Systematic Review for the Medical Applications of Meditation in Randomized Controlled Trials

Background: Meditation has been increasingly adapted for healthy populations and participants with diseases. Its beneficial effects are still challenging to determine due to the heterogeneity and methodological obstacles regarding medical applications. This study aimed to integrate the features of therapeutic meditation in randomized controlled trials (RCTs). Methods: We conducted a systematic review of RCTs with meditation for populations with diseases using the PubMed database through June 2021. We analyzed the characteristics of the diseases/disorders, participants, measurements, and their overall benefits. Results: Among a total of 4855 references, 104 RCTs were determined and mainly applied mindfulness-based (51 RCTs), yoga-based (32 RCTs), and transcendental meditation (14 RCTs) to 10,139 patient-participants. These RCTs were conducted for participants with a total of 45 kinds of disorders; the most frequent being cancer, followed by musculoskeletal and connective tissue diseases and affective mood disorder. Seven symptoms or signs were frequently assessed: depressive mood, feeling anxious, quality of life, stress, sleep, pain, and fatigue. The RCTs showed a higher ratio of positive outcomes for sleep (73.9%) and fatigue (68.4%). Conclusions: This systematic review produced the comprehensive features of RCTs for therapeutic meditation. These results will help physicians and researchers further study clinical adaptations in the future as reference data.


Introduction
Meditation is a mental practice aiming to improve the psychological capacity of selfregulation regarding attention, awareness, and emotion [1]. Its implementation has been described not only in religious and cultural beliefs but also in health promotion as itself or as a component of mind-body practices such as yoga, qigong, tai chi, and mindfulnessbased interventions (MBIs) [2]. The US National Center for Complementary and Integrative Health (NCCIH) reported the health benefits of meditation, including the regulation of blood pressure and blood glucose and stress reduction for healthy participants [3]. From 2012 to 2017, the use of meditation more than tripled from 4.1% to 14.2% in the general US adult population [4].
Additionally, meditation for patients in clinics against diverse disease conditions, such as complementary and alternative medicine (CAM), has been demanded [5]. Therapeutic meditation techniques include MBIs, yoga-based programs, and transcendental meditation 2 of 22 (TM) [2]. Mindfulness-based stress reduction (MBSR) and mindfulness-based cognitive therapy (MBCT), which are representative MBIs, have presented psychological benefits for patients with depression and cancers [6,7]. Yogic meditation and TM also produce preventive benefits regarding risk factors for cardiovascular diseases and affective disorders by regulating brain activity and the autonomic nervous system [8,9].
However, it is difficult to objectify the effect of meditation because of its methodological obstacles, such as diverse task forms, a wide range of traditional roots, and the proficiency of practicians [10]. The NCCIH pointed out that scientific research on meditation practices has a heterogeneous theoretical perspective and poor study design quality [3]. Most of the 400 clinical studies for meditation conducted before 2005 were considered to be poorly designed [11], while more recent meditation-derived clinical studies include many well-designed trials [12,13]. Physicians need to consider the characteristics of meditative intervention and its expected effect for application in their treatment plans. To date, no systematic analysis showing the current status of clinical applications for patients exists.
To facilitate these works, this study comprehensively reviewed the features of randomized controlled trials (RCTs) with meditative interventions designed for patients to date regarding patients, interventions, controls, measurements, and overall results.

Database and Searching Strategy
The literature survey of this review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [14] using the PubMed electronic database through 10 June 2021. The search term used was 'meditation [Title/Abstract]' with excluding filters [Systematic review, Review, Meta-analysis]. The literature survey was conducted by 6 individual reviewers, and the third one checked the process.

Eligibility Criteria
The literature screening process of this study used the following inclusion criteria: (1) studies that were RCTs or randomized controlled crossover trials; (2) the study participants were patients with diseases; and (3) studies that evaluated the therapeutic efficacy of meditation. The exclusion criteria were as follows: (1) articles with no full text; (2) pilot RCTs or post-analysis studies; (3) studies with a Jadad score less than 3 points; and (4) studies with results published in languages other than English.

