In this chapter, the main results of the systematic review are presented according to the five topics that were addressed. The results for topic II, the state of research for the therapeutic use of meditation practices, contain all eligible studies. Studies were then selected from this larger set to address topics III to V (see chapter two on Methods).

Topic I. The Practice of Meditation

Main Components

The main components of any meditation practice or technique refer to its most general features. These may include specific postures (including the position of the eyes and tongue), the use of a mantra, breathing, a focus of attention, and an accompanying belief system. Posture refers to the position of the body assumed for the purpose of meditation. Though traditional meditation practices prescribe particular postures (e.g., the lotus position), postures vary between practices with the only limitation being that the posture does not encourage sleep.64 Because accounts of most meditation practices describe explicitly the use and role of breathing, mantra, attention, spirituality and belief, training, and criteria for successful meditation practice, these topics are described individually.

Breathing. Breathing in meditation can be incorporated passively or actively. In passive breathing, no conscious control is exerted over inhalation and exhalation and breathing is “natural.” In contrast, active breathing involves the conscious control over inhalation and exhalation. This may involve controlling the way in which air is drawn in (e.g., through the mouth or nostrils), the rate (e.g., drawn in quickly or over a specified length of time), the depth (e.g., shallow or deep), and the control of other body parts (e.g., relaxation of the abdomen).

Mantra. A distinctive feature of some meditation practices is the use of a mantra. A mantra is a sound, word, or phrase that is recited repetitively, usually in an unvarying tone, and used as an object of concentration. The mantra may be chanted aloud, or recited silently. Mantras can be associated with particular historical or archetypal figures from spiritual or religious systems, or they may have no such associations.65

Relaxation. Relaxation is often considered to be one of the defining characteristics of meditation practices and meditation itself is often considered to be a relaxation technique.6668 Indeed, it has been suggested that the popularity of meditation practices in the West is due, at least in part, to the widely accepted plausibility of their alleged effects with respect to arousal reduction.69 Some researchers have attempted to draw a distinction between relaxation and meditation practices on the basis of intention.70

Attention and its object. The intentional self-regulation of attention is considered crucial to the practice of meditation, as is the development of an awareness in which thoughts do not necessarily disappear, but are simply not encouraged by dwelling on them, a state of so-called “thoughtless awareness.”71,72 Some meditation practices focus attention on a singular external object (e.g., mandala, candle, flame), sound (e.g., breath), word or phrase (i.e., mantra), or body part (e.g., the tip of the nose, the space between the eyebrows).71 In contrast, “mindfulness” meditation techniques aim to cultivate an objective openness to whatever comes into awareness (e.g., by paying attention to simple and repetitive activities or to the sensations of the body). In doing so, the breath may be used as an anchor (but not a focal point) to keep the meditator engaged with the present moment.65,73 Each of these techniques serves, in a different way, to discourage logical and conceptual thinking.65

Spirituality and belief. This component refers to the extent to which spirituality and belief systems are a part of meditation practices. Spirituality and belief systems are composed of metaphysical concepts and the rules or guidelines for behavior (e.g., devotional practices or interpersonal relations) that are based on these concepts.

Training. Training refers to the recommended frequency and duration of periods of practice, and how long a practitioner is expected to train before being considered proficient in a given technique.

Criteria of successful meditation practice. The criteria of successful meditation practice are understood both in terms of the successful practice of a specific technique (i.e., is the technique being practiced properly) and in terms of achieving the aim of the meditation practice (e.g., has practice led to reduced stress, calmness of mind, or spiritual enlightenment).

Five broad categories of meditation practices were identified in the scientific literature: mantra meditation (comprising Transcendental Meditation® [TM®], Relaxation Response [RR], and Clinically Stadardized Meditation [CSM]), mindfulness meditation (comprising Vipassana, Zen Buddhist meditation, Mindfulness-based Stress Reduction [MBSR], and Mindfulness-based Cognitive Therapy [MBCT]), Yoga, Tai Chi, and Qi Gong. These broad categories were used for descriptive purposes throughout the report to address the key research questions.

Mantra Meditation

The distinctive characteristic of the meditation practices included in this category is the use of a mantra. A mantra is a word or phrase repeated aloud or silently and used to focus attention. A mantra often has a smooth sound, for example, the mantras “Om” or “Mu.”74 It is thought that these sounds produce vibrations that have different effects on people, and these vibrations can be described qualitatively or quantitatively.62,75 The three mantra meditation practices described below consist of standardized techniques; that is, the techniques have been described systematically in manuals and are relatively invariant wherever, whenever, and by whomever they are taught.23

Transcendental Meditation®

TM® is a technique derived from the Vedic tradition of India by Maharishi Mahesh Yogi.76 In TM®, a meditative state is purportedly achieved in which the repetition of the mantra no longer consciously occurs and instead the mind is quiet and without thought.77 During the practice of TM®, the ordinary thinking process is said to be “transcended” (or gone beyond) as the awareness gradually settles down and is eventually freed of all mental content, remaining silently awake within itself, and producing a psychophysiological state of “restful alertness.”78,79 These periods, referred to as pure consciousness or transcendental consciousness, are said to be characterized by the experience of perfect stillness, rest, stability, order, and by a complete absence of mental boundaries.80

Main components. In the TM® technique, the meditation state is achieved by the repetition of a mantra. The mantra is a meaningless sound from the ancient Vedic tradition and is given to the meditator by an instructor in the TM® technique.81,82 TM® practitioners sit in a comfortable posture, with eyes closed, and silently repeat the mantra.83 Though there are reports of the components of the mantras and how they are assigned, it is difficult to confirm these reports as many of the details of practice, including mantras, are revealed only to those who have formal instruction in TM®. Instruction in the TM® technique is a systematic, but individualized process. It is believed that keeping the techniques confidential prevents students from having preconceptions about the technique (making the learning process simpler) and that it maintains the integrity of the technique across generations.

Breathing. TM® involves passive breathing; no breath control procedures are employed and no specific pattern is prescribed.80

Attention and its object. TM® is described as not requiring any strenuous effort, concentration, or contemplation.80,84 However, meditators are instructed to direct their attention to the mantra.83

Spirituality and belief. The TM® technique has a theoretical framework that is described in Maharishi Mahesh Yogi's writings on the nature of transcendental consciousness and the principles underlying the TM® technique.81 However, it is unclear to what extent this theoretical framework, including any of its implications for spirituality, is a part of the practice. Sources that discuss this issue contend that the practice of the technique requires no changes in beliefs, philosophy, religion, or lifestyle,78,80,85 implying that the theoretical framework plays no role in its practice.

Training. TM® is usually taught in a course comprising five to six hours of instruction over four days.84 General information about the technique and its effects is presented in a 1.5-hour lecture. More specific information is given in a second 1-hour lecture. Those interested in learning the technique meet with the teacher for a 5- to 10-minute interview. The participant learns the technique on a separate day in a 1- to 1.5-hour session, following a short ceremony in which the mantra is given to the prospective practitioner. The next three sessions consist of 1.5-hour meetings held in the 3 days following, in which further aspects of the technique are explained. The teacher explains the practice of the technique in more detail, corrects practice if necessary, and explains practical arrangements (e.g., when to practice), the benefits of practice, and personal development through the technique. In addition, the technique is regularly checked by the teacher in the first months of practice to ensure correct practice, and the student is advised to continue with periodic checks thereafter.86,87

Clinical reports indicate that this technique can be learned easily by individuals of any age, level of education, occupation, or cultural background.78,80,85 The technique requires systematic instruction by a qualified teacher to ensure effortless and correct practice.78,84,86 The technique is practiced twice daily for 15 to 20 minutes, usually once in the morning (before breakfast) and once in the afternoon (before dinner).78,85,88

Criteria of successful meditation practice. The successful practice of the TM® technique is determined by a qualified teacher. As many details of the TM® technique are restricted to those who receive instruction, a description of the criteria used by the instructor for the assessment of the technique is not available in the scientific literature.

Relaxation Response

The “relaxation response” is a term coined by Harvard cardiologist Herbert Benson in the early 1970s to refer to the self-induced reduction in the activity of the sympathetic nervous system,68,89 the opposite of the hyperactivity of the nervous system associated with the fight-or-flight response. Benson believed that this response was not unique to TM® and that all ancient meditation practices involved common components that together are capable of producing such a response.68 Basing his belief on his scientific research on hypertension and TM®, he integrated these common factors into a single technique (RR) and found that it promoted a decrease in sympathetic nervous system similar to TM®.90 Many techniques for eliciting the relaxation response have been presented in a religious context in Judaism, Christianity, or Islamic mysticism (Sufism). These techniques employ both mental and physical methods, including the repetition of a word, sound, or phrase (often in the form of a prayer); and the adoption of a passive attitude.91 Benson emphasized that the relaxation response is not simply a state of relaxation (and should not be confused with it) or a sleep-like state, but a unique state brought about by adherence to specific instructions.89

Main components. The individual is instructed to assume a comfortable posture (usually sitting, but kneeling or squatting may also be used), the eyes are closed, and the muscles are relaxed, beginning at the feet and progressing upward to the face. Once the practitioner is relaxed, the eyes may be open or remain closed. Then, breathing through the nose and focusing on the breath, the practitioner inhales and exhales, silently saying the word “one” with each exhalation.89,90,92 Like TM®, the repetition of a sound, word, or phrase is considered essential to the technique.89 Benson recommends “one” as a neutral, one-syllable word.93 When the practice is completed, the meditator sits quietly for several minutes with eyes closed and then with eyes open.89

More recent versions of the technique include a body scan (similar to that employed in MBSR, described below) in which practitioners are asked to move their attention slowly over the body focusing on relaxing different regions, and information sessions on the stress response and its effects on health.94

Breathing. Breathing is active. Practitioners breathe through the nose, cultivating an easy, natural rhythm.89

Attention and its object. Attention is focused on the breath. In addition, should distracting thoughts occur, an attempt should be made to ignore them and focus on the mantra.92 The mantra is therefore “linked” with the breath.68 It has been claimed that Benson's RR demands a greater degree of concentration than either TM® or CSM (described below).64

Spirituality and belief. Because it is believed that RR incorporates the essential components of a wide variety of meditation practices, it is conceptualized as a secular technique89,95 and does not require adopting a specific spiritual orientation or belief system.

Training. RR is learned in approximately five minutes. Patients are typically instructed to elicit the relaxation response twice daily, for 15 to 20 minutes, but not within two hours after any meal, as the digestive processes may interfere with the subjective changes induced by the technique.89,90,96

Criteria of successful meditation practice. Instructions for this technique are available in books and articles and there is no explicit recommendation that an experienced practitioner teach the technique or that individualized instruction is necessary. The criteria for successful meditation practice rest with the subjective evaluation of the meditator; the results of practice judged against the reported effects of RR. Instructions for this technique include the injunction not to worry about whether one is successful in achieving a deep level of relaxation, and instead to maintain a positive attitude and let relaxation occur at its own pace.89

Clinically Standardized Meditation

CSM was developed by Patricia Carrington while she was conducting studies on meditation at Princeton University in the early-to-mid 1970s. Believing that TM® was not flexible enough to be suitable for all clinical purposes and that the cost of its instruction put it beyond the reach of most individuals and institutions, Carrington modified a classical Indian form of mantra meditation and produced what she called CSM.64

Main components. Trainees are instructed to choose a mantra from a list of 16 Sanskrit mantras, or choose their own. In choosing their own mantra, practitioners are told to select a word that has a “pleasant ringing sound” and to avoid using words that are emotionally loaded. The word should help imbue the practitioner with a sense of serenity.64 In its original formulation, CSM used a secular ritual for transferring the mantra. CSM is practiced while sitting comfortably, with eyes open and focused on a pleasant object of some kind. The mantra is repeated aloud, slowly and rhythmically, at ever decreasing volume, until it is a whisper, at which point the mantra is no longer said aloud, but instead is only thought. The eyes are then closed as the meditator continues repeating the mantra in thought. Meditators allow the mantra to proceed at its own pace, getting faster or slower, louder or softer “as it wants.”64,97

Breathing. Breathing is passive, proceeding at its own pace and is unconnected to the repetition of the mantra.

Attention and its object. Like TM® and RR, CSM is a passive technique that requires little concentration or discipline. In contrast to RR, CSM instructs practitioners to flow with their thoughts rather than ignore them, returning periodically to the mantra.64

Spirituality and belief. CSM is designed as a secular, clinical form of meditation practice, so no specific system of spirituality or belief is required.

Training. CSM is taught in two lessons: a 1-hour individual lesson and a group meeting. CSM is practiced twice daily for 20 minutes.64 As with RR, the contemporary version of CSM differs slightly from its original form, with perhaps the most important difference being that trainees are given a manual and an audio recording of instructions rather than individual instruction.64

Criteria of successful meditation practice. The criteria for successful meditation practice rest with the subjective evaluation of the meditator, the results of practice judged against the reported effects of CSM. Books and audiotapes for self-instruction in CSM are readily available, and there is no explicit statement that an experienced practitioner teach the technique or that individualized instruction is necessary.

Mindfulness Meditation

Mindfulness has been described as a process of bringing a certain quality of attention to moment-by-moment experience and as a combination of the self-regulation of attention with an attitude of curiosity, openness, and acceptance toward one's experiences.98 Mindfulness meditation, the core practice of Vipassana meditation, has been incorporated into several clinically-based meditation therapies.76 The capacity to evoke mindfulness is developed using various meditation techniques that originated in Buddhist spiritual practices;99 however, general descriptions of mindfulness vary from investigator to investigator and there is no consensus on the defining components or processes.98

Mindfulness approaches are not considered relaxation or mood management techniques,98 and once learned, may be cultivated during many kinds of activities. Mindfulness increases the chances that any activity one is engaged in will result in an expanded perspective and understanding of oneself.76 In a state of mindfulness, thoughts and feelings are observed on par with objects of sensory awareness, and without reacting to them in an automatic, habitual way.98,99 Thus, mindfulness allows a person to respond to situations reflectively rather than impulsively.98 Mindfulness meditation practices include the traditional Vipassana, and Zen meditation and the clinically-based techniques MBSR and MBCT. Of the four practices described below, the last two, MBSR and MBCT have standardized techniques (i.e., the techniques have been described systematically in manuals and are relatively invariant wherever, whenever, and by whomever they are taught).

Vipassana

Considered by some to be the form of meditation practiced by Gautama the Buddha more than 2,500 years ago,100 Vipassana, or insight meditation, is practiced primarily in south and southeast Asia but is also a popular form of meditation in Western countries. Vipassana is the oldest of the Buddhist meditation techniques that include Zen (Soto and Rinzai schools) and Tibetan Tantra.47,99 The Pali term “Vipassana”, though not directly translatable to English roughly means “looking into something with clarity and precision, seeing each component as distinct, and piercing all the way through so as to perceive the most fundamental reality of that thing.”47 The goal of Vipassana is the understanding of the 3 characteristics of nature which are impermanence (anicca), sufferings (dhuka), and non-existence (anatta). Vipassana meditation helps practitioners to become more highly attuned to their emotional states.47 Through the technique, meditators are trained to notice more and more of their flowing life experience, becoming sensitive and more receptive to their perceptions and thoughts without becoming caught up in them. Vipassana meditation teaches people how to scrutinize their perceptual processes, to watch thoughts arise, and to react with calm detachment and clarity, reducing compulsive reaction, and allowing one to act in a more deliberate way.47

Main components. Vipassana meditation requires the cultivation of a particular attitude or approach: (1) don't expect anything, (2) don't strain; (3) don't rush, (4) observe experience mindfully, that is, don't cling to or reject anything, (5) loosen up and relax, (6) accept all experiences that you have, (7) be gentle with yourself and accept who you are, (8) question everything, (9) view all problems as challenges, (10) avoid deliberation, and (11) focus on similarities rather than differences.47

Vipassana meditation is practiced in a seated position when focusing on the breath; otherwise, no posture is prescribed and the meditator may sit, stand, walk, or lie down. Traditionally, if a static position has been taken, it is not to be changed until the meditation session has ended. However, many Western teachers allow students to move, though mindfully, to avoid persistent pain caused by being in the same position for too long.47 The time devoted to seated meditation should be no longer than one can sit without excruciating pain. The eyes should be closed.47

Breathing. Air is inhaled and exhaled freely through the nose. There is a natural, brief pause after inhaling and again after exhaling.47

Attention and its object. The focus of attention or awareness in Vipassana can be categorized into 4 groups: body, emotions and feelings, thoughts, and mental processes.101 In focusing attention on the breath, novice Vipassana meditators attain a degree of “shallow concentration.”47 This is not the deep absorption or pure concentration of the mantra meditation techniques. Gradually, the focus of attention is shifted to the rims of the nostrils, to the feeling of the breath going in and out. When attention wanders from the breath, the meditator brings it back and anchors it there.47,100 To help concentrate on the breath, a novice meditator may silently count breaths or count between breaths.47 The meditator notices the feeling of inhaling and exhaling and ignores the details of the experience. The movement of the abdominal wall while inhaling and exhaling may also be used as a focus of attention.47

The primary technique for focusing on bodily sensations is the body scan.102 Beginning with the top of the head, the practitioner observes the sensations as if for the first time, and then scans the scalp, the back of the head, and the face. When visualizations of the body distract the meditator, the thoughts are simply directed back to the sensations. The focus of attention is moved continuously over the body, moving down the neck, to the shoulders, arms, hands, trunk, legs and feet. Throughout the entire scan, an attitude of nonanticipation and acceptance is maintained.102

Mindfulness can be practiced during any activity and practitioners are encouraged to practice being mindful and fully aware during other activities such as walking, stretching, and eating.100

Spirituality and belief. Though often described as a profound religious practice, no particular spiritual or philosophical system is required to practice Vipassana meditation.47

Training. Vipassana should be practiced twice daily, morning and evening, for about 5 to 10 minutes.100 Western interpreters of Vipassana have recommended that novice meditators should be instructed to sit motionless for no longer than 20 minutes.47 Ideally, a meditator works up to at least two 1-hour sessions per day, and does at least one 10-day retreat per year.102 Longer meditation sessions allow for deeper periods of meditation.102 The length of time required to become proficient in Vipassana meditation varies by individual, some students progress rapidly, others slowly.

Criteria of successful meditation practice. As instructions for this technique are available in books and articles and there is no explicit instruction in the literature that an experienced practitioner teach the technique or that individualized instruction is necessary, it is presumed that the criteria for successful meditation practice rests with the subjective evaluation of the meditator. However, instruction may be given and, if this is the case, presumably successful practice is judged by an experienced meditator.

Zen Buddhist Meditation

Zen Buddhist meditation, or Zazen, perhaps one of the most well-known forms of meditation, is a school of Mahayana Buddhism103 that employs meditation techniques that originated in India several thousand years ago and were introduced to Japan from China in 1191 A.D.104 Zen Buddhist meditation is typically divided into the Rinzai and Soto schools.

Main components. The harmony of the body, the breath, and the mind is considered essential to the practice of Zen. In the traditional forms of Zen meditation, physical preparation involves eating nutritious food in modest amounts.104

Posture is of great importance in Zen meditation. In traditional forms, Zen meditation is performed while seated on a cushion in either the full-lotus or half-lotus position; however, many Western practitioners practice in a variety of ways from chair sitting to full lotus.104 In the full-lotus position, the legs are crossed and the feet rest on top of the thighs. In the half-lotus position, only one foot is brought to rest on top of the thigh, the other remaining on the ground as in the regular cross-legged position.104,105 The hands are held in one of two prescribed ways, either with the left hand placed palm up on the palm of the right hand with the tips of the thumbs touching, or with the right hand closed in a loose fist and enclosed in the left hand, the left thumb between the web of the thumb and the index finger of the right hand.104 The spine is held straight and with the top of the head thrust upward, with the chin drawn in and the shoulders and abdomen remain relaxed. The body should be perpendicular and the ears, shoulders, nose, and navel should be in line. The tongue should touch the upper jaw and the molars should be in gentle contact with one another. The eyes should be half closed and the gaze focused on a point on the floor approximately 3 feet in front.104,105

Breathing. Breathing in Zen meditation is active and many breathing patterns are used. One deep breathing pattern begins with exhaling completely through an open mouth and letting the lower abdomen relax. Air is then inhaled through the nose and allowed to fill the chest and then the abdomen. This breathing pattern is repeated 4 to 10 times. The mouth is then closed, and air is inhaled and exhaled through the nose only. By the use of abdominal and diaphragmatic pressures, air is drawn in and pushed out. Both inhalation and exhalation should be smooth, with long breaths.104 After practitioners have learned to focus on their breath by counting, counting is omitted and meditators practice “shikantaza,” which means “nothing but precisely sitting.”106 Shikantaza is the most advanced form of Zen meditation.106 With practice, the frequency of breathing becomes about three to six breaths per minute.104

Attention and its object. Attention is focused on counting breaths or on a koan, a specific riddle that is unsolvable by logical analysis.106 The frequency of breathing is silently counted in one of three ways: counting the cycles of inhalation and exhalation, counting inhalations only, or counting exhalations only.104 Though some koans have become famous in the West (e.g., what is the sound of one hand clapping?), in practice, beginners often silently repeat the sound “mu” while counting. As a student advances, there are many koans that may be worked on over a period of years.47 This silent repetition allows the meditator to become fully absorbed in the koan. In both counting of breaths and focusing on a koan, it is essential that the concentration of the mind is on the counting or on the koan and not on respiration as such.104 No attempt is made to focus the mind on a single idea or experience; the meditator sits, aware only of the present moment.49

Spirituality and belief. It is generally accepted that Buddhist metaphysical beliefs are not essential to the practice of Zen. At a spiritual level, Zen is considered a recognition of or, more accurately, the constant participation of all beings in the reality of each being.49 Sitting should be based on the compassionate desire to save all sentient beings by means of calming the mind; however, this belief is not essential to practice. Only the wish to save all sentient beings and the strength to be disciplined in practice is necessary.104

Training. Depending on the purpose, Zen meditation may be practiced for a few minutes or for many hours.103

