notably, cardiovascular and genitourinary symptom improvements contributed most to the overall effect of Silexan in reducing somatic anxiety. Among patients with cardiovascular disease, anxiety disorders are common and associated with poor cardiovascular health, including the development and progression of coronary artery disease and heart failure (Celano et al., 2016;Roest et al., 2010). In patients with stable coronary artery disease, two- year follow-up data revealed that patients with comorbid GAD had a twofold increased risk of major adverse cardiac events compared to those without GAD (Frasure-Smith & Lesperance, 2008).
The mechanisms mediating the underlying association between anxiety disorders and cardiac disease are poorly understood, although behavioral and physiologic factors have been proposed (Celano et al., 2016). Another important public health concern is the fact that anxiety disorders are common in patients with noncardiac chest pain (Celano et al., 2016;Ortiz-Garrido et al., 2015).
Such patients tend to have similar levels of anxiety and low QoL as patients diagnosed with chest pain of cardiac origin (Webster et al., 2012). Patients with cardiovascular disease are likely to be taking other drugs, notably antihypertensives, lipid-lowering drugs and antiarrhythmic drugs; hence, prescribing anti-anxiety agents such as SSRIs is of some concern because of possible pharmacokinetic interactions (Davies et al., 2004). We have shown that Silexan exerts a clinically relevant small to moderate effect on cardiovascular symptoms and therefore is a well-tolerated alternative to treatment with SSRIs in GAD.
Evidence from the literature suggests that anxiety and genitourinary symptoms might share biological pathways, for example, the serotonergic pathway. Indeed, the SNRI duloxetine has been shown to improve symptoms of overactive bladder in women (Steers et al., 2007). Interestingly, in our study, treatment for 10 weeks with Silexan also improved genitourinary symptoms in patients with anxiety.
Impairment in HRQoL has been noted in patients with anxiety disorders (Beard et al., 2010). Our results show that Silexan is efficacious in improving physical HRQoL in patients with anxiety disorder.
The greatest improvements were observed for the SF-36 physical health components role-physical and bodily pain. The latter finding is of particular interest as anxiety shares the same pathophysiological pathways as pain with mutual effects on each other (de Heer et al., 2014;Means-Christensen et al., 2008). Both anxiety disorders and pain facilitate the central modulation of the pain response at multiple sites in the brain (e.g., periaqueductal gray, amygdala, and hypothalamus) (Ossipov et al., 2010). They also share underlying cognitive and behavioral processes, such as increased attention toward threat and anxious avoidance of physical exertion (Asmundson & Katz, 2009;Sareen et al., 2005). Furthermore, anxiety-induced stress increases the production of pro-inflammatory cytokines (Hou et al., 2017), which may also increase pain (de Oliveira et al., 2011).
In a study by de Heer et al. (2014), anxiety symptom severity was associated with more disabling and severely limiting pain (de Heer et al., 2014). This study also showed that anxiety disorders have a similar and strong association with musculoskeletal pain, cardio-respiratory pain, and gastrointestinal pain compared to a control group without depressive or anxiety disorder (de Heer et al., 2014).
Silexan does not share the same mechanism of action as other antidepressant or anxiolytic drugs (e.g., benzodiazepines). This may explain why particular side effects typical of benzodiazepines, such as sedation, are absent (Schuwald et al., 2013). The SNRIs duloxetine and venlafaxine are used to achieve reduction in chronic pain associated with depression (Jann & Slade, 2007). It is unclear whether the mechanisms through which Silexan treatment improves somatic symptoms and physical health in anxiety are similar to those attributed to dual-action antidepressants. However, a study by Baldinger et al. (2014) reported that effects on the serotonin-1A receptor may contribute to the therapeutic action of Silexan (Baldinger et al., 2014). Alternatively, it is plausible that Silexan reduces anxious arousal thus decreasing patient sensitivity to anxiety and perception of somatic symptoms, including pain. Indeed, some individuals with elevated anxiety sensitivity may be more likely to perceive somatic sensations associated with anxiety as dangerous, which can lead to a perpetuated cycle of increased perception to and misinterpretation of bodily cues (Reiss, 1991), thereby increasing healthcare-seeking behavior.
Insomnia, commonly accompanied by fatigue, is highly prevalent in anxiety disorders and may also contribute to the maintenance and/or exacerbation of anxiety through its impact on anticipatory brain function (Cox & Olatunji, 2016). Insomnia (according to the HAMA insomnia item) has been reported to be the most frequent complaint among patients presenting with GAD in primary care, with an incidence of 32.5% (Wittchen et al., 2002).
In our study, there was an overall improvement in insomnia and fatigue in patients treated with Silexan; this is a secondary effect resulting from anxiolysis (Seifritz et al., 2019). Better sleep may also lead to reduced hyperarousal and perception of pain and/or somatic symptoms associated with anxiety as well as a decrease in fatigue.
In general, women report more bodily distress in terms of more numerous, more intense, and more frequent somatic symptoms than men (Barsky et al., 2001). The fact that the trials analyzed in our meta-analysis recruited significantly more women than men supports this finding. There was no significant difference in the effect of Silexan treatment between male and female patients included in our analysis, suggesting that Silexan is of clinical benefit for somatic anxiety irrespective of gender.
In addition to gender, other socio-demographic factors such as age have been identified as correlates of insomnia (Ohayon, 2002;Wennberg et al., 2013). From our data the effects of Silexan on somatic complaints are found independent of age, that is, age did not influence the variance of the results. Evidence from two studies with Silexan in subthreshold anxiety patients with a wide range of ages, including elderly patients > 65 years, support our findings (Kasper, 2015;Seifritz et al., 2019), and however, further randomized studies will be required to confirm this.
Another limitation of this meta-analysis is that the studies found and included were published by the same research group, which also overlaps with the authors of this work. Therefore, effect sizes could potentially be more similar than effect sizes from studies of different research groups. This is because the effect sizes might be influenced by, for example, the way the variables were analyzed, the subjects were sampled, and the observers or interviewers who collected the data (Cooper, 2009;Van den Noortgate et al., 2013). This is in-

| CON CLUS IONS
In conclusion, this meta-analysis in patients with anxiety reports statistically significant and clinically meaningful advantages of 80 mg Silexan compared to placebo in improving somatic symptoms, including pain and insomnia, and physical health. This study adds to the empirical evidence supporting a role for the lavender oil preparation Silexan in the treatment armamentarium for anxiety disorders, including those presenting with somatic symptoms.

E TH I C A L S TATEM ENT
There are no ethical concerns with regard to this meta-analysis as only data from other randomized trials were re-analyzed.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.1997.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Mean SD
A Silexan 80 mg