Data Extraction and Quality Assessment
Data extraction dealt with the number of participants, mean age, subjects' diseases, intervention category, treatment period, control, and outcome measurements. We also obtained the result data of the original article with a statistical analysis of the treatment effect compared to the control.
To assess the quality of the RCTs, the Jadad scale was used [15]. The Jadad scale is a five-point checklist scale that allocates points for descriptions of randomization, doubleblinding, withdrawals, and drop-outs. Trials with ≥3 points were considered high quality and were included in this study.

Categorization of Meditation
We categorized meditative interventions into 3 groups (Mindfulness-and yoga-based meditation and TM). Mindfulness-based or yoga-based meditation included interventions with the terms "mindfulness" or "yoga" used in each original article. The TM group included TM, mantras, and spiritual meditation.

Judgment of the Statistical Efficiency of the Intervention
We judged the benefit of intervention with statistical significance based on the data presentations of the original articles. In general, 'statistical significance' meant that the intervention showed statistically significant improvement (intervention vs. co 0.05 or Cohen's d > 0.5) according to the outcome measurement at the planned or the closest time to the end of the treatment period. We defined 'benefit with significance' for the following cases: (1) One or more of the outcomes regard symptoms or signs were statistically significant, or (2) statistical significance was only at the planned time point or the closest point to the period after treatme any description was noted.

Data Analysis
This systematic review did not need to apply statistical analysis. Regarding ber of participants, age, and treatment periods in the demographic features, th presented as the mean and standard deviation (SD) computed by Microsoft ware.

Items
Count N. of RCT 104  A This is the mean of the participants' ages presented as the median or mean in the original articles. B These items have been applied multiple times, thus the sum of percentages is larger than 100%. C These items shown in the table are the most frequently used in RCTs.

Diseases of the Participants and Types of Meditation in the RCTs
Among a total of 104 RCTs, 73 trials (70.2%) were performed for participants with physical diseases, and 31 (29.8%) were performed for participants with mental disorders (Tables 2 and 3). The physical diseases included cancer (16 RCTs), followed by musculoskeletal and connective tissue (14 RCTs), nervous system (9 RCTs), circulatory system and gynecological diseases (7 RCTs each). On the other hand, affective disorder (13 RCTs) and post-traumatic stress disorder (PTSD, 7 RCTs) were the most frequent mental diseases in the subjects (Table 1).       In terms of the interventions used in the RCTs, 34 kinds of meditations were employed. Ninety-three trials (89.4%, out of 104 total) employed at least one of the three major meditative interventions: mindfulness-based (51 RCTs), yoga-based (32 RCTs), or TM (14 RCTs) ( Table 1). Further detailed information for the diseases/disorders and types of meditations is shown in Table 2.

Target Measurements in the RCTs
From a total of 104 RCTs, the mean number of assessed outcomes per RCT was 3.9 ± 2.5, which presented 105 kinds of outcomes (Table 1). A total of 76 RCTs (73.1%) reported primary outcome measurements, which produced 99 primary outcomes for 41 kinds of measurements. Ten kinds of measurements were reported in at least three RCTs, including pain (11 RCTs), depressive mood (10 RCTs), feeling anxious (8 RCTs), quality of life (QoL, 7 RCTs), stress (7 RCTs), sleep (6 RCTs), PTSD symptoms (5 RCTs), blood pressure, intraocular pressure (IOP), and fatigue (3 RCTs each) (Figure 2). In terms of the interventions used in the RCTs, 34 kinds of meditations were employed. Ninety-three trials (89.4%, out of 104 total) employed at least one of the three major meditative interventions: mindfulness-based (51 RCTs), yoga-based (32 RCTs), or TM (14 RCTs) ( Table 1). Further detailed information for the diseases/disorders and types of meditations is shown in Table 2.
As shown in Figure 3, RCTs for sleep and fatigue showed 100% positive effects in RCTs with participants with mental disorders and showed a relatively high ratio of positive outcomes in all the RCTs (73.9% and 68.4%, respectively), followed by stress (48.4%), depressive mood (42.9%), pain (40.9%), QoL (40.6%), and feeling anxious (35.0%). The positive ratios of depressive mood and stress symptoms were greater than 50% only for mental and physical diseases (57.9 and 56.5% each) (Figure 3). In the sum of all measurements (406 measurements in 104 RCTs), depressive mood symptoms were most frequent (56 RCTs, 53.8%), followed by feeling anxious (
As shown in Figure 3, RCTs for sleep and fatigue showed 100% positive effects in RCTs with participants with mental disorders and showed a relatively high ratio of positive outcomes in all the RCTs (73.9% and 68.4%, respectively), followed by stress (48.4%), depressive mood (42.9%), pain (40.9%), QoL (40.6%), and feeling anxious (35.0%). The positive ratios of depressive mood and stress symptoms were greater than 50% only for mental and physical diseases (57.9 and 56.5% each) (Figure 3).