Criteria of successful meditation practice. Successful meditation practice is judged in terms of the internal changes that are brought about by cultivating awareness. The practice of Zen meditation should not be done with the aim of accomplishing some purpose or acquiring something.104 Examples of incorrect aims or approaches include (1) sitting in order to tranquilize the mind, (2) sitting to be empty in one's mind, (3) attempting to solve a koan as if playing a guessing game, and (4) being motivated by a wish to escape from everyday conflicts.104 Some Zen masters believe that it is acceptable for prospective students to be motivated by desires for good health, composure, iron nerves, etc., because in time their attachment to these less important purposes will be recognized.104 The successful practice of Zen meditation is often described in terms of an awareness of the “true nature” of reality, of discovering the extent to which ordinary experience is constructed and manipulated by our interests, fears, and purposes. Thus, successful practice results in the realization that a dreamlike absorption in personal intentions is actually the principal content of daily mental life,49 freeing the practitioner from circumstance and emotion.104

Mindfulness-Based Stress Reduction

The MBSR program emerged in 1979 as a way to integrate Buddhist mindfulness meditation into mainstream clinical medicine and psychology.107 Originally designed by Dr. Jon Kabat-Zinn at the University of Massachusetts Medical Center, the MBSR program was a group-based program designed to treat patients with chronic pain. Since then, MBSR has also been used to reduce morbidities associated with chronic illnesses such as cancer and acquired immunodeficiency syndrome and to treat emotional and behavioral disorders.98

Main components. The mindfulness component of the program incorporates three different practices: a sitting meditation, a body scan, and Hatha yoga. In addition to the mindfulness meditation practice that forms the basis of the intervention, patients are taught diaphragmatic breathing, coping strategies, assertiveness, and receive educational material about stress.96 The foundation for the practice of MBSR is the cultivation of seven attitudes:

1.

nonjudgment, becoming an impartial witness to your own experience;

2.

patience, allowing your experiences to unfold in their own time;

3.

beginner's mind, a willingness to see everything as if for the first time;

4.

trust, in your own intuition and authority and being yourself;

5.

nonstriving, having no goal other than meditation itself;

6.

acceptance, of things as they actually are in the present moment; and

7.

not censoring one's thoughts and allowing them to come and go.48

In addition to these attitudes, a strong motivation and perseverance are considered essential to developing a strong meditation practice and a high degree of mindfulness.48 These attitudes are cultivated consciously during each meditation session.48 As with other mindfulness practices, posture and breathing are essential.48 The practitioner sits upright, either on a chair or cross-legged on the floor, and attempts to focus attention on a particular object, most commonly on the sensations of his or her own breath as it passes the opening of the nostrils or on the rising and falling of the abdomen or chest.48 Whenever attention wanders from the breath, the practitioner will simply notice the distracting thought and then let it go as attention is returned to the breath. This process is repeated each time that attention wanders from the breath. The MBSR program incorporates formal meditation (i.e., seated, walking, Yoga) and informal meditation (i.e., the application of mindfulness to the activities of daily life). In informal practice, practitioners are reminded to become mindful of their breath to help induce a state of physical relaxation, emotional calm, and insight.48

The seated meditation is done either on the floor or on a straight-backed chair.48 When sitting on the floor, a cushion approximately 6 inches thick should be placed beneath the buttocks. The practitioner may use the “Burmese” posture in which one heel is drawn in close to the body and the other leg is draped in front, or a kneeling posture, placing the cushion between the feet.48 The sincerity of effort matters more than how one is sitting.48 Posture should be erect with the head, neck, and back aligned. The shoulders should be relaxed and the hands are usually rested on the knees or on the lap with the fingers of the left hand above the fingers of the right and the tips of the thumbs just touching each other.48

The body scan is the first formal mindfulness technique that meditators do for a prolonged period and is practiced intensively for the first 4 weeks of the program. Body scanning involves lying on your back and moving the mind through the different regions of the body, starting with the toes of the left foot and moving slowly upwards to the top of the head. Scanning is done in silence and stillness.

The third formal meditation technique used in the MBSR program is mindful Hatha yoga. It consists of slow and gentle stretching and strengthening exercises along with mindfulness of breathing and of the sensations that arise as the practitioner assumes various postures48

Breathing. Breathing is passive and without any specific pattern.48

Attention and its object. During sitting meditation, the attention is focused on the inhalation and exhalation of the breath or on the rising and falling of the abdomen. When the mind becomes distracted with other thoughts, the attention is gently, but firmly returned to the breath or abdomen. During the body scan, attention is focused on the bodily sensations. When the mind wanders, attention is brought back to the part of the body that was the focus of awareness.48 In contrast to other Yoga practices, mindful Hatha yoga is focused less on what the body is doing and more on maintaining moment-to-moment awareness. As in the seated meditation and body scan, the attention is focused on the breath and on the sensations that arise as the various postures are assumed.

Spirituality and belief. MBSR was designed as a secular, clinical practice and its practice does not require adopting any specific spiritual orientation or belief.

Training. The program consists of an 8-week intervention with weekly classes that last 2 to 3 hours. There is a day-long intensive meditation session between the sixth and seventh sessions.48,96 Participants also complete 45-minute sessions at home, at least 6 days a week for 8 weeks.48 During the 2-hour weekly sessions, participants are instructed in the informal and formal practice of mindfulness meditation. Participants must commit to a daily, 45-minute home practice of the skills taught during the weekly meetings.48 The components of practice change as participants become more adept in sitting meditation, body scan, and Yoga. Body scan is initially practiced at least once per day for 45 minutes for about 4 weeks. It is then practiced every other day, alternating with Yoga.48

Criteria of successful meditation practice. The proper practice is determined by an experienced teacher. In the absence of any religious or spiritual component, the measure of success is the achievement of successful outcomes, whether subjective (reduced perceived stress, reduced anxiety, etc.) or objective (reduced blood pressure, reduction in medication usage, etc.).

Mindfulness-Based Cognitive Therapy

Developed by Zindel Segal, Mark Williams, and John Teasdale in the 1990s as a method for preventing relapse in patients with clinical depression, MBCT combines the principles of cognitive therapy with a framework of mindfulness to improve emotional well-being and mental health.98,108 Based on the MBSR program developed by Jon Kabat-Zinn, the original aim of the MBCT program was to help individuals alter their relationship with the thoughts, feelings, and bodily sensations that contribute to depressive relapse, and to do so through changes in understanding at a deep level.108

Main components. Like MBSR, the MBCT program incorporates seated meditation and body scan. The practice teaches patients decentering (the ability to distance oneself from one's mental contents), how to recognize when their mood is deteriorating, and techniques to help reduce the information channels available for sustained ruminative thought-affect cycles and negative reactions to emotions and bodily sensations.108 The core skill that the MBCT program aims to teach is the ability, at times of potential relapse, to recognize and disengage from mind states characterized by self-perpetuating patterns of ruminative, negative thought

Breathing. Breathing is passive and without any specific pattern.108

Attention and its object. During seated meditation, the attention is focused on the inhalation and exhalation of the breath or on the rising and falling of the abdomen. When the mind becomes distracted, the attention is gently, but firmly, returned to the breath or abdomen. During the body scan, attention is focused on the bodily. When the mind wanders, attention is brought back to the part of the body that was the focus of attention.

Spirituality and belief. Like MBSR, MBCT was developed as a secular, clinical intervention and does not require adopting any specific spiritual orientation or belief system.

Training. The program consists of an 8-week program, with one 2-hour session per week. Classes contain approximately 12 students. The program is divided into two main components: in sessions one to four, participants are taught to become aware of the constant shifting of the mind and how to bring the mind to a single focus using a body scan technique and breathing. Participants also learn how the wandering mind can give rise to negative thoughts and feelings. In sessions five to eight, participants learn how to handle mood shifts, either immediately or at a future time.

Like the MBSR program, participants must continue the sessions at home for 6 or 7 days and complete various homework exercises that teach and reinforce mindfulness skills and help participants to reflect on their mindfulness practice.108

Criteria of successful meditation practice. The presence of an instructor who is adept in the practice of mindfulness is crucial to the success of the program. It is generally believed that if instructors are not mindful as they teach, the extent to which class members can learn mindfulness will be limited.108 The proper technique is determined by an experienced practitioner. The measure of success is the achievement of successful prevention of relapse based on clinical criteria.

Yoga

The philosophy and practice of Yoga date back to ancient times, originating perhaps as early as 5,000 to 8,000 years ago.1,109,110 It has been argued that the rules or precepts set down in the first systematic work on Yoga, Patanjali's Yoga Sutras, do not set forth a philosophy, but are practical instructions for attaining certain psychological states.111,112 It is important to acknowledge the diversity of techniques subsumed under the term “Yoga.” Over many millenia, different yogic meditative techniques had been developed and used to restore and maintain health, and to elevate self-awareness and to also transcend ordinary states of consciousness, and ultimately to attain states of enlightenment.110

Yogic meditative techniques have been transmitted through Kundalini yoga, Sahaja yoga, Hatha yoga and other yogic lineages.113 Though there are numerous styles of Yoga;114 the styles vary according to the emphasis and combination of four primary components: asanas, pranayamas, mantras, and the various meditation techniques.115 In Kundalini yoga, there are thousands of different postures, some dynamic and some static, and also thousands of different meditation techniques, many of which are disorder specific.116,117 Kundalini yoga meditation techniques are usually practiced while maintaining a straight spine, and employ a large number of specific, and highly structured breathing patterns, various eye and hand postures, and a wide variety of mantras. All of these techniques supposedly have different effects and benefits in their respective combinations.

Within Hatha yoga, many “schools” have developed, each differing slightly in its emphasis on the use of breathing and postures: in Bikram Yoga, practitioners perform the same sequence of 26 asanas in each session; in Vini Yoga, emphasis on the breath makes for a slower-paced practice. Iyengar Yoga is distinguished from other styles by its emphasis on precise structural alignment, the use of props, and sequencing of poses.118,119 There are also two Tibetan yogic practices, Tsa Lung and Trul Khor, that incorporate controlled breathing, visualization, mindfulness techniques, and postures.120 In Yoga, it is also believed that the practice of meditation techniques can be enhanced by the proper cleansing and conditioning of the body through the asanas and breathing exercises, or pranayama techniques121 (though pranayama places particular emphasis on techniques of breathing, some pranayama also employ physical movements).122

In addition to the schools of Yoga described above, TM® and the secular meditation techniques RR and CSM are derived from classical yogic techniques.123 It is important to note that the techniques in any given school or type of Yoga represent distinct interventions, in much the same way that psychodynamic, cognitive-behavioral, and interpersonal therapies each involve different approaches to psychotherapy.124

The purpose of asanas, pranayams, and pratyahar (emancipation of the mind from the domination of the senses) is to help rid the practitioner of the distractions of body, breath, and sensory activity and to prepare the body and mind for meditation and spiritual development.114 The use of mantras is said to help cleanse and restructure the subconscious mind, and to help prepare the conscious mind to experience the various states of superconsciousness. The more advanced Yoga practices lie in dharana (concentration), dhyana (yogic meditation) and samadhi (absorption). Concentration involves attention to a single object or place, external or internal (e.g., the space between the eyebrows, the tip of the nose, the breath, a mantra [chanted loudly, softly, or silently] or attention to all of these elements simultaneously). When the mind flows toward the object of concentration uninterruptedly and effortlessly, it is meditation. When it happens for a prolonged period of time it leads to samadhi, the comprehension of the true nature of reality that ultimately leads to enlightenment and emancipates the practitioner from the bonds of time and space.123,125

Main components. Classical Yoga is an all-encompassing lifestyle incorporating moral and ethical observances (yamas and niyamas), physical postures (asanas), breathing techniques (pranayams), and four increasingly more demanding levels of meditation (pratyahar, dharana, dhyana, and samadhi).126,127 Due to the incredible diversity of techniques in yogic meditation practice, it is impossible to describe them in adequate detail here. Instead, we have attempted to provide the reader with a very general description of the main components of many yogic meditation techniques. The reader is directed to the reference list for more detailed information on specific Yoga styles or techniques.110,116,117,119,128,129

The most common translation of “asana” is “posture” or “pose” and it refers to both specific postures for gaining greater strength and flexibility and those used specifically to help achieve proper concentration for meditation. Asanas are practiced either standing, sitting, supine, or prone.130 The postures for strength and flexibility take each joint in the body through its full range of motion, stretching, strengthening, and balancing each body part.114 Depending on the particular yogic technique one follows and the individual level of practice, each asana is held anywhere from a few breath cycles (as long as 2 minutes) to as long as 10 minutes or, in the case of some advanced practices, even 2.5 hours.

In most schools, during each posture attention is directed to the breath—to the deep, in-out, rhythmic sensation—and awareness is brought to the area of the body that is being stretched or strengthened.130 Though poses may be held for a few seconds to a few minutes, the body can also be in constant dynamic motion. Muscles relax and loosen, changing the shape of the pose, and the in and out breath moves in rhythm with the body. The practitioner simply observes the physical or psychical sensations and emotions arising while suspending judgment. The asanas are interspersed with brief moments of relaxation during which the practitioner attempts to redirect or maintain an inward focus.130

In postures used specifically for meditation, for example in Kundalini yoga, the spine is kept straight and the practitioner can be seated in a chair with the feet flat on the floor or seated in a cross-legged posture, and specific directions are given regarding the positioning of the arms, hands, and eyes, (e.g., the palms of the hands can be pressed together with the fingers together pointing up at a 60-degree angle, and the sides of the thumbs rest on the sternum in what is called “prayer pose,”129 and the eyes are closed as if looking at a central point on the horizon, the “third eye,” or the notch region between the eyes). A mantra (again technique specific) may also be chanted, and/or a simple or complex breathing pattern may be employed.129 Alternately, the eyes might be kept open and focused on the tip of the nose or closed and focused on the tip of the chin or top of the head, again in conjunction with any number of a wide variety of breathing patterns, and/or mantras.129 In Sahaja yoga, practitioners sit in a relaxed posture with hands in front, palms upward. Attention is directed to a picture placed in front with a candle lit before it. Gradually when thoughts recede, meditators close their eyes and direct their attention to the “sahasrara chakra” or top of the head. The individual sits in meditation for about 10 to 15 minutes.131 The amount to which the eyes are open or closed also varies; eyes may be fully open, fully closed, or half-closed.

Breathing. A central focus for most yogic meditation techniques is the breathing pattern.119 Pranayams, or breathing exercises, involve the conscious regulation of rhythmic breathing patterns, where some or all of the inspiration, breath retention, expiration, and breath out phases are regulated according to specific ratios or times. The inspiration and expiration phases can also be regulated by breaking each breath of the inspiration and expiration into 4 parts, 8 parts, or 16 parts or only the inspiration may be broken while the expiration remains unbroken.132 In addition, a breath pattern may be employed selectively through either the left or right nostril (or a sequential combination of both), or specific combinations of the nose and mouth. A wide variety of broken breath patterns have been discovered that have varying effects. Some techniques may also require holding attention on the imagined flow of energy along the spinal column collaterally with the breathing rhythm, on the sensation of inhaled air touching and passing through the nasal passage, on other parts of the body, or on a mantra.129,133

In Hatha yoga, various patterns of respiration are closely coordinated with the body in either a static posture or with movement.134 There are many pranayama techniques described in Hatha yoga texts; however, the practice of pranayama in this tradition has four primary objectives: (1) a stepwise reduction in breathing frequency, (2) attainment of a 1:2 ratio for the duration of inspiration and expiration respectively, (3) holding the breath for a period at the end of inspiration that lasts twice the length of expiration, i.e., a 1:4 ratio between inhalation and retention, and (4) mental concentration on breathing.121,135 The four objectives are united in the achievement of a single purpose, namely, the slowing down of respiration to achieve an immediate intensification of consciousness through the elimination of external stimuli.136

Practices such as Sudarshan Kriya Yoga involve rhythmic breathing at different rates following ujjayi pranayama (long and deep breaths with constriction at the base of throat) and bhastrika (fast and forceful breaths through the nose along with arm movements).137,138 Other practices, such as Iyengar Yoga, instruct the practitioner to breath through the nostrils only while performing the asanas.139 Some varieties of pranayama require the practitioner to inhale and exhale through one nostril selectively, a practice called unilateral forced nostril breathing.119,140 These breathing exercises are often practiced in combination with different postural locks (bandhas). Bandhas are restrictive positions or muscle maneuvres that exercise certain parts of the body. The most common of these are the abdominal lift (uddiyana bandha), the root lock (mula bandha), and the chin lock (jalandhara bandha).123

In Kundalini yoga, there are hundreds of different breathing patterns, each having unique and specific benefits and effects. In “Sodarshan Chakra Kriya,” considered one of the most powerful pranayama meditation techniques in Kundalini yoga, a unilateral forced nostril breathing pattern is employed selectively with inspiration through the left nostril, with breath retention, and with selective expiration through the right nostril. During the breath retention phase the abdomen is pumped in and out 48 times and a three-part mantra is mentally repeated 16 times in phase with the abdominal pumping (one repetition of the three-part mantra with three pumps), and the eyes are open and focused on the tip of the nose. As the technique is mastered, the rate of respiration is eventually reduced to less than one breath per minute and practiced for a maximum of 2 hours and 31 minutes.129

Attention and its object. Inherent in the practice of Yoga is an effortful progression toward increased concentration, or, more precisely, toward entering a state in which the mind is highly stable and still, consciously and purposely focused, and ordinary thoughts are suspended, and the meditator is more aware of the present moment (samadhi).141,142 This state has been described as the complete merging of the subjective consciousness and the object of focus.130 Hatha yoga has been defined as gentle stretching and strengthening exercises with constant awareness of breathing and of the sensations that arise as the meditator assumes various postures.76,128 By manipulating the body and making minute, detailed adjustments to perfect each posture, a person develops “one-pointed” concentration and ceases to become distracted by extraneous thoughts.130

One Hatha yoga technique, Shavasana, or corpse pose, involves lying on the back, with legs resting on the floor slightly apart, arms at the sides, palms facing up, and eyes closed. This seemingly simple pose is actually one of the most demanding to perfect because of the practitioner's need to achieve absolute stillness and total concentration as well as control over the breath.119 If drowsiness occurs, practitioners are told to increase the depth of their breathing. If the mind is restless, attention to the breathing cycle or other bodily sensations is encouraged. The goal is to rest in a state of relaxation, yet be aware of raw, sensory information and to let go of any reactions or judgments.121

In Kundalini yoga, one complex meditation technique called “Gan Puttee Kriya”, with multiple aspects of focus, is said to help eliminate negative thoughts, “psychic scarring,” and acute stress.116 The practitioner sits with a straight spine, either on the floor or in a chair. The backs of the hands are resting on the knees with the palms facing upward. The eyes are open only one-tenth of the way, but looking straight ahead into the darkness, not the light below. The practitioner chants consciously from the heart center in a natural, relaxed manner at a rate of one sound per second. The practitioner begins by chanting “SA” (the A sounding like “ah”), and touching the thumbtips and index fingertips together quickly and simultaneously then chanting “TA” and touching the thumbtips to the middle fingertips, then chanting “NA” and touching the thumbtips to the ring fingertips, then chanting “MA” and touching the thumbtips to the little fingertips, then chanting “RA” and touching the thumbtips and index fingertips, then chanting “MA” and touching the thumbtips to the middle fingertips, then chanting “DA” and touching the thumbtips to the ring fingertips, then chanting “SA” and touching the thumbtips to the little fingertips, then chanting “SA” and touching the thumbtips and index fingertips, then chanting “SAY” (like the word “say”) and touching the thumbtips to the middle fingertips, then chanting “SO” and touching the thumbtips to the ring fingertips, then chanting “HUNG” and touching the thumbtips to the little fingertips. The thumbtips and fingers touch with about 2 to 3 pounds of pressure with each connection which supposedly helps to consolidate a circuit created by each thumb-finger link. The techniques can be practiced for 11 minutes (or less) to a maximum of 31 minutes. When finished, the practitioner remains in the sitting posture and inhales and holds the breath for 20 to 30 seconds while shaking and moving every part of the body vigorously, with the hands and fingers moving very loosely, then exhaling and repeating this two additional times, immediately followed by opening the eyes and focusing them on the tip of the nose and breathing slowly through the nose for one minute

Spirituality and belief. Yoga is a science and philosophy of the human mind and body; it is a way of life, moral as well as practical.143 Yoga predates all formal religions,1,129 and, perhaps for this reason, the practice of Yoga does not presuppose an individual's commitment to a particular philosophical or religious system.144,145

Training. The ethical principles of Yoga describe the essential attitudes and values that are needed to undertake the safe practice of Yoga. The physical practice of Yoga focuses on the development of the strength, flexibility, and endurance of the body, strengthening of the respiratory and nervous systems, development of the glandular system, and increasing the ability to concentrate. In its complete form, Yoga combines rigorous physical training with meditation practices, breathing, and sound/mantra techniques that lead to a mastery of the body, mind, and consciousness. Both ancient commentaries on Yoga and more modern books of instruction stress the importance of learning under the guidance of an experienced teacher, Guru or Master.110,121,139 However, some Yoga techniques, especially asanas, pranayams, and meditation techniques, have been described and illustrated in books and videos produced for the purpose of self-study.139 In terms of specific training requirements, it is recommended that Yoga exercises be practiced daily, preferably in the morning, and on an empty stomach.139 Exercises can last from 15 minutes to several hours and it can take several years of consistent practice before a practitioner is able to practice properly the more demanding asanas and meditation techniques.121

Criteria of successful meditation practice. The ideal instruction in and assessment of Yoga techniques comes from a Guru or Master. Nevertheless, as books and video instruction are available, it can be assumed that the practitioner is able, to varying degrees, to assess the correctness of at least some asanas, pranayams, and a wide variety of meditation techniques

Yoga is ultimately a tradition of spiritual self-discipline and practice for the pursuit of enlightenment.136 Like Vipassana and Zen Buddhism, the success of meditation practice is judged on the basis of the practitioner achieving this state of enlightenment or other intermediate psychological or spiritual states. For example, the central experience achieved through Sahaja yoga meditation is a state called “thoughtless awareness” or “mental silence” in which the meditator is alert and aware but is free of any unnecessary mental activity.12 The state of thoughtless awareness is usually accompanied by emotionally positive experiences of bliss. In general, the outcome of the meditative process is associated with a sense of relaxation and positive mood and a feeling of benevolence toward oneself and others.146

As Yoga also involves exercises to strengthen the body and voluntarily control different aspects of breathing, success in these techniques can be evaluated against the standards for practice (e.g., achieving a 1:4:2 ratio in inhalation, retention, and exhalation), or developing the ability to reduce the rate of respiration to one breath per minute for 1 or 2 hours. Successful practice can also be determined by a subjective and objective evaluation of the achievement of some of the reported health benefits

Tai Chi

Tai Chi (also romanized as Tai Chi Ch'uan, T'ai Chi Ch'uan, Taijiquan, Taiji, or T'ai Chi) has a history stretching back to the 13th century A.D. to the Sung dynasty.147 There are five main schools, or styles, of Tai Chi, each named for the style's founding family: Yang, Chen, Sun, Wu (Jian Qian), and Wu (He Qin).148 Each style has a characteristic technique that differs from other styles in the postures or forms included, the order in which the forms appear, the pace at which movements are executed, and the level of difficulty of the technique.148 Though differing in focus on posture and the position of the center of gravity, all styles emphasize relaxation, mental concentration, and movement coordination.147 Tai Chi practice usually involves the need to memorize the names associated with each posture and the sequence of postures.148

Main components. The practice of Tai Chi encompasses exercises that promote posture, flexibility, relaxation, well-being, and mental concentration.148,149 It is characterized by extreme slowness of movement, absolute continuity without break or pause, and a total focusing of awareness on the moment.150 Unlike most exercises that are characterized by muscular force and exertion, the movements of Tai Chi are slow, gentle and light. The active concentration of the mind is instrumental in guiding the flow of the body's movements.151 Thus, Tai Chi is not only a physical exercise, but also involves training the mind, and this has prompted some to consider the practice “moving meditation.”148150 Although Tai Chi follows the principles of other types of martial arts that focus on self-defense, its primary objective is to promote health and peace of mind. In contrast to other martial arts, Tai Chi is performed slowly, with deep and consistent breathing.151 The movements should be performed in a quiet place that will help the practitioner to achieve a relaxed state. The muscles and joints are relaxed and the body is able to move easily from one position to another. The spine is in a natural erect position, and the head, torso, arms, and legs should be able to move freely and gently. The upper body is straight, never bending forward or backward, or leaning left or right.152

Breathing. Several different breathing techniques are employed in Tai Chi; however, the principal breathing technique, called “natural breathing,” is the foundation for all other breathing techniques. In natural breathing, the practitioner takes a slow, deep (but not strained) breath, inhaling and exhaling through the nose. The mouth is closed, but the teeth are not clenched. The tip of the tongue is held lightly against the roof of the mouth. As the air is taken in, the lower abdomen expands. Once the lungs are adequately filled with air, the person exhales and the lower abdomen contracts. The breath is never held. The eyes should be lightly closed.152

The movements of Tai Chi are coordinated with the breath, and the pattern of breathing follows the succession of opposing movements of the arms: inhalation takes place when the arms are extended outward or upward, exhalation occurs as arms are contracted or brought downward. Breathing eventually becomes an unconscious part of the exercise; however, its importance in the practice never diminishes.150

Attention and its object. Throughout the practice, the mind remains alert but tranquil, directing the smooth series of movements and focusing on one's internal energy. This active concentration is integral to the practice.149,151 It has been argued that if Tai Chi movements are performed without concentration, Tai Chi is no different from other forms of exercise. The variety and distinctiveness of the movements ensure that one concentrates on the execution of the movements.151

Spirituality and belief. Tai Chi derives its philosophical orientation from the opposing elements of yang (activity) and yin (inactivity) and from qi (breath energy).147 In accordance with the symbols of yin and yang, Tai Chi movements are circular. The movements are designed to balance the qi, or vital energy, in the meridians of the body, and strengthen the qi, thus preventing illness.153 Like Yoga, the practice of Tai Chi does not require adopting a specific spiritual or belief system and has been used clinically as a therapeutic intervention

Training. The exercise routines of the different forms of Tai Chi vary in the number of postures and in the time required to complete the routine,147 with some Tai Chi programs being modified to suit the abilities of practitioners with declining physical and mental function.148 Classical Yang Tai Chi includes 108 postures with some repeated sequences. Each training session includes a 20-minute warm-up, 24 minutes of Tai Chi practice, and a 10-minute cooldown. The warm-up consists of 10 movements with 10 to 20 repetitions. However, the exercise intensity depends on training style, posture, and duration.154

When practiced solely as an exercise form, sessions should occur twice a day and last about 15 minutes, 4 or more days per week.147 Practitioners are not required to continue training permanently with a Tai Chi teacher, and can continue practice as a form of self-therapy.152 When used as a system of self-defense, Tai Chi must be practiced with a Master and long enough to develop a deep understanding and “body memory” of the movements.155 However, as a healing practice, years of study are not required and the typical practitioner may be able to learn the fundamental movements within a week.155

Criteria of successful meditation practice. The overall aim is not to “master” the movements, but to appreciate a developing sense of inner and outer harmony as the movements become more fluid, yet controlled, and the mind more alert, yet peaceful.149 To learn and practice Tai Chi successfully, practitioners must adopt and practice specific traditional principles of posture and movement such as holding the head in vertical alignment, relaxing the chest and straightening the back, using mental focus instead of physical force, and seeking calmness of mind in movement.148

Qi Gong

Qi Gong is classified as one of the practices known as “energy healing,” a category that includes Reiki, therapeutic touch,156 and the Korean practice of Chundosunbup. Dating back more than 3,000 years to the Shang Dynasty (1600 to 1100 B.C.), Qi Gong is believed to be the basis for traditional Chinese medicine.157 Qi Gong is intimately connected with the practice of Tai Chi in that both exercises utilize proper body positioning, efficient movement, and deep breathing. A quiet focused mind is also essential to both. The main difference between Qi Gong and Tai Chi is that Tai Chi is a martial art. Usually practiced slowly, Tai Chi movements can be sped up to provide a form of self-defense, whereas this is not the case with the forms of Qi Gong. As a result, the visualization that accompanies a particular form is different: for a movement in Tai Chi that might involve visualizing the external consequences of a motion (e.g., disabling one's adversary), the same movement in Qi Gong would involve the visualization of an internal consequence of qi flow (e.g., qi flowing down your arm, healing your arthritis).155 There are two forms of Qi Gong practice: internal (nei qi), consisting of individual practice, and external (wai qi), whereby a Qi Gong practitioner “emits” qi for the purpose of healing another person.156,158 External Qi Gong is not a meditative practice according to the working definition developed for this report. Specifically, is not a self-applied practice, and there is a relationship of dependency between the practitioner and the person being treated. For this reason, this review is restricted to studies using internal Qi Gong.

Qi Gong is said to have several thousand forms. There are five main schools or styles of Qi Gong, each emphasizing a different purpose for practice157 and incorporating different exercises: Taoist, Buddhist, Confucian, Medical, and Martial.155 It is believed that every Qi Gong style has its own special training methods, objectives, and compatibility with an individual's constitution and physique.159 Despite this variation in technique, the main function of Qi Gong is to regulate the mind.160

Main components. Qi Gong, literally “breathless exercise,” consists primarily of meditation, physical movements, and breathing exercises. The main components of Qi Gong vary, but most emphasize correct posture and body alignment, regulation of respiration, posture, and mind, as well as self-massage and movement of the limbs.155,160 In general, Qi Gong consists of two aspects: (1) dynamic or active Qi Gong, which involves visible movement of the body, typically through a set of slowly enacted exercises, usually performed in a relaxed stationary position;155 and (2) meditative or passive Qi Gong, which comprises still positions with inner movement of the diaphragm.156 In some concentration practices, the eyes are closed and the tip of the tongue touches the front of the upper palate.160 Essential to both aspects of practice are alert concentration, precise control of abdominal breathing, and a mental concentration on qi flow.156

Qi Gong, as a practice of self-regulation, includes regulation of the body (e.g., relaxation and posture), breath (to breathe deeply and slowly), and mind (thinking and emotion). Methods for the regulation of the mind vary. Some forms of Qi Gong stress thinking, e.g., focusing on a specific object or visualization. Other forms emphasize regulation of the emotions (e.g., a peaceful and calm mood), but let thinking go or remain “no-thought.” Accordingly, Qi Gong techniques may be classified as one of two forms: concentrative Qi Gong and nonconcentrative Qi Gong.161 Self-practice of Qi Gong consists of three major forms: guided movement (dynamic form), pile standing, and static meditation.162 Whether with motion or without, the aim of Qi Gong is to remove all thoughts and focus on a region of the body known as “dantian” (the elixir field). As the body relaxes, the mind concentrates on the elixir field and all other thoughts are erased, while respiration becomes deeper and gradually decreases in frequency. When the respiration rate is decreased to four or five times per minute, the subject falls into the so-called Qi Gong state.161 It is recommended that a student practice only one type of Qi Gong before learning another as not all techniques are congruent.155

Breathing. Qi Gong breathing is characterized by a concentration of attention on dantian in concert with inhalation, exhalation, and holding of breath in order to stimulate qi and blood, and to strengthen the body.159 There are many ways to regulate the breath in Qi Gong including natural breathing, chest breathing, abdominal and reverse-abdominal breathing, holding the breath, and one-sided nostril and alternating nostril techniques.160

Attention and its object. A main tenet of Qi Gong is that intention can direct the qi within the body; the mind leads the qi, and qi leads the blood.158 To exert this control over qi, the practitioner must calm the mind and clear it of thoughts. A person's success Qi Gong is directly related to the ability to concentrate in this way. This is done by focusing the mind and body on correct breathing, and the visualization of qi as a substance moving through the body.160

Spirituality and belief. Qi Gong posits the existence of a subtle energy (qi) that circulates throughout the entire human body. Pain and disease are considered to be the result of qi blockage or imbalance; strengthening and balancing qi flow can improve health and ward off disease.159,162 Taoism, an ancient spiritual tradition in East Asia, is a philosophical perspective underlying the practice of Qi Gong. The Tao is the indefinable ultimate reality—the process involving every aspect in nature and in the entire universe. Similar to the worldviews of Buddhism and Hinduism, Taoism emphasizes harmony with nature. The universe is viewed in a dynamically continuous flow and constant change.163

Basic concepts considered essential to the understanding of Qi Gong include qi, vital energy, and gong, the skill, control, training, cultivation and practice of adjusting physical, mental and spiritual phenomena. Yin and yang, two other crucial concepts, are complementary opposites: yin signifies decrease, stillness, darkness, the six solid organs (lungs, spleen, heart, kidneys, pericardium, and liver), and bodily substances; yang signifies increase, activity, lightness, the upper and exterior parts of the body, the six hollow organs (large intestine, stomach, small intestine, urinary bladder, gallbladder, san jiao [not an organ, but the sum of the functions of transformation and interpenetration of various densities and qualities of substance within the organism]), and bodily functions.160

Training. Because of the possibility of Qi Gong-induced disorders from improper practice, or from the combination of incongruent forms, proper coaching is considered mandatory for safe Qi Gong practice.159 Qi Gong should be practiced twice daily for 20 to 30 minutes160,164 with no single session exceeding 3 hours.159

Criteria of successful meditation practice. Correctness of technique is judged by a Qi Gong Master. No statement of the criteria for evaluating successful outcomes was available in the literature.

Characteristics of Meditation Practices

Main Components

What are the main components of the various meditation practices? Which components are universal and which ones are supplemental?.

The variety of meditation practices is an indication of the diversity of the combination of main components and the way in which a given component may be emphasized in a practice. Given the multitude of practices and the many variations or techniques within these practices, it is impossible to select components that might be considered universal or supplemental across practices. Some practices prescribe specific postures (e.g., Zen Buddhist meditation, Tai Chi, Yoga) while others are less concerned with the exact position of the body (e.g., TM®, RR, CSM). Some practices (e.g., Vipassana, Zen Buddhist meditation, Yoga, Tai Chi, and Qi Gong) incorporate moving meditation, while others are strictly seated meditations (e.g., TM®, RR, and CSM). Some clinically-based practices (e.g., MBSR, MBCT), though guided by the underlying practice of mindfulness, combine several techniques. In this, however, they are not substantially different from older multifaceted meditation practices such as Yoga.

More detailed summaries addressing the main components used to describe individual practices are described below and summarized in Table 4. However, it is worth noting here some general conclusions that can be drawn from them. Though some statement about the use of breathing is universal across the practices, this seems more indicative of the ubiquitousness of breathing in humans rather than a universal feature of meditation practices per se. The control of attention is putatively universal; however, as noted below, there are at least two aspects of attention that might be employed and a wide variety of techniques for anchoring the attention, no one of which is universal. In terms of the spiritual or belief component of meditation, no meditation practice required the adoption of a specific religious framework. However, if Taoist metaphysical assumptions of Qi Gong are crucial to correctly understanding, visualizing, and guiding qi flow, then at least this practice would seem to require the adoption of a particular belief system. Nevertheless, this aspect of all meditation practices is poorly described, and it is unclear in what way and to what extent spirituality and belief play a role in the successful practice of meditation at all levels. The amount of variation in the described frequency and duration of practice make it difficult to draw generalizations about the training requirements for meditation techniques. Lastly, the criteria for successful meditation, for both the correct practice of the technique and the achievement of successful outcomes, have not been described well in the literature.

Table 4. Characteristics of included meditation practices.

Table 4

Characteristics of included meditation practices.

Breathing

How is breathing incorporated in these practices? Are there specific breathing patterns that are integral elements of meditation? Is breathing passive or directed?

The use of the breath is ubiquitous in all practices; however, the importance and attention given to it vary from practice to practice. Each meditation practice and technique has a breathing pattern or element that can be considered integral to that technique, whether the breath is actively controlled in terms of its timing and depth (e.g., Zen Buddhist meditation, Yoga, Tai Chi), or passive and “natural” (e.g., TM®, RR, CSM, Vipassana, MBCT). The practice of Yoga, which covers thousands of techniques, uses both active and passive breathing. Though the direction for active breathing may be relatively uniform across the techniques in a given practice (e.g., Zen Buddhist meditation), other practices use a wide array of breathing techniques that change according to the outcome desired (e.g., Kundalini yoga). For those practices that utilize passive breathing, there is no consistent pattern or rhythm as “breathing naturally” will vary from practitioner to practitioner.

Attention and Its Object

For each type of meditation practice, where is the attention directed during meditation (e.g., mantra, breath, image, nothing)?

The purposeful focusing of attention is considered crucial in all meditation practices. However, like breathing, the techniques for anchoring attention vary and there is no single method shared by all practices. For those practices that use a mantra (e.g., TM®, RR, CSM), in some the mantra may be repeated silently, and in some aloud. The factors surrounding the choice of the mantra vary and the nature of the mantra chosen will influence the number of associations brought forth by the word and the vibrations caused by the vocalization of the mantra. Some mantras will have no meaning to Western practitioners unfamiliar with Sanskrit (e.g., TM®, CSM, Yoga), while others will (e.g., RR).

Other forms of meditation practice focus attention on bodily sensations (e.g., Vipassana, MBSR, MBCT) or a body part (e.g., Tai Chi) to the exclusion of other thoughts. The so-called mindfulness techniques focus on the breath and cultivate an objective openness to whatever comes into awareness.72 Though this may be interpreted as not focusing attention, or, as it is sometimes paradoxically phrased, as focusing on nothing, the attention is controlled and directed with the aim of achieving a distance from one's emotional and cognitive responses to the objects in the field of attention. The difference between mindfulness meditation and other practices lies in the acceptance of these other thoughts into the field of awareness.

Though the distinction between concentrative and mindfulness meditation has prima facie validity, the reality is somewhat more complicated because some practices, such as Zen and Vipassana, have phases where concentration is used, and for which certain techniques such as counting or concentrating on a mantra are employed, while at other stages broad spaced mindful attention is encouraged.

Spirituality and Belief

To what extent is spirituality a part of meditation? To what extent is belief a part of meditation?

The one common feature of all meditation practices examined in this review is the apparent ability to practice meditation without adopting a specific system of spiritual or religious belief. However, the extent to which spirituality and belief are part of any given meditation practice is poorly described. Furthermore, if the Taoist metaphysical assumptions of Qi Gong are crucial to successfully understand, visualize, and guide qi, then at least this practice requires adoping a specific belief system

The extent to which spirituality or belief play a role in any meditation practice appears to depend in large part on the individual practitioner. Though the traditional practices were developed within specific spiritual or religious contexts (Vipassana, Zen Buddhist meditation, Yoga, Tai Chi, Qi Gong), and therefore have spiritual or religious aspects, this does not mean that a practitioner must adopt the belief systems upon which they were based. In addition, some practices developed for purposes other than spiritual enlightenment; for example, Tai Chi and Qi Gong were developed within a system martial exercise and Traditional Chinese Medicine, respectively. Though Yoga, too, has spiritual and religious components, it is often considered more properly a system of metaphysics and psychology, especially when the ethical instructions are ignored. In summary, it appears that all meditation practices can be performed, to some degree, without adopting a specific system of spirituality or belief.

Training

What are the training requirements for the various meditation practices (e.g., the range of training periods, frequency of training, individual and group approaches)?

Training refers to the specific periods of practice, the frequency and duration of practice, and how long a practitioner is expected to train before becoming proficient in a given technique. The training for meditation varies with periods of practice, ranging from 5 minutes (RR, Vipassana) to several hours (Yoga). The frequency of practice ranges from daily (MBSR, MBCT, Tai Chi, Vipassana, Yoga) to twice daily (TM®, RR, CSM, Qi Gong). Zen meditation does not specify a frequency of practice. Few practices give a required duration of practice; however, some (Yoga, Zen Buddhist meditation) give an indication of the time required to master a given technique.

Criteria of Successful Meditation Practice

How is the success of the meditation practice determined (i.e., was it practiced properly)? What criteria are used to determine successful meditation practice?

The criteria of successful meditation practice is understood both in terms of the successful practice of a specific technique (i.e., the technique is practiced properly) and in terms of achieving the aim of the meditation practice (e.g., leading to reduced stress, calmness of mind, or spiritual enlightenment).

The successful practice of a specific technique is sometimes judged by an experienced or master practitioner (TM®, MBSR, Yoga, Tai Chi, Qi Gong), and in some cases it can be judged by the individual (RR, CSM). However, the proliferation of self-instruction books and videos for some of the practices that also recommend an experienced teacher implies that individuals may judge, to some degree, the success of a practice.

Search Results for Topics II to V

The combined search strategies identified 11,030 citations. After screening titles and abstracts, 2,366 references were selected for further examination. The manuscripts of 81 articles were not retrieved (Appendix E).* The majority of the unretrieved studies were abstracts from conference proceedings and articles from nonindexed journals and were requested through our interlibrary loan service, but did not arrive within the 9-month cutoff that we established for article retrieval. Therefore, the full text of 2,285 potentially relevant articles was retrieved and evaluated for inclusion in the review. The application of the selection criteria to the 2,285 articles resulted in 911 articles being included and 1,374 excluded. Figure 2 outlines study retrieval and selection for the review.

Figure 2. Flow-diagram for study retrieval and selection for the review.

Figure

Figure 2. Flow-diagram for study retrieval and selection for the review.

The primary reasons for excluding studies were as follows: (1) the study was not primary research on meditation practices (n= 909), (2) the study did not have a control group (n= 280), (3) the study did not report adequately on measurable data for health-related outcomes relevant to the review (n= 170), (4) the study did not examine an adult population (n= 9), and (5) the study sample included less than 10 participants (n= 6) (Appendix E)*. The level of agreement between reviewers for inclusion and exclusion of studies was substantial (kappa = 0.84, 95% CI, 0.80 to 0.87).

From 911 included articles, 108 were identified as multiple publications;165 that is, cases in which the same study was published more than once, or part of data from an original report was republished.166 The multiple publications were not considered to be unique studies and any information that they provided was included with the data reported in the main study (Appendix F).* The report that was published first was regarded as the main study. In total, 803 articles were included in this report 10 of which each reported on two studies. Therefore, this report included 813 unique studies reported in 803 articles.

Topic II. State of Research on the Therapeutic Use of Meditation Practices in Healthcare

General Characteristics

Eight hundred and thirteen studies provided evidence regarding the state of research on the therapeutic use of meditation. Tables G1 to G3 of Appendix G* summarize the key characteristics of studies included in topic II.

The studies were published between 1956 and 2005, with 51 percent of the studies (n = 417) published after 1994. Most of the studies (86 percent, n = 701) were published as journal articles. Seventy-nine (10 percent) were theses or dissertations, 25 (3 percent) were abstracts from scientific conferences, and 5 (0.5 percent) were book chapters or letters. Three unpublished studies (0.5 percent) were identified by contacting investigators. Studies were conducted in North America (61 percent), Asia (24 percent), Europe (11 percent), Australasia (3 percent) and other regions (1 percent).

Of the 813 studies included, 67 percent (n = 547) were intervention studies (286 RCTs, 114 NRCTs and 147 before-and-after studies), and 33 percent (n = 266) were observational analytical studies (149 cohort and 117 cross-sectional studies).

Methodological Quality

Intervention studies. Overall, the methodological quality of the 286 RCTs was poor (median Jadad score = 2/5; IQR, 1 to 2). Only 14 percent (n = 40) of the RCTs were considered of high quality (i.e., Jadad scores greater than or equal to 3 points). Three studies167169 obtained 4 points on the Jadad scale, and none obtained a perfect score (5 points). The remaining 246 RCTs had a high risk of bias.

The methodological quality of the RCTs was analyzed by the individual components of the Jadad scale. We found that 21 percent (n = 60) described how the randomization was carried out. Among these 60 trials, 75 percent (n = 45) reported adequate methods to randomize study participants to treatment groups, whereas 25 percent (n = 15) used inappropriate and unreliable methods (i.e., alternation or methods based on patient characteristics) that might have introduced imbalances and jeopardized the estimates of the overall treatment effect.

The vast majority of RCTs (97 percent, n = 278) did not use double blinding to hide the identity of the assigned interventions from the participant and assessor, or hide the hypothesis from the instructor and participant or participant and assessor. One of them170 described an inadequate method of double blinding while the others did not provide any description about the double-blinding procedures. Finally, 51 percent (n = 145) of the RCTs provided a description of withdrawals and dropouts from the study.

Concealment of treatment allocation (separating the process of randomization from the recruitment of participants) was adequately reported in 12 (4 percent) RCTs and was inadequate in 2 (1 percent) RCTs. The majority of RCTs (272, 95 percent) failed to describe how they concealed the allocation to the interventions under study. Finally, funding source was disclosed in 41 percent (n = 118) of the RCTs. A summary of the methodological quality of RCTs is presented in Table 5.