Clinical Outcomes According to the Type of Meditation
Among the three major meditative interventions, yoga-based interventions (53.7% of 67 outcomes) showed slightly higher positive outcomes than mindfulness-based interventions (45.3% of 113 outcomes) and TM interventions (31.0% of 29 outcomes). Compared to other symptoms, sleep-targeted RCTs most frequently employed mindfulness-based meditation (16 of 25 RCTs), which presented over half the ratio of positive results (75.0%). Yogabased interventions showed higher positive results regarding depressive mood (56.3%). RCTs for sleep and fatigue showed 100% of significant benefits by mindfulness-based meditation and TM (Figure 4).

Clinical Outcomes According to the Type of Meditation
Among the three major meditative interventions, yoga-based interventions (53.7% of 67 outcomes) showed slightly higher positive outcomes than mindfulness-based interventions (45.3% of 113 outcomes) and TM interventions (31.0% of 29 outcomes). Compared to other symptoms, sleep-targeted RCTs most frequently employed mindfulness-based meditation (16 of 25 RCTs), which presented over half the ratio of positive results (75.0%). Yogabased interventions showed higher positive results regarding depressive mood (56.3%). RCTs for sleep and fatigue showed 100% of significant benefits by mindfulness-based meditation and TM (Figure 4).

Figure 4. The features of significant outcomes in the RCTs according to the types of meditations.
Regarding all measurements, the seven most frequently assessed outcomes are displayed according to the types of meditations. The circle's size and level of darkness indicate the number of RCTs and the ratio of positive outcomes, respectively. The positive outcome indicates that treatment achieved statistical significance compared to the control (p < 0.05 or Cohen's d > 0.5). */# indicates the total number of outcomes measured for the seven symptoms in the RCTs for patients with given disorders and positive outcomes, and their ratio.

Discussion
Research on the physiological effects of meditation was first conducted in the 1950s, and a clinical study was initiated in the 1970s [118]. While CAM is becoming increasingly popular, so does the employment of meditation for disease therapy currently representing a quarter of the total CAM uses [119]. Although numerous studies have been conducted to investigate the clinical benefits of meditation, the standardization of its benefits is still challenging because of the heterogeneity and methodological weaknesses [11]. To provide fundamental information for future clinical use and study, this review aimed to produce features of clinical applications with meditation using RCTs for especially diseased populations to date.
The trials included in this study gradually increased from 3 RCTs over a 5 year period from 2000 to 2004 to 52 RCTs in the period from 2015 to 2019 (data not shown). The sex ratio of participants showed an approximately twofold higher distribution in females (7022 out of a total of 10,139), generally correlating with sex differences in the prevalence of meditation use [120]. The 104 RCTs included in this study involved 45 kinds of disease conditions, and 43 RCTs assessed the primary outcomes directly related to the target diseases/disorders, such as pain for arthritis [33,44], sleep quality for sleep disturbance [114], and the PTSD-symptom scale for people living with PTSD [13,[101][102][103]. Several RCTs evaluated objective changes by assessing blood markers, heart rate variability [12,32,59,61,72,74,76,77,92,93,103], or brain imaging scans [53], as summarized in Tables 2  and 3. The others mainly focused on disease-related comorbid symptoms, including depressive mood, feeling anxious, QoL, and stress ( Figure 2). The top 7 most frequently assessed clinical outcomes included depressive mood, feeling anxious, QoL, stress, sleep, pain, and fatigue, and at least 1 of them was measured in 87.5% (91 RCTs) of the 104 RCTs. These symptoms have a high prevalence in both diseased