Table 5. Methodological quality of RCTs.

Table 5

Methodological quality of RCTs.

Overall, the quality of the 114 NRCTs was low (median modified Jadad score: 0/3; IQR, 0 to 1). Forty-six percent (n = 52) of the NRCTs obtained only 1 point out of 3 for the individual components of the Jadad scale, most frequently for the description of withdrawals and dropouts. The remaining 54 percent (n = 62) of the NRCTs did not score any points. Finally, the source of funding was cited in 26 percent (n = 30) of the NRCTs. A summary of the methodological quality of NRCTs is presented in Table 6.

Table 6. Methodological quality of NRCTs.

Table 6

Methodological quality of NRCTs.

The quality of the 147 before-and-after studies was poor. Only 16 percent (n = 23) of the before-and-after studies included representative samples of the target population. Descriptions of the number of study withdrawals (31 percent, n = 45), reasons for study withdrawals (14 percent, n = 20), and blinding of outcome assessors to intervention and assessment periods (2 percent, n = 3) were also infrequent. Better quality results were obtained for the homogeneity in the methods for outcome assessment for the pre- and postintervention periods for all participants. Finally, funding source was disclosed in 28 percent (n = 41) of the before-and-after studies. A summary of the methodological quality of the before-and-after studies is presented in Table 7. Studies that were included in the analysis of the methodological quality of RCTs, NRCTs, and before-and-after studies are summarized in Table G4 in Appendix G.

Table 7. Methodological quality of before-and-after studies.

Table 7

Methodological quality of before-and-after studies.

Observational analytical studies. The quality of reporting of cohort studies was evaluated with the individual components of the NOS scale regarding the selection and comparability of the cohorts, and outcome assessment. Overall, the methodological quality of the 149 cohort studies was poor (median NOS score = 3/9 stars; IQR; 2 to 4), suggesting a high risk of bias in these studies. Table 8 displays the methodological quality of the cohort studies assessed with the NOS scale.

Table 8. Methodological quality of cohort studies (NOS scale).

Table 8

Methodological quality of cohort studies (NOS scale).

In general, the cohort studies failed to protect against selection bias when assembling the exposed and nonexposed cohorts. Participants in 60 percent (n = 94) of the studies were not representative of the target population about which conclusions were to be drawn. The selection of the nonexposed cohort was equally compromised (62 percent, n = 93).

Detection bias affecting the ascertainment of both exposure and outcome was introduced in 139 (93 percent) studies. These studies did not use reliable methods to ensure that no differences in accuracy of exposure data between the cohorts existed. A similar proportion was found for studies that failed to demonstrate that the outcomes of interest were not present at the start of the study. Similarly, 105 (71 percent) cohort studies did not provide enough information to assess whether the length of the followup period was sufficient for outcomes to occur.

Attrition bias was substantial; only 20 percent (n = 29) of the studies reported followup rates unlikely to introduce differences between the comparison groups. The only methodological component that did not appear to be severely jeopardized was the control of confounders in the design or analysis. Sixty-six percent (n = 99) of the cohort studies adjusted for potential confounders either in the design or analysis. Finally, 28 percent (n = 41) of the cohort studies reported the source of funding.

The methodological quality of the cross-sectional studies was poor (median NOS total score = 2/6 stars; IQR, 1 to 3). The methodological characteristics of cross-sectional studies are summarized in Table 9. The cross-sectional studies had less prominent methodological weaknesses than the cohort studies.

Table 9. Methodological quality of cross-sectional studies (NOS scale).

Table 9

Methodological quality of cross-sectional studies (NOS scale).

Over half of the cross-sectional studies (53 percent, n = 62) chose study groups that were at least somewhat representative of the target population. However, only 21 percent of the studies (n = 24) drew the comparison groups from the same population as the study group. None of the studies used secure methods for ascertainment of exposure. Half of the cross-sectional studies (54 percent, n = 63) adjusted for potential confounders either in the design or analysis and used relatively reliable methods for assessing the outcomes (53 percent, n = 62). Finally, only 27 (23 percent) cross-sectional studies disclosed their source of funding.

Studies that were included in the analysis of the methodological quality of cohort and cross-sectional studies are summarized in Table G5 in Appendix G.*

Meditation Practices Examined in Clinical Trials and Observational Studies

Eight hundred and thirteen studies described meditation practices examined in intervention studies (RCTs, NRCTs, and before-and-after studies) and observational analytical studies (cohort and cross-sectional studies with control groups).

Overall, 86 percent (n = 698) of the studies reported on single interventions, whereas 14 percent (n = 115) reported on composite interventions. The composite interventions included either meditation practices combined with each other, or with other therapeutic strategies within holistic treatment programs. Table 10 reports the type of meditation practices that have been examined in intervention studies and observational analytical studies. Table G6 in Appendix G provides the references of studies included for this question along with their distribution by meditation practice and study design.

Table 10. Meditation practices examined in intervention and observational analytical studies.

Table 10

Meditation practices examined in intervention and observational analytical studies.

Mantra meditation. Forty-one percent (n = 337) of the included studies reported on interventions involving the use of a mantra as a pivotal component for the practice of meditation. The studies were published from 1972 to 2005, with 1986 the median year of publication (IQR, 1978 to 1991). Study sample sizes ranged from 10 to 602,000 participants with a median of 40 participants per study (IQR, 24 to 68).

A variety of mantra meditation techniques were assessed in the studies. The majority of the studies (68 percent, n = 230) focused on TM® or the TM®-Sidhi program. Fifteen percent (n = 51) reported on Benson's RR, and nine percent (n = 31) assessed practices in which words or phrases (mantra) were chanted aloud or silently and used as objects of attention. Mantra meditation techniques such as CSM, and SRELAX that are similar to TM®, but developed specifically for clinical purposes, were assessed in four percent (n = 12) of the studies”. Acem meditation, an amalgam of traditional meditation techniques and Western psychological theory and practices, was evaluated in two percent (n = 7) of the studies. Finally, three percent of the studies focused on other mantra techniques such as Ananda Marga (n = 3), concentrative prayer (n = 2), and Cayce's meditation (n = 1)

Design and methodology. Thirty-three percent (n = 111) of the studies on mantra meditation were RCTs, 31 percent (n = 105) were cohort studies, 18 percent (n = 60) cross-sectional studies, and 9 percent for each of before-and-after studies (n = 31) and NRCTs (n = 30). The methodological quality of intervention studies on mantra meditation was poor: The median Jadad score for RCTs was 1/5 (IQR, 1 to 2). Only 13 out of 111 RCTs (12 percent) scored 3 points or more on the Jadad scale and thus could be considered high quality. The median modified Jadad score for NRCTs was 0.5/3 (IQR, 0 to 1). The quality of before-and-after studies was poor. The methodological quality of observational studies was also low, with a median NOS total score for cohort studies of 3/9 stars (IQR, 2 to 4) and a median NOS total score for cross-sectional studies of 2/6 stars (IQR, 1 to 3). There were major deficiencies in the selection and comparability of the study groups.

Mindfulness meditation. Sixteen percent of the studies (n = 127) described the use of mindfulness meditation techniques, such as MBSR (n = 49), mindfulness meditation techniques not further described (n = 37), Zen Buddhist meditation (n = 28), MBCT (n = 7), and Vipassana meditation (n = 6). The studies were published from 1964 to 2005, with a median year of publication of 2001 (IQR, 1992 to 2003). Study sample sizes ranged from 10 to 719 with a median number of 39 participants per study (IQR, 23 to 73).

Design and methodology. Thirty-nine percent (n = 50) of the studies on mindfulness meditation were RCTs, 22 percent (n = 28) were before-and-after studies, 20 percent (n = 25) NRCTs, and 9 percent for each of cohort (n = 12) and cross-sectional studies (n = 12). The methodological quality of intervention studies on mindfulness meditation was low (RCTs median Jadad score = 2/5; IQR, 1 to 2; NRCTs median modified Jadad score: 0.5/3; IQR, 0 to 1). The quality of before-and-after studies was poor. Only 7 of 50 RCTs (14 percent) scored 3 or more points in the Jadad scale and were thus considered high quality. The observational studies also exhibited major methodological shortcomings (cohort studies median NOS total score: = 3/9 stars; IQR, 2 to 4; cross-sectional studies median NOS total score: 3/6; IQR, 1 to 3), particularly in the areas of selection and comparability of the study groups.

Meditation practices not described. Three percent of the included studies (n = 21) reported on meditation practices that were not described. The studies were published from 1974 to 2004, with a median year of publication of 1998 (IQR, 1990 to 2002). Study sample sizes ranged from 10 to 230 with a median number of 46 participants per study (IQR, 27 to 97).

Design and methodology. Almost half (n = 11) of the studies were RCTs, six were NRCTs, two before-and-after studies, one cohort and one cross-sectional study. The methodological quality of the intervention studies was low (RCTs median Jadad score = 1.5/5; IQR, 1 to 2; NRCTs median modified Jadad score = 0/3; IQR, 0 to 0). Only 1 out of 11 RCTs scored 3 or more points on the Jadad scale and thus was considered high quality. The quality of before-and-after studies was poor. The cohort and cross-sectional studies obtained three and two stars on the NOS scales, respectively. Both studies failed to select unbiased study samples, thus compromising the comparability of the groups.

Miscellaneous meditation practices. One percent of the included studies (n = 11) reported on interventions that combined different meditation techniques in a single intervention. The studies were published from 1980 to 2005, with a median year of publication of 1985 (IQR, 1981 to 1993). Sample sizes ranged from 11 to 340 with a median number of participants per study of 84 (IQR, 20 to 181).

Design and methodology. Three out of 11 studies were RCTs, 3 were before-and-after studies, 2 were cohort studies, and 3 were cross-sectional studies. The methodological quality of studies on miscellaneous meditation practices showed important flaws. All RCTs scored 2 points on the Jadad scale and were considered low quality. The quality of before-and-after studies was also poor. The observational studies exhibited the same methodological flaws as the studies of other interventions described above (cohort studies median NOS total score = 3/9 stars; IQR, 1 to 3; cross-sectional studies median NOS total score = 2/6; IQR, 2 to 3).

Qi Gong. Five percent of the included studies (n = 37) reported on Qi Gong interventions. The studies were published between 1956 and 2005, with a median year of publication of 2000 (IQR, 1996 to 2004). Study sample sizes varied from 10 to 254 with a median number of 36 participants per study (IQR, 22 to 73).

Design and methodology. Thirty-five percent (n = 13) of the studies on Qi Gong were RCTs, 24 percent (n = 9) were before-and-after studies, 19 percent cohort (n = 7), and 22 percent cross-sectional studies (n = 8). The methodological quality of studies on Qi Gong was poor (RCTs median Jadad score = 1/5; IQR, 1 to 2), all scoring less than 3 points on the Jadad scale. The quality of before-and-after studies was also poor. The quality of observational studies was low (cohort studies median NOS total score = 2/9 stars; IQR, 2 to 4; cross-sectional studies median NOS total score = 2.5/6; IQR, 2 to 3). Major deficiencies were found in the selection and comparability of the study groups.

Tai Chi. Eleven percent of the included studies (n = 88) reported on Tai Chi interventions. The studies were published from 1977 to 2005, with a median year of publication of 2002 (IQR, 1998 to 2004). Study sample sizes ranged from 10 to 311 with a median number of participants per study of 39 (IQR, 25 to 65).

Design and methodology. Thirty-three percent (n = 29) of the studies on Tai Chi were RCTs, 23 percent (n = 20) were before-and-after studies, 19 percent (n = 17) NRCTs, 20 percent (n = 18) cross-sectional studies, and 4.5 percent (n = 4) were cohort studies. The methodological quality of studies on Tai Chi was poor (RCTs median Jadad score = 2/5; IQR, 1 to 3; NRCTs median modified Jadad score = 1/3; IQR, 0 to 1). Nine out of 29 RCTs scored 3 or more points on the Jadad scale and thus were considered high quality. The quality of before-and-after studies was also low. The observational studies exhibited major flaws and were likely to be affected by bias (cohort studies median NOS total score = 2/9 stars; IQR, 2 to 4; cross-sectional studies median NOS total score = 2/6; IQR, 2 to 4).

Yoga. Twenty-four percent of the included studies (n = 192) reported on interventions involving Yoga practices. The studies were published between 1968 and 2005, with a median year of publication of 1998 (IQR, 1991 to 2002). Study sample sizes ranged from 10 to 335 with a median of 40 participants (IQR, 23 to 70).

Design and methodology. Thirty-six percent (n = 69) of the studies on Yoga interventions were RCTs, 28 percent (n = 54) were before-and-after studies, 19 percent (n = 36) NRCTs, 9 percent (n = 18) cohort studies, and 8 percent (n = 15) were cross-sectional studies. The methodological quality of studies on Yoga was low (RCTs median Jadad score = 1/5; IQR, 1 to 2; NRCTs median modified Jadad score = 0/3; IQR, 0 to 1). Fourteen percent (n = 10) of the RCTs on Yoga scored 3 points or more on the Jadad scale and were considered high quality. The quality of before-and-after studies was also poor. The methodological quality of observational studies was low (cohort studies median NOS total score: = 3.5/9 stars; IQR, 2.5 to 5; cross-sectional studies median NOS total score = 3/6; IQR, 1 to 3).

Tables 11 to 15 provide a comparative summary of the methodological quality of the studies classified according to the seven categories of meditation practices described in this report.

Table 11. Methodological quality of RCTs by meditation practice*.

Table 11

Methodological quality of RCTs by meditation practice*.

Table 12. Methodological quality of NRCTs by meditation practice*.

Table 12

Methodological quality of NRCTs by meditation practice*.

Table 13. Methodological quality of before-and-after studies by meditation practice*.

Table 13

Methodological quality of before-and-after studies by meditation practice*.

Table 14. Methodological quality of cohort studies by meditation practice*.

Table 14

Methodological quality of cohort studies by meditation practice*.

Table 15. Methodological quality of cross-sectional studies by meditation practice*.

Table 15

Methodological quality of cross-sectional studies by meditation practice*.

Control Groups Used in Studies on Meditation Practices

Six hundred and sixty-eight studies contributed data for this question (402 intervention studies [RCTs and NRCTs] and 266 observational analytical studies [cohort studies and cross-sectional studies with control groups]). One hundred and forty-five studies were excluded from this analysis because they were uncontrolled before-and-after studies. Only two before-and-after studies171, 172 had controlled comparisons and were considered for the analysis of the type of control groups used in studies on meditation practices.

Overall, the number of control groups per study ranged from one to four. The median number of control groups per study was one (IQR, 1 to 2). Table 16 shows the distribution of the number of control groups by study design. The majority of studies (72 percent, n = 482) included one control group per study, 21 percent (n = 139) used two control groups, 5 percent (n = 33) used three control groups, and, 2 percent (n = 14) used four control groups.

Table 16. Number of control groups by study design.

Table 16

Number of control groups by study design.

The majority of intervention studies and observational analytical studies considered in this review used single control groups (n = 482, 72 percent) as compared to the number of studies that used multiple control groups (n = 186, 28 percent). Tables 17 and 18 display the distribution of the number of control groups used in the intervention and observational analytical studies for each meditation practice.

Table 17. Controlled intervention studies: number of control groups by meditation practice*.

Table 17

Controlled intervention studies: number of control groups by meditation practice*.

Table 18. Observational analytical studies: number of control groups by meditation practice.

Table 18

Observational analytical studies: number of control groups by meditation practice.

Control groups from intervention studies (RCTs, NRCTs, and controlled before-and-after studies) were grouped into six categories according to the type of control group.173 As some studies used more than one control group as a comparator, the number of intervention studies reported below does not match the number of control groups. Tables 19 and 20 describe the types of control groups for intervention and observational studies along with their distribution by meditation practice. Table G7 in Appendix G* lists the references for studies included in the description of the type of control groups for intervention studies along with their distribution by meditation practice.

Table 19. Types of control groups for intervention studies on meditation practices.

Table 19

Types of control groups for intervention studies on meditation practices.

Table 20. Types of control groups for observational analytical studies on meditation practices.

Table 20

Types of control groups for observational analytical studies on meditation practices.

Sham meditation or placebo concurrent controls. Eighteen of 402 intervention studies (four percent) compared meditation practices with elaborately designed and executed sham procedures such as sitting in a comfortable position without being instructed in the use of any sound or in directing the attention in certain way. Half of the studies (n = 9) using sham meditation or placebo control groups were conducted on mantra meditation (three on TM®, three on mantra techniques not specified, two on RR, and one on SRELAX, a technique adapted from TM®). Evaluation of other practices that used sham meditation or placebo groups included three studies on meditation practices not further described, two studies on Qi Gong, two on Yoga, one study on mindfulness meditation (Zen meditation), and one on Tai Chi.

No-treatment concurrent controls. Two types of no-treatment conditions were included in the studies: no intervention and waiting lists (WL).

No intervention controls. One hundred and twenty-four out of 402 studies (31 percent) used control groups that received no intervention of any kind. Thirty-five percent (43/123) of these studies were conducted on mantra meditation (25 studies on TM®, 8 on mantra techniques not specified, 6 on RR, 1 on Acem meditation, 1 on Cayce's meditation). There were 30 intervention studies on Yoga that used a no-intervention condition as comparator. There were 22 studies no-intervention studies on mindfulness meditation (9 studies on MBSR, 7 on mindfulness meditation practices not further specified, 5 on Zen Buddhist meditation, and 1 study on MBCT), 19 on Tai Chi, 6 on meditation practices not further described, 2 on Qi Gong, and 1 on a miscellaneous technique called “coloring mandalas.”

Waiting list controls. Sixty-two (15 percent) of the intervention studies utilized a WL control group. Twenty-four were conducted on mantra meditation (10 studies on TM®, 5 on CSM, 5 on RR, 3 on mantra techniques not specified, and 1 on SRELAX, a technique modeled after TM®); 21 on mindfulness meditation (11 studies on MBSR, 6 on mindfulness meditation practices not further specified, 2 on MBCT, and 2 on Zen Buddhist meditation); 10 on Yoga, 3 on meditation practices not further described, 2 on Qi Gong, and 2 on Tai Chi.

Active (positive) concurrent controls: interventions other than meditation. Active concurrent controls, as opposed to placebo or no treatment concurrent controls (i.e., no intervention, and waiting list conditions) were used as comparisons in 306 intervention studies (90 percent). A wide variety of active comparison groups were employed.

Exercise and other physical activities. The practice of exercise and other physical activities constituted the most frequently used comparator (45 studies). Physical activities included, but were not limited to, aerobics, running, swimming, fencing, and stretching. Eighteen studies using exercise and other physical activities as controls were conducted on Yoga, 14 on Tai Chi, and 10 on mantra meditation (3 on mantra techniques not specified, 3 on RR, 2 on TM®, 1 on Acem meditation, 1 on CSM). One study was conducted on MBSR, one on meditation practices not specified, and one on Qi Gong.

Rest and states of relaxation. Conditions involving states of rest and relaxation were used as controls in 45 studies. There were 28 studies on mantra meditation (14 on RR, 9 on TM®, 3 on mantra techniques not specified, and 2 on CSM), 9 on Yoga, 6 on mindfulness meditation (3 on Zen Buddhist meditation, 2 on mindfulness meditation techniques not further specified, and 1 on MBSR), and 2 on other meditation practices not further described.

Educational activities. Forty-four studies used educational activities such as lectures and courses on stress management, nutrition, health, and wellness as comparators. Seventeen of these studies were conducted on mantra meditation (9 on TM®, 5 on RR, 2 on mantra techniques not specified, and 1 on CSM), 10 studies on mindfulness meditation (5 studies on MBSR, 3 on Zen Buddhist meditation, and 2 on mindfulness meditation techniques not further specified), 8 on Yoga, 6 on Tai Chi, 2 on meditation practices not further described, and 2 on miscellaneous meditation techniques.

Progressive muscle relaxation. The practice of progressive muscle relaxation (PMR) was chosen as a control group in 39 intervention studies. The majority of studies (n = 27) using PMR as a control were conducted on mantra meditation (10 on TM®, 8 on RR, 5 on mantra techniques not specified, 3 on CSM, and 1 on Acem meditation). There were also six studies on Yoga, five on mindfulness meditation (two on MBSR, two on mindfulness meditation techniques not further specified, and one on Zen Buddhist meditation), and one study on a meditation practice not further described.

Cognitive behavioral techniques. Twenty studies employed cognitive behavioral interventions as comparison groups. Nine of these studies were conducted on mantra meditation (three on TM®, three on RR, two on CSM, and one on mantra techniques not specified). There were seven intervention studies on mindfulness meditation (four on mindfulness meditation techniques not further specified, and three on MBSR). There were two studies on meditation practices not further described, and two studies on Yoga.

Pharmacological interventions. Eight studies used comparators involving pharmacological interventions such as antihypertensive medication, lipid-lowering medication, antidepressants, and other medications that were not described. There were six studies on Yoga, and two on Qi Gong that used a pharmacological intervention as a control.

Miscellaneous active controls. Nineteen studies reported on the use of control groups that involved a heterogeneous collection of active interventions, such as charting, creativity techniques, herbal therapy, visualization and other imagery, and cognitive tasks. Six of these studies were conducted on mantra meditation (three on RR, two on mantra techniques not specified, and one on TM®). There were also six studies on Yoga, four studies on mindfulness meditation (two on MBSR, one on Zen Buddhist meditation, and one on mindfulness meditation techniques not further specified), two on miscellaneous meditation practices, one on Tai Chi, and one on a meditation practice not further described.

Group therapy and psychotherapy. Sixteen studies used psychotherapeutic interventions such as group therapy (13 studies) and individual psychotherapy (3 studies) as comparison groups. Among the 13 studies that used group therapy as a control, 6 were on mantra meditation (3 on RR, 2 on TM®, and 1 on Acem meditation), 3 on mindfulness meditation (2 on mindfulness meditation techniques not further specified, and one on MBSR). There were also three studies on Tai Chi and two on Yoga that used group therapy as a comparator. Generally, group therapy was delivered as a form of group counseling and psychosocial support. Individual psychotherapeutic approaches were used as control groups in one study on mantra meditation (TM®), one study on MBSR, and one study on Yoga.