Discussion
Research on the physiological effects of meditation was first conducted in the 1950s, and a clinical study was initiated in the 1970s [118]. While CAM is becoming increasingly popular, so does the employment of meditation for disease therapy currently representing a quarter of the total CAM uses [119]. Although numerous studies have been conducted to investigate the clinical benefits of meditation, the standardization of its benefits is still challenging because of the heterogeneity and methodological weaknesses [11]. To provide fundamental information for future clinical use and study, this review aimed to produce features of clinical applications with meditation using RCTs for especially diseased populations to date.
The trials included in this study gradually increased from 3 RCTs over a 5 year period from 2000 to 2004 to 52 RCTs in the period from 2015 to 2019 (data not shown). The sex ratio of participants showed an approximately twofold higher distribution in females (7022 out of a total of 10,139), generally correlating with sex differences in the prevalence of meditation use [120]. The 104 RCTs included in this study involved 45 kinds of disease conditions, and 43 RCTs assessed the primary outcomes directly related to the target diseases/disorders, such as pain for arthritis [33,44], sleep quality for sleep disturbance [114], and the PTSD-symptom scale for people living with PTSD [13,[101][102][103]. Several RCTs evaluated objective changes by assessing blood markers, heart rate variability [12,32,59,61,72,74,76,77,92,93,103], or brain imaging scans [53], as summarized in Tables 2 and 3. The others mainly focused on disease-related comorbid symptoms, including depressive mood, feeling anxious, QoL, and stress ( Figure 2). The top 7 most frequently assessed clinical outcomes included depressive mood, feeling anxious, QoL, stress, sleep, pain, and fatigue, and at least 1 of them was measured in 87.5% (91 RCTs) of the 104 RCTs. These symptoms have a high prevalence in both diseased and healthy populations. For example, US survey data presented prevalence rates of 21%, 25%, and 33% for depressive mood, fatigue, and pain, respectively, in subjects with a cancer history and 18%, 18%, and 29% in adults without a cancer history, respectively [121]. Patients usually complained of clusters of several symptoms simultaneously; accordingly, most RCTs had multiple outcomes, with an average of 3.9 ± 2.5 measurements (Table 1). Mind-body interventions such as relaxation, cognitive behavioral therapy, coping skills training, or meditation helped patients manage symptom clusters of pain-fatigue-sleep disturbance in cancer patients [122].
Meditative interventions are usually expected to promote strengths in tolerating individual torments by reinforcing the psychological capacity of self-regulation [123]. In our data, meditations most frequently targeted depressive mood and feeling anxious, which were known as the most common comorbid symptoms in the population under disease conditions [124,125]. Both complaints were significantly improved by meditation practices by 42.9% and 35.0%, respectively, and the anti-depressive effect was more prominent in patients with mental disorders (57.9%) than in those with physical disorders (35.1%) ( Figure 3). The therapeutic mechanisms of meditation against major depressive disorders are understood as upregulation of serotonin and dopamine along with neuro-immunoendocrinological modulation [126]. Sleep quality and fatigue-related benefits were the most positive in the current study by 73.9% and 68.4%, respectively, which increased further in subjects with mental disorders (100.0% both) (Figure 3). In disease conditions, sleep-requiring symptoms such as fatigue and lethargy are common complaints as pathophysiological defense responses [127]. The role of meditation in sleep quality is understood as changing sleep architecture by regulating brain activities [128]. Accumulated evidence indicates that meditative interventions could lead to functional and structural changes within the brain, such as the prefrontal region, cingulate cortex, striatum, and amygdala, suggesting its promising therapeutic effect on mental disorders [129].