Biofeedback techniques. The practice of biofeedback (BF) techniques such as electromyographic (EMG) BF, and blood pressure BF was used as comparators in 12 intervention studies. The majority of the studies (n = 11) were conducted on mantra meditation (six on RR, three on mantra techniques not specified, and two on TM®) and one was conducted on Yoga.

Reading. Activities involving reading were utilized as controls in eight studies. There were six studies on mantra meditation (four on RR, and two on TM®), one on Tai Chi, and one on Yoga.

Hypnosis. Hypnosis was selected as a control group in four intervention studies: two on mantra meditation (TM®) and two on meditation practices not further described.

Therapeutic massage and acupuncture. Three studies used complementary interventions such as massage (two studies; one on RR, and another on MBSR) and acupuncture (one on Tai Chi) as comparison groups.

Usual care Thirty-seven intervention studies included a group of usual care in their comparisons. Nine of these studies were conducted on mindfulness meditation (3 on MBCT, 2 on mindfulness meditation techniques not further specified, 2 on MBSR, and 1 on Zen Buddhist meditation), 3 on Qi Gong, 3 on mantra meditation (2 on TM® and 1 on RR), 4 on Tai Chi, 16 on Yoga, and one on meditation practices not further described.

Other control groups. Six studies reported on the comparison groups in terms of controls without providing further comprehensive details. Two of these studies were conducted on mantra meditation (one on RR and one on TM®), two on Qi Gong, one on mindfulness meditation not further specified, and one on Tai Chi.

Active (positive) concurrent controls: meditation practices as comparison groups. Forty-three studies used meditation practices as control groups. Fourteen of these studies compared two different meditation practices against each other. Twenty-nine studies compared two versions of the same meditation practice but varied certain components of the practice, e.g., method of delivery, intensity, and length of session, of the comparison group. The former category of studies is described first and the latter is described under the category of “different dose or response concurrent control groups.”

Yoga practices. Four studies (three on TM® and one on mantra techniques not specified) compared mantra meditation techniques versus Yoga techniques such as Savasana. One study compared Hatha yoga versus a meditation practice not further described.

Mantra meditation. Three studies on Yoga (Kundalini, Sahaja, and Hatha yoga) used a mantra meditation technique for their comparison groups; two of them used RR169,174 the third 175 used a mantra technique not further described.

Mindfulness meditation. Two studies on mantra meditation (TM® and a mantra technique not further described) used interventions described as “mindfulness training” as comparison groups. Another study on a meditation practice not further described used mindfulness meditation as the comparison group.

Meditation practices not described. Two studies on mantra meditation (one on RR, and the other on TM®) failed to describe the type of meditation practice chosen for the comparison group.

Tai Chi: One study on mantra meditation (RR) used a Tai Chi-based intervention for the comparison group.

Different dose or regimen: concurrent control groups. Twenty-nine studies compared similar meditation practices but modified certain components of the practices to create the comparison groups.

Yoga practices. Fourteen studies compared different types of Yoga practices with each other. Nine studies140,176183 compared different patterns of yogic nostril breathing techniques (e.g., unilateral versus bilateral nostril breathing, left versus right forced unilateral nostril breathing), whereas five studies compared different modalities of yoga practice such as Hatha versus Astanga,127 different formats for practice (e.g., full Sudarshan Kriya versus partial Sudarshan Kriya),184 or combinations with other therapeutic strategies.111,175,185

Mantra meditation. Nine studies on mantra meditation compared different formats for the delivery of practice. Three studies186188 on TM® examined either short- versus long-term or regular versus irregular practice. Two other studies on TM®189,190 included RR as one of the comparators. There were two studies on RR that used TM®191 or modifications of the RR technique192 as comparison groups. One study on CSM193 used a RR control group. The remaining study on mantra meditation194 did not describe the practices being compared.

Mindfulness meditation. Four studies on mindfulness meditation used other mindfulness meditation techniques as control groups. There were two studies on MBSR,195,196 one on Zen Buddhist meditation,197 and one197 that did not describe the mindfulness techniques being compared.

Meditation practice not described. Two studies198,199 failed to provide a clear description of the meditation practices being compared.

Multiple control groups. As was shown in Table 16, 275 out of 402 intervention studies used a single control group, whereas 127 used more than one kind of control (e.g., used one active and one inactive control). Sixty-five of the intervention studies with multiple controls were conducted on mantra meditation (25 on TM®, 22 on RR, 12 on mantra techniques not further described, 4 on CSM, 1 on Acem meditation, and 1 on SRELAX). There were 26 studies with multiple controls conducted on Yoga, 20 studies on mindfulness meditation (8 on MBSR, 6 on mindfulness meditation techniques not further specified, and 6 on Zen Buddhist meditation), 7 studies on meditation practices not further described, 6 on Tai Chi, 2 on miscellaneous meditation practices, and 1 on Qi Gong.

Control groups from observational analytical studies (cohort and cross-sectional studies) were also classified according to the type of comparison used.173 As some studies used more than one control group as a comparator, the number of observational analytical studies reported below is less than the number of control groups. Table G8 in Appendix G provides the references for studies included in the description of the type of control groups in observational analytical studies along with their distribution by meditation practice.*

Unexposed controls. The vast majority of observational analytical studies (92 percent, 244/266) used comparison groups consisting of individuals that were not been exposed to any type of meditation practice. Sixty-three percent (153/244) of these studies examined mantra meditation (140 studies on TM®, 6 on mantra techniques not specified, 4 on Acem meditation, and 3 on Ananda Marga meditation). There were 29 observational analytical studies on Yoga that used a group of unexposed individuals as a comparator, 21 studies where the exposed group practiced mindfulness meditation (12 on Zen Buddhist meditation, 6 on mindfulness meditation techniques not further specified, and 3 studies on Vipassana meditation), 21 on Tai Chi, 13 on Qi Gong, 5 on miscellaneous practices combining different meditation practices, and 2 on meditation practices not further described.

Active (positive) controls using interventions other than meditation practice. Thirty-seven observational analytical studies utilized control groups consisting of practitioners of techniques other than meditation.

Exercise and other physical activities. Practitioners of exercise and other physical activities constituted the most frequent active comparator (14 studies). Four studies examined Tai Chi practitioners, four studies examined Yoga practitioners, and two studies examined subjects practicing a miscellaneous group of meditation techniques. Two studies examined TM® practitioners, one examined practitioners of meditation techniques not specified, and one examined Qi Gong practitioners. The type of physical activities practiced by the control groups included aerobic and anaerobic exercises, swimming, running, and golfing.

Miscellaneous active controls. Five studies used control groups consisting of practitioners of martial arts, concentration, and creativity techniques. Three of these studies used practitioners of mantra meditation, specifically TM®, as exposed groups. One study examined practitioners of Tai Chi and one practitioners of miscellaneous meditation techniques.

Other comparison groups consisted of individuals exposed to a variety of practices not considered meditation. Four studies on TM® used a group of practitioners of PMR as a control group. Three studies on TM® included participants that underwent hypnosis therapy. Three studies on TM® used groups of participants exposed to conditions of rest and relaxation for their comparisons. One study on Qi Gong and one on Yoga included participants in educational activities. Group therapy participants were included for comparison in one study on TM® and in one on Yoga. Individuals involved in reading activities were used as controls in one study of Zen Buddhist meditation, and in one study of Yoga. Finally, practitioners of BF and cognitive behavioral techniques such as sensitivity training acted as controls in, respectively, one study of RR and one study of TM®.

Active (positive) controls exposed to other meditation practices. Forty-seven studies used active control groups of practitioners of a variety of meditation techniques. Eleven of these studies compared groups of practitioners of different meditation techniques against each other. Thirty-six observational analytical studies compared groups of practitioners of the same meditation technique but with different lengths of practice. The former group of studies is described immediately below and the latter is described under “Concurrent control groups exposed to different dose or regimen of the same meditation practice.”

Practitioners of mantra meditation (TM® and a mantra technique not specified) were used as the comparison group in two observational studies on mindfulness meditation (one on Zen Buddhist meditation and the other on a mantra technique not further described).

There were two studies (one on TM® and the other on a mantra technique not further described) that used mindfulness meditation practitioners as control groups. Two other studies (one on Yoga, and the other on a meditation practice not described) failed to describe the type of meditation technique practiced by the comparison group. One study on Qi Gong used Tai Chi practitioners for comparisons, and Yoga practitioners were used as control groups in two studies on TM®, one on Zen Buddhist meditation, and one on Qi Gong.

Concurrent control groups exposed to different dose or regimen of the same meditation practice. Thirty-six studies made comparisons between groups of practitioners of the same meditation practice but using different lengths of practice (e.g., short-term versus long-term). Twenty of these studies were on mantra meditation (17 on TM®, 2 on Ananda Marga, and 1 on a mantra technique not further described), 6 on mindfulness meditation (4 on Zen Buddhist meditation, 1 on Vipassana meditation, and 1 on a Mindfulness meditation technique not further specified), 6 on Qi Gong, 3 on Yoga, and 1 on Tai Chi.

Historical controls. Fourteen out of 266 observational analytical studies used historical controls consisting of groups of participants external to the study or of the same single group of participants with data collected at an earlier period of time. Eleven of these studies compared mantra meditation (nine on TM® and one on Ananda Marga) to data from nonmeditators collected earlier for other purposes. Three studies on Qi Gong also used nonconcurrent data from nonpractitioners,200,201 Yoga practitioners,202 groups of athletes and participants in educational lectures.202

Multiple control groups. As shown earlier in Table 16, 207 out of 266 observational analytical studies used a single control group, whereas 59 used more than one kind of control per study (e.g., use of either active controls or inactive interventions).

Twenty-nine of the observational analytical studies with multiple controls were conducted on mantra meditation (25 on TM®, 2 on Ananda Marga, and 2 on mantra techniques not further described). There were seven studies with multiple controls conducted on mindfulness meditation (five on Zen Buddhist meditation, one on mindfulness meditation techniques not further described, and one on Vipassana meditation), seven on Yoga, seven on Qi Gong, six on Tai Chi, two on miscellaneous interventions, and one on meditation practices not further described

Meditation Practices Separated by the Diseases, Conditions, and Populations for Which They Have Been Examined

Eight hundred and thirteen studies contributed to the description of the diseases, conditions, and populations for which meditation practices have been examined.

Overall, 69 percent (n = 564) of the studies included healthy participants only, whereas 30 percent (n = 244) reported on clinical populations. Five studies (0.6 percent) included both healthy and clinical participants in the study populations. Overall, the median number of participants per study was 40 (IQR, 23 to 71), with a median age of 37 years (IQR, 26 to 50; n = 536). Both male and females were equally represented in the studies (median number of males per study, 19; IQR, 10 to 36; median number of females per study, 19; IQR, 7 to 39).

Table 21 displays the diseases, conditions, and populations that have been examined in intervention and observational analytical studies on meditation practices.

Table 21. Types of populations and conditions included in studies on meditation.

Table 21

Types of populations and conditions included in studies on meditation.

In general, the majority of studies (68 percent) on meditation practices have been conducted in healthy populations such as college and university students, healthy elderly participants from the community, army and military personnel, prison inmates, workers, athletes, and smokers (553 studies comprising 196 intervention studies and 257 observational analytical studies). Individuals with mental health disorders constituted the second most studied population (and the most frequently studied category of clinical conditions) examined in studies on meditation practices (66 studies: 65 intervention studies, and 1 observational analytical study). Mental health conditions included substance abuse, anxiety disorders, depression, and binge eating disorders, among others.

People with cardiovascular and circulatory conditions were the third most studied population and the second most frequently studied clinical condition (61 studies comprising 59 intervention studies and 2 observational analytical studies). There were 37 studies on hypertensive participants (35 intervention studies and 2 observational analytical studies). Cardiovascular conditions (24 intervention studies) included hypertension and a group of heterogeneous cardiovascular diseases (diseases of the circulatory system—the heart, the blood vessels of the heart, and the veins and arteries throughout the body and within the brain) such as coronary artery disease, chronic heart failure, ischemic heart disease, and myocardial infarction.

Forty-two studies on meditation practices (41 intervention studies and 1 observational analytical study) have been conducted in musculoskeletal conditions including chronic pain, fibromyalgia, rheumatoid arthritis, and osteoarthritis. Respiratory conditions (e.g., asthma and chronic obstructive pulmonary disease) have been examined in 16 intervention studies. Twelve intervention studies in oncology have been conducted using different types of cancer populations, such as breast, prostate, skin and lymphoma. Endocrine diseases such as type II diabetes mellitus (DM) and obesity conditions have been examined in 11 intervention studies on meditation practices. Heterogeneous patient populations with a variety of medical conditions not specified have been examined in 11 intervention studies.

Gynecological conditions such as postmenopause, menopause, premenstrual syndrome, pregnancy, and infertility have been examined in 10 intervention studies. Populations with gastrointestinal disorders have been examined in three intervention studies. Three intervention studies have examined the effect of meditation practices in dermatological disorders, such as psoriasis, and on vestibular problems, such as tinnitus. Finally, patients with dental problems (one intervention study, one observational study), end-stage renal disease (one intervention study), and organ transplants (one intervention study) have been used as study populations for studies on meditation practices.

After excluding healthy populations, the distribution of conditions or disorders for which meditation practices have been examined was

1.

hypertension (35 intervention studies and 2 observational analytical studies);

2.

other cardiovascular diseases (24 intervention studies);

3.

substance abuse disorders (18 intervention studies);

4.

anxiety disorders (14 intervention studies);

5.

cancer (12 intervention studies);

6.

asthma (11 intervention studies);

7.

chronic pain (10 intervention studies and 1 observational analytical study);

8.

type II DM (10 intervention studies);

9.

fibromyalgia (10 intervention studies); and

10.

miscellaneous psychiatric conditions (six intervention studies and one observational analytical study)

Table G9 in Appendix G* provides a comparative summary of the number and study references by meditation practice, separated by the conditions and populations for which they have been examined

Mantra meditation. Among the intervention studies on TM®, the majority (72 percent, 57/80) have been conducted in healthy populations (college and university students [24 studies], healthy volunteers from the community [19 studies], prison inmates [4 studies], elderly [3 studies], smokers [2 studies], and athletes [1 study]). The second largest group of TM® studies examined its effects on mental health disorders (nine studies) such as substance abuse (five studies), anxiety disorders (two studies), posttraumatic stress disorder (one study), and other miscellaneous psychiatric conditions (one study). Participants with circulatory or cardiovascular diseases such as hypertension (9 studies) and coronary artery disease (1 study) have been included in 10 studies on TM®. Other conditions such as asthma (two studies), chronic insomnia (one study), and a miscellaneous group of cancer patients (one study) have also been included in intervention studies on TM®.

The vast majority of observational analytical studies on TM® (98 percent, 148/151) have been conducted in healthy populations (healthy volunteers from the community [91 studies], college and university students [48 studies], prison inmates [3 studies], and workers [1 study]). Conditions such as pregnancy (one study), postmenopause (one study), and dental problems (e.g., periodontitis, one study) have been also examined.

Intervention studies on RR have included mainly healthy populations (31 studies), in addition to circulatory and cardiovascular conditions (hypertension [4 studies], other cardiovascular conditions [5 studies] including chronic heart failure, congestive heart failure, ischemic heart disease, premature ventricular contractions, and peripheral vascular disease), mental health disorders (substance abuse [2 studies], anxiety disorders [1 study], schizophrenia or antisocial personality disorders [1 study]), gynecological conditions (menopause [1 study], premenstrual syndrome [1 study]), and other clinical conditions such as irritable bowel syndrome (1 study), total knee replacement (1 study), skin cancer (1 study), and a group of patients with heterogeneous clinical conditions (1 study). The only observational analytical study on RR has been conducted in a population of hypertensive patients

Nineteen intervention studies on mantra meditation techniques not further described have been conducted with healthy populations. Other populations included people with mental health disorders (anxiety disorders [three studies], substance abuse [two studies], miscellaneous psychiatric conditions [one study]), hypertension (one study), and epilepsy (one study). The six observational analytical studies conducted on mantra techniques not further described have included healthy volunteers from the community

Seven intervention studies on CSM have been conducted on healthy populations, three on mental disorders such as anxiety disorders (one study), schizophrenia (one study), and substance abuse (one study), and another study on chronic insomnia

All the intervention and observational analytical studies on Acem meditation, Ananda Marga, Cayce's meditation, and Rosary prayer have been conducted with healthy populations

Yoga. Among the intervention studies on Yoga, more than half (80/158) have been conducted with healthy populations (healthy volunteers from the community [34 studies], college and university students [26 studies], army and military personnel [7 studies], workers [5 studies], prison inmates [4 studies], and athletes [1 study]). The second largest group of conditions studied is constituted by circulatory and cardiovascular diseases (21 studies) such as hypertension (13 studies), and other cardiovascular conditions (8 studies). Studies on Yoga have also included participants with mental health disorders (16 studies) such as depression (7 studies), anxiety disorders (3 studies), substance abuse (3 studies), other miscellaneous psychiatric conditions (2 studies), and obsessive-compulsive disorders (1 study). Respiratory and pulmonary conditions such as asthma (nine studies), chronic airways obstruction, chronic bronchitis, pleural effusion, and pulmonary tuberculosis (one study each) have been also examined. Participants with musculoskeletal conditions such as chronic pain, rheumatoid arthritis (two studies each), carpal tunnel syndrome, chronic rheumatic diseases, fibromyalgia, hyperkyphosis, multiple sclerosis, osteoarthritis, and postpolio syndrome (one study each) have been included in intervention studies on Yoga. Other conditions examined in Yoga studies were gastrointestinal disorders (two studies), epilepsy, migraine, pregnancy, human immunodeficiency virus (HIV), lymphoma, chronic insomnia, tinnitus, and heterogeneous patient populations (one study each). All the observational analytic studies on Yoga (33 studies) have been conducted with healthy populations

Mindfulness meditation. Among the 49 intervention studies on MBSR, 12 were conducted with healthy populations and 12 with populations with mental health disorders. Mental health disorders included anxiety disorders (three studies), mood disorders (two studies), substance abuse (two studies), binge eating disorders, burnout, personality disorders, miscellaneous psychiatric conditions, and stress-related conditions of parents of children with behavioral problems (one study each)

Participants with musculoskeletal conditions such as chronic pain (four studies) and fibromyalgia (two studies) have been also included. Cancer patients have been included in four intervention studies on MBSR. Other conditions such as psoriasis (two studies), cardiovascular diseases (two studies), traumatic brain injuries (two studies), obesity, HIV, and organ transplantation (one study each) have also been included. No observational analytic studies on MBSR were identified

Eleven intervention studies on mindfulness meditation not further specified have been conducted in healthy populations. Other populations included mental health disorders (binge eating disorders [two studies], anxiety disorders, psychosis, substance abuse [one study each])

Musculoskeletal conditions such as fibromyalgia (three studies) and chronic pain (two studies), cardiovascular diseases, cancer (three studies each), psoriasis, infertility, and heterogeneous patient populations (one study each) have been included also. The majority of observational analytical studies on mindfulness meditation techniques not further specified (six studies) have been conducted in healthy populations, with only one observational study conducted in a clinical population (individuals with chronic pain)

The majority of intervention studies (73 percent) on Zen Buddhist meditation have been conducted on healthy participants (11 studies). Clinical conditions that have been studied in intervention studies include hypertension (two studies), coronary artery disease, and insomnia (one study each). All the observational analytical studies conducted on Zen Buddhist meditation (13 studies) have included healthy volunteers

Three intervention studies on MBCT have included patients with a depressive disorder. Other populations that have been examined are individuals with fibromyalgia, stroke, tinnitus, and healthy workers (one study each). No observational studies on MBCT were identified.

Intervention studies on Vipassana meditation have involved healthy populations from the community and patients with migraine or tension headaches (one study each). The observational analytical studies conducted on Vipassana meditation (four studies) have employed healthy populations from the community (two studies), college and university students, and elderly individuals (one study each)

Tai Chi. Intervention studies on Tai Chi have mainly assessed healthy populations (38 studies), particularly the elderly (25 studies). Clinical conditions examined in intervention studies of Tai Chi include musculoskeletal conditions such as rheumatoid arthritis (four studies), osteoarthritis (three studies), chronic pain (two studies), balance disorders, fibromyalgia, multiple sclerosis, and osteoporosis (one study each). Circulatory and cardiovascular conditions have been examined in four studies. Other populations examined in studies on Tai Chi are menopause, postmenopause, depression, miscellaneous psychiatric conditions, developmental disabilities, stroke, type II DM, HIV, breast cancer, end-stage renal disease, and vestibulopathy (one study each). The majority (91 percent, 20/22) of the observational analytical studies conducted in Tai Chi have examined groups of healthy, elderly individuals or other healthy individuals from the community. Two observational studies have been conducted in groups of postmenopausal women

Qi Gong. Intervention studies on Qi Gong have examined populations of healthy participants (seven studies), patients with circulatory and cardiovascular disorders (hypertension [four studies], coronary artery disease [one study]), musculoskeletal conditions (fibromyalgia [two studies], muscular dystrophy and regional pain syndrome [one study each]), type II DM, substance abuse, miscellaneous medical conditions, migraine, and chronic obstructive pulmonary disease (COPD) (one study each). Almost all the observational analytical Qi Gong studies (14/15) were conducted with healthy populations; one was conducted with hypertensives

Meditation practices (ND).Among the 19 intervention studies that failed to describe the meditation practice under study, 12 examined healthy college and university students (nine studies), workers (2 studies), and healthy volunteers from the community (one study). Intervention studies on clinical conditions included patients with hypertension, dental problems, and insomnia (one study each). Two observational studies included respectively, healthy college and university students and individuals with miscellaneous psychiatric conditions

Miscellaneous meditation practices. Five of the six intervention studies that combined different meditation practices were conducted in healthy populations (three studies), miscellaneous psychiatric conditions, and heterogeneous populations of patients (one study each). One intervention study was conducted in patients with breast cancer. All five observational studies on miscellaneous meditation practices examined healthy populations

Tables 22 and 23 summarize the diseases, conditions, and populations for which meditation practices have been studied in intervention and observational analytical studies

Table 22. Intervention studies conducted on meditation practices by populations examined*.

Table 22

Intervention studies conducted on meditation practices by populations examined*.

Table 23. Observational analytical studies conducted on meditation practices by populations examined*.

Table 23

Observational analytical studies conducted on meditation practices by populations examined*.

Outcome Measures Used in Studies on Meditation Practices

In total, 3,665 outcome measures were reported in 813 studies on meditation practices. The median number of outcomes reported per study was four (IQR, 2 to 6). Table 24 displays the type of outcome measures that have been examined in studies on meditation practices

Table 24. Type of outcome measures examined in studies on meditation practices.

Table 24

Type of outcome measures examined in studies on meditation practices.

The most frequently studied outcomes were those of physiological functions (1,474 measures), followed by psychosocial outcomes (1,204 measures), outcomes related to clinical events and health status (698 measures), cognitive and neuropsychological functions (239 measures), and healthcare utilization (50 outcomes)

Studies on mantra meditation techniques reported the largest number of outcome measures (1,306 measures), followed by studies on Yoga (989 measures), mindfulness meditation techniques (567 measures), Tai Chi (489 measures), and Qi Gong (197 measures). Studies that did not describe the meditation practice under study reported 76 measures and studies that combined practices reported 41 measures

Table 25 provides a summary of the type and number of outcome measures examined by meditation practice

Table 25. Number of outcome measures examined by meditation practice.

Table 25

Number of outcome measures examined by meditation practice.

Physiological outcomes.Cardiovascular measures (495 measures) were the most frequently examined variables among the physiological outcomes. They included variables such as changes in systolic and diastolic blood pressure, heart rate, oxygen consumption, and electrocardiogram patterns. Other physiological measures frequently reported included pulmonary and respiratory outcomes (251 measures) such as respiratory rate, lung function testing measures (e.g., forced expiratory volume [FEV1], forced vital capacity [FVC], peak expiratory flow rate [PEFR]), and carbon monoxide levels). Nutritional biochemistry and metabolism outcomes (235 measures) included biochemical and metabolic processes measures that act as markers of certain diseases or conditions. These measures included serum levels of cholesterol, tryglicerides, glucose, lactate, potassium, calcium, sodium, and lipid profile

Endocrine and hormonal outcomes (125 measures) described changes in substances secreted by the endocrine system to regulate the activity of the organs. They included measures of cortisol levels, neurohormones, catecholamines, endorphines, adrenaline, and aldosterone. Brain and nervous system measures (112 measures) included electroencephalogram (EEG) profile, P300 latencies, and neurotransmitter levels. Electrodermal responses, also known as galvanic skin responses, skin conductance, and skin resistance (72 measures), included measures of the ability of the skin to conduct an electrical current as a sympathetic reaction to emotional arousal and stress. Muscular physiology (46 measures), as a proxy for emotional arousal, was examined for variables such as muscle tension and relaxation, frontal electromyographic activity, muscle voltage, and reflex function, among others. Outcomes related to the physiological functioning of the immune system (45 measures) included immunoglobulin (IgA, IgG, and IgM) concentrations, leukocytes, lymphocytes, monocytes, and neutrophil levels in general, natural killer cell activity, white blood cell count, and number of monoclonal antibodies. There were 28 outcomes related to blood products and hemodynamic parameters, 22 on thermoregulatory functions such as skin or body temperature, and 14 measures related to the skeletal system, for example, bone mineral density. Other physiological outcomes less frequently reported included ocular (e.g., intraocular pressure, pupillary dilatation) (13 measures), sensory, for example, auditory thresholds (8 measures), renal function tests (7 measures), and gastric measures, for example, gastric motility (1 measure)

Psychosocial outcomes. The most studied psychosocial outcomes were those measuring psychiatric and psychological symptoms (645 measures) of anxiety, depression, stress, mood states, irritability and anger expression, and abuse of psychoactive or other substances causing psychological dependence. Measures of personality (both normal and abnormal) were reported for 313 outcomes. These studies reported data on either general characteristics of the personality (e.g., personality and psychological profiles, ego strength, and coping styles) or particular traits or characteristics of the individual psychological functioning (e.g., locus of control, neuroticism, psychoticism, extraversion, self-actualization, self-esteem, and hostility traits). Positive psychology outcomes (measures of processes that contribute to flourishing or optimal functioning of individuals (e.g., empathy, assertive behavior, happiness, spirituality, autonomy) were reported in 108 outcomes). Outcomes related to social and interpersonal relationships such as marital adjustment, level of interpersonal conflicts, social adjustment, and social functioning, were examined in 50 measures. Health-related quality of life measures were reported for 42 outcomes. Other psychosocial outcomes included activities of daily living (26 measures), and other miscellaneous and nonspecific behavioral measures not further classified, such as “level of relaxation” and “hypnotic response.”

Clinical outcomes. Measures examining physical functions such as balance, strength, flexibility, mobility, and postural stability were the most frequently reported types of clinical outcomes (252 measures). They were followed by measures of discrete clinical events, or indicators of symptom improvement that were particular to the conditions under study, such as change in fibromyalgia symptoms, number of asthma episodes, and angina pectoris symptoms (154 measures). Outcomes related to the nutritional status or body composition of individuals (74 measures) included body weight, body mass index, and diet and nutritional patterns. There were 70 outcomes related to general health status and well-being, 55 outcomes for sleep characteristics, and 54 for pain-related symptoms. Seventeen outcomes reported on the frequency of falls or falls-related behaviors. Other clinical measures included adherence (12 measures), mortality (8 measures), and longevity (2 measures)

Cognitive and neuropsychological measures. Measures related to sensory perception and motor functions (103 measures) were the most frequently examined cognitive and neuropsychological outcomes. These measures included psychomotor performance, perceptual motor skills, field independence, absorption, autonomic arousal, and visual-spatial ability. Other cognitive and neuropsychological measures less frequently examined included reasoning and executive functions (40 measures) (e.g., cognitive flexibility, logical reasoning, thought categorization, and associate learning). General cognitive outcomes (37 measures) included global measures of intelligence, cognitive status, and neuropsychological functioning. Memory functions (e.g., short- and long-term, verbal and visual, declarative and procedural) were reported by 24 measures. Finally, language (e.g., verbal fluency, vocabulary, language comprehension, reading skills) and attention functions (e.g., concentration, sustained focusing capacity) were each reported by seven measures

Healthcare utilization: A number of outcomes addressed factors related to the use of healthcare resources, such as medication use (30 measures), length of hospital stay, medical utilization rates, number of sick leaves, and payments to the healthcare system (20 measures)

When the outcome measures were analyzed by the type of meditation practice under study, we found that the 10 most frequently reported outcome measures in mantra meditation studies were

1.

psychiatric and psychological symptoms (231 measures);

2.

physiological cardiovascular outcomes (196 measures);

3.

personality outcomes (146 measures);

4.

physiological pulmonary and respiratory outcomes (83 measures);

5.

physiological nutrition, biochemical and metabolic outcomes (76 measures);

6.

physiological brain and nervous system outcomes (73 measures);

7.

physiological electrodermal responses (53 measures);

8.

physiological endocrine and hormonal outcomes (49 measures);

9.

sensory perceptual and motor neuropsychological functions (48 measures); and

10.

positive psychology outcomes (37 measures)

There are no studies on mantra meditation practices that have reported skeletal, renal and excretory, or gastric physiology outcomes or the occurrence of falls or fall-related behaviors

The 10 most frequently reported outcome measures in studies on Yoga were

1.

physiological cardiovascular outcomes (151 measures);

2.

psychiatric and psychological symptoms (140 measures);

3.

physiological nutrition, biochemical and metabolic outcomes (110 measures);

4.

physiological pulmonary and respiratory outcomes (106 measures);

5.

personality outcomes (61 measures);

6.

clinical events and symptom improvement (61 measures);

7.

physical functionality outcomes (58 measures);

8.

physiological endocrine and hormonal outcomes (42 measures);

9.

outcomes of nutritional status and body composition (27 measures); and

10.

sensory perceptual and motor neuropsychological functions (24 measures)

No studies on Yoga reported on the occurrence of falls or fall-related behaviors, or on longevity of study participants

The 10 most frequently reported outcome measures in studies on Tai Chi were

1.

physical functionality (165 measures);

2.

physiological cardiovascular outcomes (87 measures);

3.

physiological pulmonary and respiratory outcomes (33 measures);

4.

psychiatric and psychological symptoms (33 measures);

5.

physiological nutrition, biochemical and metabolic outcomes (20 measures);

6.

clinical events and symptom improvement (17 measures);

7.

falls and fall-related behavior (16 measures);

8.

personality measures (14 measures);

9.

measures of health status and well-being (13 measures); and

10.

physiological skeletal outcomes (12 measures)

There are no studies on Tai Chi that reported physiological outcomes related to the brain and central nervous system, ocular, sensory, or gastrointestinal systems, or electrodermal response. Studies on Tai Chi have not examined outcomes related to mortality, longevity, healthcare utilization, or cognitive and neuropsychological functions such as reasoning, memory, attention, and language

The 10 most frequently outcome measures in studies on mindfulness meditation practices were

1.

psychiatric and psychological symptoms (183 measures);

2.

personality measures (66 measures);

3.

positive psychology outcomes (37 measures);

4.

clinical events and symptom improvement (31 measures);

5.

physiological cardiovascular outcomes (25 measures);

6.

measures of health status and well-being (23 measures);

7.

measures of pain and pain-related behavior (20 measures);

8.

sensory perceptual and motor neuropsychological functions (18 measures);

9.

physiological pulmonary and respiratory outcomes (14 measures); and

10.

social and interpersonal relationships measures (14 measures)

No studies on mindfulness meditation practices have reported outcomes of longevity, physiology of ocular, sensory, gastric, skeletal or renal systems, mortality, or the incidence of falls

The 10 most frequently reported outcome measures in studies on Qi Gong were

1.

physiological lymphatic and immunological outcomes (29 measures);

2.

physiological cardiovascular outcomes (27 measures);

3.

psychiatric and psychological symptoms (25 measures);

4.

physiological nutrition, biochemical and metabolic outcomes (22 measures);

5.

physiological endocrine and hormonal outcomes (15 measures);

6.

physiological pulmonary and respiratory outcomes (14 measures);

7.

personality measures (8 measures);

8.

clinical events and symptom improvement (8 measures);

9.

physical function (8 measures); and

10.

physiological brain and nervous system outcomes (8 measures)

There are no studies on Qi Gong that reported physiological outcomes related to the muscular, skeletal, ocular, sensory, and gastric systems or on electrodermal response. Other outcomes that have not been examined in studies on Qi Gong include positive psychology, interpersonal and social relationships, and cognitive functions such as memory, attention, language, and reasoning and executive functions

The 10 most studied outcome measures examined in studies that did not describe the meditation practice under study were

1.

psychiatric and psychological symptoms (20 measures);

2.

personality measures (20 measures);

3.

physiological cardiovascular outcomes (9 measures);

4.

reasoning and executive neuropsychological functions (5 measures);

5.

positive psychology outcomes (4 measures);

6.

sensory perceptual and motor neuropsychological functions (3 measures);

7.

memory (3 measures);

8.

muscular physiology (2 measures);

9.

physiological nutrition, biochemical and metabolic outcomes (2 measures); and

10.

physiological endocrine and hormonal outcomes (2 measures)

Finally, the most studied outcome measures in studies that combined miscellaneous approaches to the meditation practice were

1.

psychiatric and psychological symptoms (13 measures);

2.

personality measures (6 measures); and

3.

positive psychology outcomes (5 measures)

Summary of the Results

General remarks. Evidence regarding the state of research on the therapeutic use of meditation was provided in 813 studies. Half of the studies on meditation practices were published after 1994. Most of the studies have been published as journal articles, and have been conducted in North America. More than half of the studies have examined meditation practices in intervention studies. The majority of the intervention studies on meditation practices are RCTs, followed by before-and-after studies, and NRCTs. A lesser proportion of studies have used observational analytical designs, the majority being cohort studies, and compared groups of meditators versus nonmeditators or compared different groups of meditators

Methodological quality of the included studies. Overall, the methodological quality of both intervention and observational analytical studies on meditation practices is poor. A small proportion of RCTs reported adequately on the methods of randomization, blinding, description of withdrawals, and concealment of the sequence of allocation to treatment. Half of the RCTs explicitly reported the source of funding, as did a smaller proportion of NRCTs and before-and-after studies. The observational analytical studies that have been conducted on meditation practices are prone to biases affecting the representativeness of the study and comparison groups, the ascertainment of both exposure and outcome and, in the case of longitudinal studies (i.e., cohort studies), the integrity of the followup period. Compared to the cohort studies, the cross-sectional studies have less prominent methodological weaknesses. The only methodological aspect that did not appear to be severely jeopardized in the observational studies was the methods used to control for confounders in the design or analysis. More than half of observational studies have attempted to control for confounding either in the design or the analysis of the results

Meditation practices examined in intervention and observational analytical studies. The category of meditation practices that has been most frequently studied in the scientific literature is mantra meditation. This category includes a group of meditation techniques that, despite differences in principles of practice and theoretical grounds, all have a mantra as an important component of their practice. Both intervention and observational analytical studies on TM® dominate the literature on mantra meditation techniques, followed by studies on RR. Other mantra techniques such as CSM, Acem meditation, Ananda Marga, concentrative prayer, and Cayce's meditation have been examined less frequently.

The second category of meditation practices most frequently examined is Yoga. This category includes a heterogeneous group of practices rooted in yogic traditions such as Hatha, Kundalini, and Sahaja yoga. Mindfulness meditation is the third most studied group of practices. Within this category, MBSR and Zen Buddhist meditation have been most frequently examined. The practice of Tai Chi is the fourth most frequently examined practice, followed by Qi Gong. Finally, less than five percent of the studies on meditation practices did not explicitly describe the practice under study or have combined different approaches to meditation in a single intervention without describing the individual components of the intervention.

Control groups. The number of control groups per study ranged from one to four. Among the six hundred and sixty-eight studies that used control groups, the majority of them utilized an active concurrent control for their comparisons. Among the RCTs and NRCTs, the practice of exercise and other physical activities constituted the most frequent active comparator followed by conditions involving states of rest and relaxation, educational activities, and PMR. Other active control groups included cognitive behavioral techniques, pharmacological interventions, psychotherapy, BF techniques, reading, hypnosis, therapeutic massage, and acupuncture. Almost half of the RCTs and NRCTs included comparison groups consisting of participants assigned to waiting lists or participants that did not receive any intervention. A lower proportion of RCTs and NRCTs compared different meditation practices against each other, different doses of the practice, or modified formats of similar techniques.

The vast majority of observational analytical studies used comparison groups consisting of individuals that had not been exposed to any type of meditation practice. A smaller proportion of observational analytical studies compared groups of individuals that have been actively exposed to different meditation practices.

Diseases, conditions, and populations examined in studies on meditation practices. The vast majority of studies on meditation practices have been conducted in healthy populations. The three most studied clinical conditions are hypertension, other cardiovascular diseases, and substance abuse. Other diseases that have been frequently examined include anxiety disorders, cancer, asthma, chronic pain, type II DM, fibromyalgia, and a variety of psychiatric conditions studied altogether. Studies on hypertension have been conducted mainly on mantra meditation and Yoga. Studies on other cardiovascular diseases have been conducted using Yoga, mindfulness meditation techniques, and mantra meditation. Studies on substance abuse have been conducted mainly on mantra meditation.

Outcome measures examined in studies on meditation practices. Studies on meditation practices tend to report a median number of four outcomes per study. The most frequently studied outcomes were those of physiological functions, followed by psychosocial outcomes, outcomes related to clinical events and health status, cognitive and neuropsychological functions, and healthcare utilization outcomes. Cardiovascular measures were the most frequently examined variables among the physiological outcomes. The most studied psychosocial outcomes were measures of psychiatric and psychological symptoms (e.g., anxiety and depression). Other psychosocial outcomes frequently reported include personality measures, positive psychology outcomes, and others related to social relationships, quality of life, and activities of daily living. Outcomes related to clinical events focused on measures of physical functionality, and the incidence of discrete clinical events. Among the cognitive and neuropsychological outcomes, measures of sensory perceptual and motor functions, and reasoning and executive functions were frequently examined. Finally, measures reporting healthcare utilization were uncommon.

Topic III. Evidence on the Efficacy and Effectiveness of Meditation Practices

The three most studied diseases identified in topic II were hypertension, cardiovascular diseases, and substance abuse disorders. Sixty-five RCTs and NRCTs (27 on hypertension, 21 on cardiovascular diseases, and 17 on substance abuse disorders) were included in the review on the efficacy and effectiveness of meditation practices. All qualifying studies are presented in summary tables in the appropriate sections. Details regarding these studies are available in Appendix H.*

Hypertension

Description of the Included Studies

Twenty-seven trials (24 RCTs185,203225 and 3 NRCTs226228) were identified that evaluated the effects of meditation practices in hypertensive individuals (see Appendix H*). The included trials evaluated eight meditation practices aimed to ameliorate a variety of outcomes associated with hypertension. The group of studies comprised eight trials on yoga,185,204,212,,216,,217,,219,,224,226 five trials on TM®,205,206,210,220,221,222 four trials on RR,208,209,218,228 four trials on Qi Gong,207,211,213,214 two trials on Zen Buddhist meditation,225,227 one trial on a technique modeled after TM®,222 one trial on Tai Chi,223 one trial on a mantra technique not further described,203 and one trial on a meditation practice that did not specify the technique.215

The trials were published between 1975 and 2005 (median year of publication, 1995; IQR, 1982 to 2003). Twenty-four of these trials have been published in journals185,203,204,206209,211214,216228 while three205,210,215 were identified from the gray literature. Nine trials205,206,208210,220,221,227,228 were conducted in the United States, four204,212,217,226 in India, three185,218,219 in the United Kingdom, two211,225 in China, two213,214 in South Korea, and one each in Germany,215 Hong Kong,207 The Netherlands,224 New Zealand,222 Russia,203 Taiwan,223 and Thailand.216 The trials contained a total of 1,940 participants. The median sample size was 65 participants per study (IQR, 23 to 392; data from 19 trials). Seven203,205,206,218,220,221,225 out of 19 trials had study sample sizes greater than 100 participants. The mean age of participants was 50.7 ± 9.6 years (range, 28 to 68 years; data from 20 trials). Two trials203,227 were conducted in samples with an average age between 20 and 40 years. Sixteen trials185,205208,210,213,214,216,219,220,222226 were conducted in samples with mean ages ranging from 41 to 60 years. Two trials221,228 included study populations with mean ages of 61 years and above. Seven trials204,209,211,212,215,217,218 did not report the age of participants

When the trials that reported the gender of participants were combined (n = 23), 54 percent of the participants were male and 46 percent were female. Samples in four trials203,204,211,226 were entirely male while none of the trials included entirely female samples. Four trials185,209,212,217 failed to report the gender of participants. Six trials explicitly indicated the ethnicity of their samples. Five of them205,206,210,220,221 were conducted in African-American samples, whereas one trial227 stated that only white participants took part in the study.

All the trials were conducted in patients with a diagnosis of essential hypertension. All trials except five185,204,208,212,217 provided a definition of hypertension in their selection criteria. Half of the trials (n = 14)203,207,209,210,213216,218,220,221,225,226,228 included participants diagnosed with Stage 1 hypertension (mean systolic blood pressure [SBP] between 140 and 159 mm Hg and/or mean diastolic blood pressure [DBP] between 90 and 99 mm Hg) and with Stage 2 hypertension (mean SBP 160 mm Hg and above and/or mean DBP 100 mm Hg and above). One study206 included participants with prehypertension (mean SBP between 120 and 139 mm Hg, and/or DBP between 80 and 89 mm Hg), Stage 1, and Stage 2 hypertension. Another study205 was conducted in patients with prehypertension or with Stage 1 hypertension. Five trials211,219,222,224,227 included only participants with Stage 2 hypertension, whereas one trial223 included only participants with Stage 1 hypertension.

All 27 trials employed a parallel study design. The length of the trials varied from 8 days204 to 1 year.203,210,211,221,224 The median duration of the trials was 3 months (IQR, 2 to 6). Twelve studies204,208,209,213217,219,225,226,228 were short-term trials (less than 3 months), nine trials185,205,207,212,218,220,222,223,227 had a duration from 3 to 6 months, and six trials203,206,210,211,221,224 lasted longer than 6 months

The 27 trials comprised six comparisons between meditation practices and no intervention,185,203,204,215,217,225 four comparisons between meditation practices and waiting list,208,213,214,222 and one comparison222 between meditation practices and placebo. There were 29 comparisons between meditation practices and active therapies other than no intervention, WL, or placebo. Because some trials had more than one comparison arm, the total number of comparisons exceeded the number of trials. Of the 29 active comparisons, the comparative treatments were health education (HE),205,206,210,212,216,218,220,221,225 BF,208,209,228 PMR,204,220,221 rest or relaxation,204,219,223 antihypertensive medication,211,217 blood pressure checks,225,227 exercise,207 orthostatic tilt,226 and meditation practice plus BF.185 The median number of comparisons per study was one (IQR, 1 to 2).

Methodological Quality of the Included Studies

A summary of the methodological quality of the included trials is provided in Table 26. As a measure of methodological quality for included trials, the overall median Jadad score was 2/5 (IQR, 1 to 2). Only two trials220,221 obtained 3 points and were considered of high quality. Twelve trials185,203,206,207,210,214,216,218,219,222224 obtained 2 points, nine trials204,205,208,209,211,213,215,217,225 obtained 1 point, and four trials212,226228 did not obtain any points. All the trials except three226228 were described as randomized; however, the details of the description of randomization varied. The majority of trials (n = 19)185,203205,207211,214219,222225 did not describe how the randomization was performed. Three trials206,220,221 described an appropriate method to generate the sequence of randomization, whereas two trials212,213 reported the use of inadequate approaches to sequence generation. None of the trials were described as double-blind. The adequacy of allocation concealment was unclear in all trials.

Table 26. Methodological quality of trials of meditation practices for hypertension.

Table 26

Methodological quality of trials of meditation practices for hypertension.

An intention-to-treat statistical analysis was specified in five trials.203,206,207,220,221 Nineteen trials185,203,205207,209,210,212214,216,218225 reported the number of dropouts for the total study sample (mean dropout rate: 21 percent; range 3 to 57 percent). Seven trials205,206,209,212,213,220,225 had a dropout rate of more than 20 percent. Withdrawals and dropouts per treatment group were clearly described in 14 trials.185,203,207,210,213,214,216,218224 On average, 14 percent of participants (range 0 to 26 percent) dropped out of the meditation groups. The mean dropout rate for the control groups was also 14 percent (range 4 to 25 percent; 16 control groups).