In terms of QoL and stress symptoms, they are the main reasons for the use of meditation by meditation practitioners in the US [120]. The benefits of mindfulness-and yoga-based programs for stress reduction and global wellness in nonclinical populations are widely and well-known [130,131]. However, our data suggest that meditative interventions did not consistently improve QoL or stress symptoms in disease conditions. Perhaps disease-related QoL and stress differ from those of healthy conditions due to the intensity of root matters such as histological and functional impairments [132]. Several RCTs on gynecological disease [62,64,67] and PTSD [103,104], however, showed fragmentary evidence to support the stress reduction effect without ineffective outcomes (Tables 2 and 3).
In general, RCTs for physical disease included more than two mental disorders (74 physical vs. 31 mental diseases). Among them, cancer, musculoskeletal or connective tissue, and nervous system diseases were dominant (16, 14, and 9 RCTs, respectively) ( Table 1). For patients with cancer, meditation is one of the preferred CAMs with the highest satisfaction [133]. In our data, the interventions showed a decent ratio of positive results (positive, 8, vs. not significant, 2) in terms of cancer-related fatigue (CRF) ( Table 2). Regarding the pain relief effect, the most of RCTs evaluating pain were for diseases of the musculoskeletal, connective tissue, and nervous system (13 RCTs out of 22). Although they failed to present consistent benefits (7 positive vs. 6 not significant), this ratio was higher than that in other diseases (2 vs. 6) (Figure 3, Tables 2 and 3).
Overall, according to our data, meditative therapies appear to be more beneficial for improving sleep quality and fatigue management as well as anti-depressive effects, especially in patients suffering from mental disorders. The conceptual basis of meditations includes attention, awareness, point of focus, or self-transcendence, even though they vary depending on historical and regional beliefs and methodological heterogenicity [134]. When we compared the gross features of three major meditation classifications, approx-imately 50% of the RCTs (51 RCTs) employed mindfulness-based meditation (Table 1). Compared to other symptoms, sleep-targeted RCTs (64%) mainly employed mindfulnessbased meditation, resulting in relatively high positive results (75.0%). The second most frequently applied type of intervention was yoga-based interventions, which were more beneficial for depressive mood (56.3%) than other interventions (Figure 4). In general, physical exercise therapies are recommended for patients suffering from most diseases, including cancer, fibromyalgia, and multiple sclerosis [135][136][137]. However, they could often worsen certain symptoms, such as fatigue, or occasionally cause other physical adverse effects [138]. Some studies have shown that managing concurrent symptoms well has a medical impact, including on the final clinical outcome; for example, reducing fatigue severity affects the survival or recurrence rate in breast cancer patients [139]. In this regard, meditation could be a promising alternative to manage subjective complaints in people living with various diseases. This review has several limitations. First, a single database (PubMed) was used for the literature survey. Although the majority of data were extracted from the PubMed database, more evidence could be provided by other databases. We also found that the generally poor design quality of the studies remains a weakness. In this review, the number of articles were reduced by two-thirds in the screening process with Jadad scores (Figure 1). Further rigorously designed clinical trials are needed to objectify the clinical benefits of meditation. To provide confident information, pilot RCTs were excluded from this review. They commonly enrolled a small number of participants, focusing mainly on the feasibility of further trials. However, this tactic also has a risk of losing any valuable data.
Despite the limitations mentioned above, this review produced the comprehensive features regarding patients, interventions, controls, measurements, and their overall benefits in diseased populations. These results provide fundamental information to patients, practitioners, and researchers as reference data for future studies.

Conflicts of Interest:
The authors declare no conflict of interest.