Fifteen trials185,205209,218222,224,225,227,228 disclosed their source of funding. Nine trials 205,206,209,218220,225,227,228 received funding from government sources, six studies185,207,208,221,222,224 received funding from a private donor or foundation, and one214 received internal funding.

Results of Direct Comparisons

Table 27 summarizes the meditation practices, comparison groups, and outcomes that were available for direct meta-analyses on the efficacy and effectiveness of meditation practices to treat hypertension. Direct meta-analyses were conducted when two or more studies assessed the same meditation practice, used similar comparison groups, and had usable data for common outcomes of interest. No single diagnostic criterion was chosen for categorizing study populations as hypertensive; rather, we included all studies conducted in hypertensive patients, as defined by the authors of the primary studies. Fifteen comparisons (14 studies) were not suitable for direct meta-analyses because no more than one study was available for statistical pooling: SRELAX (technique modeled after TM®) versus waiting list (WL),222 SRELAX versus placebo,222 RR versus HE,218 RR versus WL,208 Qi Gong versus antihypertensive medication (AHM),211 Qi Gong versus exercise,207 Tai Chi versus rest,223 Yoga versus AHM,217 Yoga versus orthostatic tilt,226 Yoga versus progressive muscle relaxation (PMR),204 Yoga versus relaxation,224 Yoga versus Yoga plus BF,185 Zen Buddhist meditation versus NT,225 mantra meditation not specified versus NT,203 and meditation practice not further specified versus NT.215 Data from 16 studies were available for direct meta-analyses that involved eight comparisons: TM® versus HE, TM® versus PMR, RR versus BF, Qi Gong versus WL, Yoga versus NT, Yoga versus HE, Yoga versus rest, and Zen Buddhist meditation versus blood pressure checks. Outcomes of interest and comparisons for which data could be combined into a direct meta-analysis were

Table 27. Summary of outcomes by meditation practice and by comparison group included in meta-analyses of the efficacy and effectiveness of meditation practices for hypertension.

Table 27

Summary of outcomes by meditation practice and by comparison group included in meta-analyses of the efficacy and effectiveness of meditation practices for hypertension.

1.

blood pressure: TM® versus HE, TM® versus PMR, RR versus BF, Qi Gong versus WL, Yoga versus NT, Yoga versus HE, Zen Buddhist meditation versus blood pressure checks;

2.

body weight: TM® versus HE;

3.

heart rate: TM® versus HE;

4.

stress: TM® versus HE, Yoga versus HE;

5.

anger: TM® versus HE;

6.

self-efficacy: TM® versus HE;

7.

total cholesterol (TC): TM® versus HE;

8.

high-density lipoprotein cholesterol (HDL-C): TM® versus HE;

9.

low-density lipoprotein cholesterol (LDL-C): TM® versus HE;

10.

dietary intake (caloric intake, total fat intake, and sodium intake): TM® versus HE; and

11.

physical activity: TM® versus HE

Results from individual studies not included in a meta-analysis of clinical trials of meditation practices in hypertension are summarized in Table H1 in Appendix H.*

Transcendental Meditation®

Five RCTs assessing the effects of TM® in hypertensive patients were identified. Five trials205,206,210,220,221 compared TM® versus HE, and two trials220,221 compared TM® versus PMR. Meta-analyses were conducted for the comparisons TM® versus HE, and TM® versus PMR.

TM® versus HE

Blood pressure. Five trials205,206,210,220,221 totaling 337 participants (TM® = 175, HE = 162) provided data on the effects of TM® versus HE on SBP and DBP (Figure 3). The combined estimate of changes in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) in favor of TM® (WMD = -1.10; 95% CI, -5.24 to 3.04). There was evidence of heterogeneity among the studies regarding the mean change in SBP (p = 0.05; I2 = 56.9 percent).

Figure 3. Meta-analysis of the effect of TM® versus HE on blood pressure (SBP and DBP).

Figure

Figure 3. Meta-analysis of the effect of TM® versus HE on blood pressure (SBP and DBP).

Possible causes of heterogeneity in the outcome of SBP were explored. The five trials were similar in terms of the type of participants, severity of hypertension, characteristics of the interventions, and methodological quality. There were differences, however, in the duration of the trials and followup period. All but one study221 were medium- or long-term trials (more than 3 months). The study with the shortest duration221 (3 months) was the only trial that reported statistically significant changes in SBP favoring TM®. The medium- or long-term trials did not find statistically significant differences between TM® and HE for changes in SBP. A subgroup analysis based on the duration of the studies (Figure 4) showed that greater homogeneity (p = 0.64, I2 = 0 percent) was observed for the studies that assessed the medium- and long-term effects of TM® and HE on SBP. After excluding the short-term study,221 the direction of the effect changed to a small, nonsignificant reduction of SBP in favor of HE (WMD = 0.70; 95% CI, -2.29 to 3.68).

Figure 4. Subgroup analysis by study duration of the effect of TM® versus HE on SBP.

Figure

Figure 4. Subgroup analysis by study duration of the effect of TM® versus HE on SBP.

The combined estimate of changes in DBP (mm Hg) indicated a small, nonsignificant improvement (reduction) in favor of TM® (WMD = -0.58; 95% CI, -4.22 to 3.06). We found significant heterogeneity (p = 0.003; I2 = 74.8 percent) among the studies for this outcome, which may be attributed to variations in the duration of the studies. The study with the shortest duration221 (3 months) was the only trial that reported statistically significant changes in DBP favoring TM®. The other medium- or long-term trials did not find statistically significant differences between TM® and HE for changes in DBP. A subgroup analysis based on the duration of the studies (Figure 5) showed that greater homogeneity (p = 0.26, I2 = 25.2 percent) was observed for the studies assessing the medium- and long-term effects of TM® and HE on DBP. After excluding the short-term study,221 the magnitude of the effect estimate changed to a small, nonsignificant reduction of DBP in favor of HE (WMD = 1.02; 95% CI, -1.41 to 3.44).

Figure 5. Subgroup analysis by study duration of the effect of TM® versus HE on DBP.

Figure

Figure 5. Subgroup analysis by study duration of the effect of TM® versus HE on DBP.

Body weight. Three trials205,206,210 totaling 166 participants (TM® = 86, HE = 80) provided data on the effects of TM® versus HE on changes in body weight (lbs) (Figure 6). The results of the trials for changes in body weight were homogeneous (p = 0.96; I2 = 0 percent), and the combined WMD of 1.72 (95% CI, -2.29 to 5.74) showed a greater nonsignificant improvement (reduction) in body weight in favor of HE.

Figure 6. Meta-analysis of the effect of TM® versus HE on body weight.

Figure

Figure 6. Meta-analysis of the effect of TM® versus HE on body weight.

Heart rate. Three trials205,206,210 totaling 165 participants (TM® = 85, HE = 80) provided data on the effects of TM® versus HE on heart rate (bpm) (Figure 7). The results were statistically homogeneous (p = 0.34; I2 = 8.3 percent). The combined WMD of -0.43 (95% CI, -4.17 to 3.31) indicated a small, nonsignificant reduction in pulse rate with TM®.

Figure 7. Meta-analysis of the effect of TM® versus HE on heart rate.

Figure

Figure 7. Meta-analysis of the effect of TM® versus HE on heart rate.

Stress. Two trials205,210 totaling 105 participants (TM® = 54, HE = 51) contributed data on the effects of TM® versus HE on measures of stress (Figure 8). The combined estimate (SMD = 0.12; 95% CI, -0.27 to 0.50) indicated a small, nonsignificant reduction in stress scores with HE. There was evidence of homogeneity between the studies regarding the outcome of stress (p = 0.38; I2 = 0 percent).

Figure 8. Meta-analysis of the effect of TM® versus HE on measures of stress.

Figure

Figure 8. Meta-analysis of the effect of TM® versus HE on measures of stress.

Anger. Two trials205,210 totaling 105 participants (TM® = 54, HE = 51) examined the effects of TM® versus HE on measures of anger (Figure 9). The results of the trials for changes in measures of anger were homogeneous (p = 0.64; I2 = 0 percent), and the combined SMD of -0.06 (95% CI, -0.45 to 0.32) showed a small and nonsignificant reduction in scores of anger with TM®

Figure 9. Meta-analysis of the effect of TM® versus HE on measures of anger.

Figure

Figure 9. Meta-analysis of the effect of TM® versus HE on measures of anger.

Self-efficacy. Data on changes in measures of self-efficacy were available from two trials205,210 with a total of 105 participants (TM® = 54, HE = 51) (Figure 10). The combined SMD in measures of self-efficacy for trials of TM® compared with HE was -0.36 (95% CI, -0.92 to 0.19), and showed a nonsignificant improvement in self-efficacy in favor of TM®. The results of the trials for changes in self-efficacy were moderately heterogeneous (p = 0.18; I2 = 44.8 percent).

Figure 10. Meta-analysis of the effect of TM® versus HE on measures of self-efficacy.

Figure

Figure 10. Meta-analysis of the effect of TM® versus HE on measures of self-efficacy.

Total cholesterol (TC). Information on TC changes (mg/dL) was available from two trials205,206 with a total of 126 participants (TM® = 65, HE = 61) (Figure 11). The combined effect estimate showed no differences between TM® and HE in TC changes (WMD = -0.94; 95% CI, -11.49 to 9.62). The results of the trials were homogeneous (p = 0.80; I2 = 0 percent).

Figure 11. Meta-analysis of the effect of TM® versus HE on TC.

Figure

Figure 11. Meta-analysis of the effect of TM® versus HE on TC.

High-density lipoprotein cholesterol (HDL-C). Two trials205,206 totaling 126 participants (TM® = 65, HE = 61) provided data on the effects of TM® versus HE on changes in HDL-C (mg/dL) (Figure 12). The results of the trials were homogeneous (p = 0.35; I2 = 0 percent), and the combined WMD of -2.58 (95% CI, -6.12 to 0.96) showed a nonsignificant benefit (increase) with HE for HDL-C.

Figure 12. Meta-analysis of the effect of TM® versus HE on HDL-C.

Figure

Figure 12. Meta-analysis of the effect of TM® versus HE on HDL-C.

Low-density lipoprotein cholesterol (LDL-C). Two trials205,206 totaling 126 participants (TM® = 65, HE = 61) contributed data on the effects of TM® versus HE on changes in LDL-C (mg/dL) (Figure 13). The pooled results of the trials were homogeneous (p = 0.90; I2 = 0 percent), and the combined WMD of 1.08 (95% CI, -8.65 to 10.81) showed a nonsignificant benefit (reduction) with HE for LDL-C.

Figure 13. Meta-analysis of the effect of TM® versus HE on LDL-C.

Figure

Figure 13. Meta-analysis of the effect of TM® versus HE on LDL-C.

Dietary intake. Two trials205,210 totaling 49 participants (TM® = 30, HE = 19) provided data on the effects of TM® versus HE on dietary intake, expressed as caloric intake, total fat intake, and sodium intake (Figure 14). The results of the trials for caloric intake were homogeneous (p = 0.97; I2 = 0 percent), and the combined SMD of 0.28 (95% CI, -0.30 to 0.86) showed a nonsignificant reduction in caloric intake in the HE group. The results of the trials for total fat intake were homogeneous (p = 0.23; I2 = 30.7 percent), and the combined SMD of 0.50 (95% CI, -0.21 to 1.21) showed a nonsignificant reduction in fat intake in the HE group. The results of the trials for sodium intake were homogeneous (p = 0.64; I2 = 0 percent), and the combined SMD of 0.14 (95% CI, -0.44 to 0.72) showed a nonsignificant reduction in sodium intake in the HE group.

Figure 14. Meta-analysis of the effect of TM® versus HE on dietary intake.

Figure

Figure 14. Meta-analysis of the effect of TM® versus HE on dietary intake.

Physical activity. Three trials205,206,210 totaling 138 participants (TM® = 68, HE = 70) provided data on the effects of TM® versus HE on changes in physical activity (Figure 15). The combined results showed a nonsignificant reduction in changes in favor of the HE group (SMD = -0.20; 95% CI, -0.14 to 0.53). The results of the trials for changes in physical activity were homogeneous (p = 0.57; I2 = 0 percent).

Figure 15. Meta-analysis of the effect of TM® versus HE on physical activity.

Figure

Figure 15. Meta-analysis of the effect of TM® versus HE on physical activity.

TM® versus PMR

Blood pressure. Two trials220,221 totaling 179 participants (TM® = 90, PMR = 89) provided data on the effects of TM® versus PMR on SBP and DBP (Figure 16). The combined estimate of changes in SBP (mm Hg) indicated a significant improvement (reduction) in favor of TM® (WMD = -4.30; 95% CI, -8.02 to -0.57). The results of the trials for changes in SBP were homogeneous (p = 0.25; I2 = 25.6 percent)

Figure 16. Meta-analysis of the effect of TM® versus PMR on blood pressure (SBP and DBP).

Figure

Figure 16. Meta-analysis of the effect of TM® versus PMR on blood pressure (SBP and DBP).

The combined estimate of changes in DBP (mm Hg) indicated a significant improvement (reduction) in favor of TM® (WMD = -3.11; 95% CI, -5.00 to -1.22). The results of the trials for changes in DBP were homogeneous (p = 0.67; I2 = 0 percent).

Relaxation Response

Five trials assessing the effects of RR in hypertensive patients were identified. Three trials208,209,228 compared RR versus BF, one trial compared RR versus HE,218 and one trial compared RR versus WL.208 A meta-analysis was conducted for the comparison between RR and BF

RR versus BF

Blood pressure. Three trials208,209,228 totaling 53 participants (RR = 28, BF = 25) provided data for a meta-analysis of the effects of RR versus BF on SBP and DBP (Figure 17). The combined estimate of changes in SBP (mm Hg) showed that BF produced a greater but nonsignificant reduction in SBP when compared to RR (WMD = 2.39; 95% CI, -5.13 to 9.91). The results were homogeneous across the trials (p = 0.55; I2 = 0 percent). Likewise, the combined estimate of changes in DBP (mm Hg) indicated a small, nonsignificant improvement (reduction) in favor of BF (WMD = 4.44; 95% CI, -4.00 to 12.88). The results of the trials for changes in DBP were homogeneous (p = 0.42; I2 = 0 percent).

Figure 17. Meta-analysis of the effect of RR versus BF on blood pressure (SBP and DBP).

Figure

Figure 17. Meta-analysis of the effect of RR versus BF on blood pressure (SBP and DBP).

Qi Gong

Four trials assessing the effects of Qi Gong in hypertensive patients were identified. Two trials213,214 compared Qi Gong versus WL, one trial compared Qi Gong versus AHM,211 and another trial207 compared Qi Gong versus exercise. A meta-analysis was conducted for the comparison between Qi Gong and WL

Qi Gong versus WL

Blood pressure. Two trials213,214 totaling 94 participants (Qi Gong = 46, WL = 48) provided data for a meta-analysis of the effects of Qi Gong versus WL on SBP and DBP (Figure 18). The combined estimate of changes in SBP (mm Hg) indicated a significant improvement (reduction) in favor of Qi Gong (WMD = -17.78; 95% CI, -22.03 to -13.54). The results were homogeneous across the trials (p = 0.57; I2 = 0 percent)

Figure 18. Meta-analysis of the effect of Qi Gong versus WL on blood pressure (SBP and DBP).

Figure

Figure 18. Meta-analysis of the effect of Qi Gong versus WL on blood pressure (SBP and DBP).

Likewise, the combined estimate of changes in DBP (mm Hg) indicated a significant improvement (reduction) of DBP in favor of Qi Gong (WMD = -12.06; 95% CI, -21.62 to -2.49). There was evidence of substantial heterogeneity among the studies in DBP (p <0.00001; I2 = 93.5 percent). Possible causes of heterogeneity were explored. The two trials were similar in terms of the type of participants, severity of hypertension, characteristics of the interventions, study duration, and methodological quality. Therefore, it is unknown whether clinical heterogeneity produced statistical heterogeneity between the trials for the outcome of DBP. Although each trial showed the same direction of effect, the wide confidence intervals indicate that the estimates of effect are unreliable and consistent with a broad range of possible effect sizes. Therefore, heterogeneity obscures the clinical applicability of the WMD in the analysis

Yoga

Eight trials185,204,212,216,217,219,224,226 assessing the effects of Yoga in hypertensive patients were identified. Three trials185,204,217 compared Yoga versus NT, two trials212,216 compared Yoga versus HE, two trials204,219 compared Yoga versus rest, one trial217 compared Yoga versus AHM, one trial226 compared Yoga versus orthostatic tilt, one trial204 compared Yoga versus PMR, one trial224 compared Yoga versus relaxation, and one trial185 compared Yoga versus a combination of Yoga and BF. Meta-analyses were conducted for the comparisons of Yoga versus NT, Yoga versus HE, and Yoga versus rest.

Yoga versus NT

Blood pressure. Three trials185,204,217 totaling 57 participants (Yoga = 28, NT = 29) provided data for a meta-analysis of the effects of Yoga versus NT on SBP and DBP (Figure 19). The combined estimate of changes in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) in favor of Yoga (WMD = -15.39; 95% CI, -31.97 to 1.19). There was evidence of significant (p = 0.006) and substantial (I2 = 80.2 percent) heterogeneity among the studies regarding the mean change in SBP

Figure 19. Meta-analysis of the effect of Yoga versus NT on blood pressure (SBP and DBP).

Figure

Figure 19. Meta-analysis of the effect of Yoga versus NT on blood pressure (SBP and DBP).

Possible causes of heterogeneity in the outcome of SBP were explored. The three trials were similar in duration and methodological quality. The studies failed to appropriately report some important characteristics that would have been useful for appraising the potential sources of heterogeneity in the trials. Age of participants was similar in two studies,204,217 while the remaining study185 failed to provide this information. The distribution of males and females for the total study population was also unknown in two185,217 of the three trials. None of the studies provided a critical value for the presence or severity of hypertension. Treatment in the Broota study204 consisted of practicing Shavasana consecutively for 8 days, with each session lasting 20 minutes. The intervention group in the trial of Hafner185 practiced Yoga for eight 1-hour sessions at weekly intervals. Finally, participants in the Yoga group in the study of Murugesan217 engaged in a variety of yogic practices (i.e., asanas, Om recitation, and meditation) twice a day for 1 hour, 6 days a week.

The most obvious difference among the three studies was that control participants in the Broota204 and Hafner185 trials were assigned to a NT condition in which existing medical treatment was not interrupted, whereas controls in the trial of Murugesan217 did not receive any therapy. Therefore, it is likely that the conditions of NT in the Murugesan217 study were systematically different from the other two studies. Yoga was used as an adjuvant therapy in the studies of Broota204 and Hafner185 whereas in the Murugesan trial217 it was not. Murugesan217 was the only study to report statistically significant results in favor of Yoga for changes in SBP and DBP.

A subgroup analysis by concomitant treatment (Figure 20) showed that greater homogeneity (p = 0.86, I2 = 0 percent) was observed for the studies that continued medical therapy in the NT condition. After excluding the study that did not provide any therapy,217 the direction of the effect did not change, and a nonsignificant improvement (reduction) in favor of Yoga was found for SBP (WMD = -7.15; 95% CI, -17.70 to 3.39).

Figure 20. Subgroup analysis by concomitant therapy of Yoga versus NT on SBP.

Figure

Figure 20. Subgroup analysis by concomitant therapy of Yoga versus NT on SBP.

As depicted in Figure 19, the combined estimate of changes in DBP (mm Hg) indicated a significant improvement (reduction) in favor of Yoga (WMD = -13.95; 95% CI, -27.24 to -.0.66) There was evidence of significant heterogeneity (p = 0.01; I2 = 76.5 percent) among the studies for this outcome, which may be accounted for by the use of concomitant therapy in the NT condition. A subgroup analysis based on the presence of concomitant treatment (Figure 21) showed that homogeneity (p = 0.44, I2 = 0 percent) was observed for the studies that did not interrupt existing medical therapy for the NT condition. After excluding the study that did not provide any concomitant therapy,217 the results remained similar, and a nonsignificant improvement (reduction) in favor of Yoga was found for DBP (WMD = -6.82; 95% CI, -15.51 to 1.87).

Figure 21. Subgroup analysis by concomitant therapy of Yoga versus NT on DBP.

Figure

Figure 21. Subgroup analysis by concomitant therapy of Yoga versus NT on DBP.

Yoga versus HE

Blood pressure. Two trials208,212,216 totaling 68 participants (Yoga = 34, HE = 34) provided data on the effects of Yoga versus HE on SBP and DBP (Figure 22). The combined estimate of changes in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) in favor of Yoga (WMD = -15.32; 95% CI, -38.77 to 8.14). There was evidence of heterogeneity between the studies regarding the mean change in SBP (p = 0.001; I2 = 90.3 percent). Possible causes of heterogeneity in the outcome of SBP were explored. The studies failed to report appropriately some important characteristics that would have been useful for appraising the potential sources of heterogeneity. The two trials were similar in terms of the type of participants, and methodological quality. There were differences in the duration of trials that may explain the differences in the results from the individual studies, and the heterogeneity in the pooling of the results. The Latha212study was a medium-term trial lasting 6 months, whereas the McCaffrey216 study was a short-term trial of 11 weeks. The short-term trial reported statistically significant changes in SBP for Yoga as compared to HE, whereas the effects seem to disappear at medium term, as reported by the statistically nonsignificant results of the McCaffrey216 trial

Figure 22. Meta-analysis of the effect of Yoga versus HE on blood pressure (SBP and DBP).

Figure

Figure 22. Meta-analysis of the effect of Yoga versus HE on blood pressure (SBP and DBP).

The combined estimate of changes in DBP (mm Hg) indicated a nonsignificant improvement (reduction) in favor of Yoga (WMD = -11.35; 95% CI, -30.17 to 7.47). There was evidence of significant heterogeneity (p = 0.01; I2 = 84.0 percent) between the studies for this outcome, which may be primarily accounted for by the duration of the trials. The difference in the significance of the individual study results may be a function of the duration of the trials, with the short-term trial212showing statistically significant changes in DBP, and the medium term trial reporting nonstatistically significant results

Stress. Two trials208,216 totaling 68 participants (Yoga = 34, HE = 34) examined the effects of Yoga versus HE on measures of stress (Figure 23). The results of the trials for changes in measures of stress were homogeneous (p = 0.59; I2 = 0 percent), and the combined SMD of -1.10 (95% CI, -1.61 to -0.58) showed a statistically significant reduction in scores of stress with Yoga.

Figure 23. Meta-analysis of the effect of Yoga versus HE on stress.

Figure

Figure 23. Meta-analysis of the effect of Yoga versus HE on stress.

Zen Buddhist meditation

Two trials225,227 assessing the effects of Zen Buddhist meditation in hypertensive patients were identified. The two trials were included in a meta-analysis comparing Zen Buddhist meditation versus blood pressure checks.

Zen Buddhist meditation versus blood pressure checks

Blood pressure. Two trials225,227 totaling 250 participants (Zen Buddhist meditation = 134, blood pressure checks = 116) provided data for a meta-analysis of the effects of Zen Buddhist meditation versus blood pressure checks on SBP and DBP (Figure 24). The combined estimate of changes in SBP (mm Hg) indicated a nonsignificant improvement (reduction) in favor of Zen Buddhist meditation (WMD = -3.67; 95% CI, -9.04 to 1.70). The results were homogeneous (p = 0.34; I2 = 0 percent). The combined estimate of changes in DBP (mm Hg) indicated a significant improvement (reduction) in favor of Zen Buddhist meditation (WMD = -6.08; 95% CI, -11.68 to -0.48). The results of the trials for changes in DBP were moderately homogeneous (p = 0.15; I2 = 52.4 percent)

Figure 24. Meta-analysis of the effect of Zen Buddhist meditation versus blood pressure checks on blood pressure (SBP and DBP).

Figure

Figure 24. Meta-analysis of the effect of Zen Buddhist meditation versus blood pressure checks on blood pressure (SBP and DBP).

Mixed Treatment and Indirect Comparisons

Blood pressure. Since many of the studies of meditation practices in hypertensive patients reported data on SBP and DBP, we were able to do a mixed treatment analysis56 which allowed us to compare all interventions to one another.

SBP.Table 28 and Figure 25 show the results of the mixed treatment comparisons for SBP, ordered by the point estimate of difference from NT. The interventions ranged from reducing SBP from an average of 0.3 to 21.9 mm Hg. Tai Chi, Yoga plus BF, and Qi Gong seem to be more effective than the other interventions in terms of point estimates and likelihood of being the best intervention. However, we cannot make strong inferences on which is the best intervention due to a lack of statistical power.

Table 28. Mixed treatment comparisons on SBP (mm Hg) reductions compared to NT.

Table 28

Mixed treatment comparisons on SBP (mm Hg) reductions compared to NT.

Figure 25. SBP results (point estimate and 95% credible interval) for all intervention based on mixed treatment comparisons.

Figure

Figure 25. SBP results (point estimate and 95% credible interval) for all intervention based on mixed treatment comparisons.

Tai Chi, Yoga plus BF, and Yoga alone all reduced SBP significantly compared to NT. Yoga, Tai Chi, and Yoga plus BF were also found to be significantly superior to HE, while Qi Gong was significantly superior to a WL control (not shown). No other pair-wise comparisons were statistically significant.

DBP0.Table 29 and Figure 26 show the results of the mixed treatment comparisons for DBP, ordered by the point estimate of difference from NT. Note that the study215 that reported on the CMBT intervention did not report DBP and was excluded from this analysis, giving us one less intervention than the SBP analysis. The interventions ranged from reducing DBP from an average of 1.0 to 17.1 mm Hg. Yoga plus BF and Qi Gong were slightly above the other interventions in terms of point estimates and likelihood of being the best intervention, although the differences between interventions were even less than for SBP

Table 29. Mixed treatment comparisons on SBP (mm Hg) reductions compared to NT.

Table 29

Mixed treatment comparisons on SBP (mm Hg) reductions compared to NT.

Figure 26. DBP results (point estimate and 95% credible interval) for all interventions based on mixed treatment comparisons.

Figure

Figure 26. DBP results (point estimate and 95% credible interval) for all interventions based on mixed treatment comparisons.

Yoga alone and Yoga plus BF were the only interventions that reduced DBP significantly compared to NT. The only other pair-wise comparisons (not shown) that were statistically significant were Yoga compared to HE and Qi Gong compared to WL

Other indirect comparisons. We were able to make indirect comparisons between TM® and Yoga via HE for body mass index (BMI), heart rate, and stress. Yoga was nonsignificantly superior to TM® in reducing BMI (MD: -0.69; 95% CI, -2.53 to 1.15) and significantly superior in reducing both heart rate (MD: -15.6 bpm; 95% CI, -21.7 to -9.6) and stress (MD: -0.95; 95% CI, -1.76 to -0.14).

We were also able to make an indirect comparison of TM® versus RR in reducing cigarette smoking via direct comparisons with HE. RR was found to significantly reduce smoking compared to TM® (MD: -2.8; 95% CI, 0.3 to 5.4).

Analysis of Publication Bias

Because of the very small number of trials available for each comparison, the statistical tests lacked the power to detect publication bias. Therefore the analysis of the effect of publication bias on the meta-analyses presented above was not conducted.

Cardiovascular Diseases

Description of the Included Studies

Twenty-one trials (15 RCTs91,233246 and 6 NRCTs247252) that evaluated the effects of meditation practices in individuals with cardiovascular diseases were identified. They included seven trials on Yoga,233,238240,247,250,251 three on Tai Chi,235,246,248 three on RR,91,234,236 three on mindfulness meditation (not specified),241,242,249 two on MBSR,244,245 one on Qi Gong,243 one on TM®,252 and one on Zen Buddhist meditation.237

The trials were published between 1988 and 2005 (median year of publication: 2002; IQR, 1998 to 2004). Fifteen of these trials have been published in journals233236,238240,243,244,246,248252 while six91,237,241,242,245,247 were identified from the gray literature. Eleven trials91,233,234,237,241,242,244246,251,252 were conducted in the United States, three239,240,250 in India, one236 in Brazil, one249 in China, one247 in Germany, one243 in Sweden, one248 in Taiwan, and one238 in Thailand. Characteristics of the trials are summarized in Table H2 in Appendix H.*A total of 1,358 individuals were assigned to meditation practices or control groups. The median sample size based on data from 20 trials was 48 participants per study (IQR, 31 to 106). Five235,242,243,249,250 of 20 trials that provided data on sample size had more than 100 participants assigned to the study groups. The mean age of participants based on data from 17 trials was 63 ± 7 years (range: 52 to 77 years). Eight trials91,235,238,240,241,244,248,252 were conducted in samples with mean ages ranging from 41 to 60 years. Nine trials233,234,236,237,243,246,247,249,251 included study populations with ages above 61 years. Four trials239,242,245,250 did not report on the age of participants.

Across all the trials that reported the gender of participants (n = 17), 70 percent were males and 30 percent were females. The samples in three trials240,248,252 were entirely male while samples in two trials233,244 were entirely female. Four trials239,242,245,250 did not report the gender of participants. Five trials91,234,241,244,246 explicitly indicated the race or ethnicity of their samples. Around 80 percent of their samples consisted of Caucasian participants, except for one trial249 that involved Asian subjects only.

Twelve studies237241,243245,247,249,250,252 were conducted in patients with coronary artery disease (CAD), as described by the primary study authors. Clinical conditions included history of myocardial infarction (MI), chronic stable angina, valve diseases, and arrhythmias. CAD diagnoses were confirmed either by angiography,237,238,240,252 clinical history,241,243,244,249 or combining both clinical history and electrocardiogram.239

Three studies245,247,250 failed to provide a description of the diagnosis criteria for inclusion in the trials. Three studies 233,242,251 were conducted in patients with coronary heart disease.

Three studies246,234,236 were conducted in patients with chronic hearth failure (CHF). Patients from one of the studies246 on CHF met the functional capacity criteria for New York Heart Association (NYHA) classification I–IV. Patients in another study on CHF234 met the criteria for NYHA functional class II–III. The remaining study on CHF236 included patients that met both the Vasan and Lecy criteria for CHF, and the criteria for NYHA functional class I–II. Other cardiovascular conditions that were studied included acute myocardial infarction (AMI),235 and peripheral vascular occlusive disease.91 Finally, one study248 was conducted in patients that underwent coronary artery bypass surgery.

All 21 trials employed a parallel study design. The length of the trials varied from 90 minutes237 to 1 year.240,248250 The median duration of the trials was 3 months (IQR, 2 to 9; data from 20 trials). Six studies235,237,241,244,251,252 were short-term trials (less than 3 months in duration), nine trials233,234,236,238,239,243,245247 were between 3 and 6 months, and five trials240,242,248250 were longer than 6 months.

The 21 trials comprised 5 comparisons between meditation practices and no intervention,235,244,245,247,251 and one comparison between meditation and WL.252 There were 20 comparisons between meditation and active therapies other than no intervention or WL. As some trials had more than one comparison arm, the total number of comparisons exceeds the number of trials. The 20 active comparisons comprise exercise,233,235,239,240,248 HE,234,237,243 usual care,234,242,246,249 group therapy,236,241 pharmacological interventions,238,246,250 rest,91 listening to music,91 and cognitive restructuring training.241 Four studies were three-arm trials91,234,235,241 while the remaining 17 were two-arm trials.

Methodological Quality of Included Studies

The methodological quality of the included trials as measured by the overall median Jadad score was 1/5 (IQR, 1 to 2). Two trials91,234 obtained 3 points and were considered of high quality (i.e., Jadad scores greater than or equal to 3 points). Seven trials235,236,238,241,243,244,246 obtained 2 points, 11 trials233,237,239,240,242,245,248252 obtained 1 point, and one trial247 did not obtain any points. All the trials except six247252 were described as randomized; however, the description of randomization varied. The majority of trials (n = 13)235246,253 did not provide a description on how the randomization was performed. Two trials91,234 described appropriate methods of generating the sequence of randomization. None of the trials were described as double-blind. The adequacy of allocation concealment was unclear in all the trials except one.246

An intention-to-treat analysis was specified in two trials only.234,246 Sixteen trials91,233238,241,243,244,246,248252 reported dropout information for the total study sample (mean dropout rate, 17 percent; range, 0 to 32 percent). Six trials236,241,248,249,251,252 had a dropout rate of more than 20 percent. Withdrawals and dropouts per treatment group were clearly described in 14 trials91,234236,238,241,243,244,246,248252 On average, 20 percent of participants (range, 0 to 39 percent) dropped out from the meditation groups in the 14 studies that reported dropouts. The mean dropout rate for the control groups was slightly lower (16 percent; range, 0 to 33 percent).

Eight trials233,234,239,240,243,246,248,249 disclosed their source of funding. Seven trials233,234,239,240,243,246,249 received funding from government sources; two243,246 received funds from a private donor/foundation; and one248 received internal funds. A comparative summary of the methodological quality of the included trials is provided in Table 30

Table 30. Methodological quality of trials of meditation practices for other cardiovascular disorders.

Table 30

Methodological quality of trials of meditation practices for other cardiovascular disorders.

Results of Direct Comparisons

Table 31 summarizes the type of meditation practice, comparison group, and outcomes that were available for direct meta-analyses on the efficacy and effectiveness of meditation practices to treat cardiovascular diseases. No single diagnostic criterion was chosen for categorizing study populations; rather, we included all studies conducted in patients with cardiovascular disorders, as defined by the authors of the primary studies. Direct meta-analyses were conducted when two or more studies assessed the same type of meditation practice, used similar comparison groups, and had usable data for common outcomes of interest. Briefly, the majority of the comparisons from 14 studies (16 out of 18 comparisons) were not suitable for direct meta-analyses. Common clinical outcomes were absent for the following comparisons: MBSR versus NT,244,245 mindfulness techniques not specified versus usual care,242,249 Yoga versus exercise,233,239,240 and Yoga versus NT.247,251 No more than one study was available for statistical pooling of the results for mindfulness techniques not specified versus cognitive restructuring training,241 mindfulness techniques not specified versus group therapy,241 Qi Gong versus HE,243 RR versus HE,234 RR versus group therapy,236 RR versus music,91 RR versus usual care,234 RR versus rest,91 Tai Chi versus HE,235 Tai Chi versus usual care,246 and Zen Buddhist meditation versus HE.237

Table 31. Summary of outcomes by meditation practice and by comparison group included in meta-analyses of the efficacy and effectiveness of meditation practices in cardiovascular diseases.

Table 31

Summary of outcomes by meditation practice and by comparison group included in meta-analyses of the efficacy and effectiveness of meditation practices in cardiovascular diseases.

Data from six studies were available for direct meta-analyses to compare Tai Chi versus exercise, Yoga versus medication, and Yoga versus exercise. Outcomes of interest for which data could be combined into a direct meta-analysis were

1.

heart rate; Tai Chi versus exercise;

2.

total cholesterol (TC); Yoga versus medication;

3.

low-density lipoprotein cholesterol (LDL-C); Yoga versus medication; and

4.

body weight; Yoga versus exercise

Results from individual studies that were not included in a direct meta-analysis of clinical trials of meditation practices in cardiovascular are summarized in Table H2 in Appendix H.*

Tai Chi versus exercise

Heart rate. Two trials235,248 totaling 99 participants (Tai Chi = 47, exercise = 52) provided data on the effects of Tai Chi versus exercise on heart rate (HR). After analyzing the substantial heterogeneity of the studies (I2 = 70 percent), it was considered inappropriate to combine the study results into a single effect estimate. There were substantial differences between the two studies regarding the characteristics of participants in the studies, the methods to evaluate HR, study design, and the duration of the followup period. The study by Channer235 was an 8-week RCT conducted in patients who had suffered acute MI within 3 weeks prior to enrolling in the trial. Measures of HR were taken at rest. Individual study results showed a significant benefit (reduction) in resting heart rate that favored Tai Chi over exercise. The study of Lan248 was a 1-year NRCT conducted in patients that underwent coronary artery bypass surgery. Measures of HR were taken during exercise. Individual study results showed a nonsignificant improvement (increase) in HR during exercise as compared to Tai Chi.

Yoga versus lipid lowering medication (LLM)

Total cholesterol (TC). Two trials238,250 totaling 157 participants (Yoga = 93, LLM = 64) provided data on the effects of Yoga versus LLM on TC. After analyzing the substantial heterogeneity of the studies (I2 = 97.4 percent), it was considered inappropriate to combine the study results into a single effect estimate. There were substantial differences between the two studies regarding the characteristics of participants in the studies, study design, and the duration of the followup period. The study of Jatuporn238 was a 4-month RCT conducted in patients with coronary artery disease that compared the practice of Yoga and the administration of LLM. Individual study results showed a significant benefit (reduction) over the short-term in TC that favored LLM over Yoga. The study of Yogendra250 was a 1-year NRCT conducted in patients with coronary artery disease that compared Yoga versus LLT. Individual study results showed a nonsignificant improvement (reduction) over the long-term in TC that favored Yoga over LLM.

Low-density lipoprotein cholesterol (LDL-C). Two trials238,250 totaling 157 participants (Yoga = 93, LLM = 64) provided data on the effects of Yoga versus LLM on TC. As mentioned before, there was considerable clinical heterogeneity between the studies (I2 = 97.3 percent) that precluded the pooling of the results. The short-term RCT of Jatuporn238 reported a significant reduction in LDL-C with LLM. The long-term NRCT of Yogendra250 showed a nonsignificant decrease in LDL-C that favored Yoga over LLM.

Yoga versus exercise

Body weight. Two trials233,239 totaling 95 participants (Yoga = 51, exercise = 44) provided data on the effects of Yoga versus exercise on body weight changes (Figure 27). The combined estimate of changes in body weight (kg) indicated a nonsignificant improvement (reduction) in favor of Yoga (WMD = -2.14; 95% CI, -7.30 to 3.02). The results were statistically homogeneous (p = 0.61; I2 = 0 percent).

Figure 27. Meta-analysis of the effect of Yoga versus exercise on body weight.

Figure

Figure 27. Meta-analysis of the effect of Yoga versus exercise on body weight.

Indirect Comparisons

We were able to indirectly compare changes in measures of anxiety in Yoga versus MBSR (i.e., each was compared to NT in separate studies). There was no significant difference between the two interventions in terms of measures of anxiety (SMD = 0.03; 95% CI, -1.16 to 1.22).

Analysis of Publication Bias

Because of the very small number of trials available for each comparison, the statistical tests lacked the power to detect publication bias. Therefore the analysis of the effect of publication bias on the meta-analyses presented above was not conducted.

Substance Abuse

Description of the Included Studies

Seventeen trials (13 RCTs258270 and 4 NRCTs271274) that evaluated the effects of meditation practices in individuals with substance abuse disorders were identified. They included five trials on TM®,259,261,267,270,271 three on Yoga,266,269,273 two on MBSR,263,272 two on RR,265,268 one on CMS,264 one on a medical meditation practice involving the use of mantra and breathing techniques,260 one on Qi Gong,262 one on mindfulness meditation not further specified,258 and one on a meditation practice not further described.274

The trials were published between 1956 and 2004 (median year of publication: 1986; IQR, 1979 to 1999). All the trials were published in journals, except for two,260,263 which were identified from the gray literature. The majority of trials (n = 13) 258260,263265,267272,274 were conducted in the United States; two studies were conducted in India,266,273 one study was conducted in China,262 and one in Sweden.261 Characteristics of the trials are summarized in Table H3 in Appendix H.*

A total of 825 individuals were assigned to meditation practices or control groups. The median sample size based on data from 16 trials was 45 participants per study (IQR, 30 to 77). Two273,274 of the 16 trials had more than 100 participants assigned to the study groups. The mean age of participants based on data from 13 trials was 33 ± 7 years (range: 21 to 45 years). All the trials except two265,270 were conducted in samples with mean ages ranging from 20 to 40 years. Four trials259261,273 did not report the age of participants.

Across all the trials that reported the gender of participants (n = 16), 87 percent were males and 13 percent were females. Samples in nine trials262266,268,270,273,274 were entirely male; none of the trials included entirely female samples. One trial259 failed to report the gender of participants. The race of ethnicity of samples was reported in five trials.258,263,269,270,274 African American participants constituted more than 60 percent of the study population in three trials,258,263,270 whereas Caucasian participants constituted more than 80 percent of the study population in two trials.269,274

All the trials except five217,258,266,268,269 attempted to use formal criteria or validated instruments to select participants in their studies. Two studies used the Addiction Severity Index,258,269 one study266 used the DSM-III criteria for alcohol dependence, and another used the Drinking Practices Questionnaire. The remaining 13 trials selected the study participants based on their reported history of substance abuse.

Participants in the studies were recruited in addiction treatment centers,258,261,262,265,269272,274 prisons,259,263,267 psychiatric wards,266 universities264,268 or from Alcoholics Anonymous.260,273 Abused substances included alcohol,258,260,261,263266,268,270273 cocaine,258,263,272 heroin,258,262,269,272 marijuana,261,272 inhalants,272 hashish,261 amphetamines,261 and lysergic acid diethylamide (LSD).261 Three studies did not provide details about the type of substances abused.

All 17 trials employed a parallel study design. The length of the trials varied from 1 day260 to 18 months.270 The median duration of the trials based on data from 16 trials was 4 months (IQR, 1 to 6). Seven studies259,260,262265,272 were short-term trials (less than 3 months), seven trials258,261,266,268,269,271,274 had a duration between 3 and 6 months, and two trials270,273 lasted longer than 6 months.

The 17 trials comprised four comparisons between meditation practices and no intervention,259,262,264,272 and two comparisons between meditation practices and WL.259,271 There were 20 comparisons between meditation practices and active therapies other than no intervention or WL. As some trials had more than one comparison arm, the total number of comparisons exceeds the number of trials. Of the 20 active comparisons, the comparative treatments were BF,270,273 exercise,264,266 group therapy,261,269 PMR,263,265 rest,260,265 counseling,270 psychotherapy,273 relaxation,274 neurotherapy,270 stereotaxic surgery,273 low frequency pulsed magnetic field therapy,273 and pharmacotherapy.262 Two studies267,268 failed to provide a description of the control group, and one study258 reported the comparison group as “usual care” without providing further details. The median number of comparisons per study was one (IQR, 1 to 2).

Methodological Quality of Included Studies

As a measure of methodological quality for included trials, the overall median Jadad score was 1 (IQR, 1 to 2). Three trials258,265,266 obtained 3 points and were considered high quality (i.e., Jadad scores of 3 points or more). Three trials260,263,264 obtained 2 points, seven trials259,261,267271 obtained 1 point, and four trials262,272274 did not obtain any points.. All the trials except four271274 were described as randomized; however, the description of randomization varied. The majority of trials (8 out of 13259,261,263,264,267270 did not provide a description on how the randomization was performed. Four trials258,260,265,266 described an appropriate method of generating the sequence of randomization, whereas one trial262 reported an inadequate method of sequence generation. None of the trials were described as double-blind. The adequacy of allocation concealment was unclear in all included trials.

None of the studies reported the use of intention-to-treat analysis. Eight trials258,263268,271 reported dropout information for the total study sample (mean dropout rate: 34 percent; range: 0 to 87 percent). Four trials263,264,266,271 had a dropout rate of more than 20 percent. Withdrawals and dropouts per treatment group were clearly described in six trials.258,263266,271 Among the six studies that reported dropouts per treatment group, 24 percent of participants (range: 0 to 48 percent) dropped out from the meditation groups. The mean dropout rate for the control groups was similar (21 percent; range: 0 to 44 percent; eight control groups).

Seven trials264,268270,272274 reported their source of funding. Five trials268270,273,274 received government funding and two264,272 received internal funding. A comparative summary of the methodological quality of the included trials is provided in Table 32.

Table 32. Methodological quality of trials of meditation practices for substance abuse.

Table 32

Methodological quality of trials of meditation practices for substance abuse.

Results of Quantitative Analysis

Table 33 summarizes the type of meditation practice, comparison group, and outcomes that were available for meta-analysis. No single diagnostic criterion was chosen for categorizing study populations. Rather, we included all studies conducted in patients with substance abuse, as defined by the authors of the primary studies. Studies were too dissimilar in type of meditation practice, comparison group, and data for common outcomes of interest to allow direct or indirect comparisons of the effectiveness of meditation practices for substance abuse. No more than one study was available for statistical pooling of any of the 23 comparisons.

Table 33. Summary of outcomes by meditation practice and by comparison group included in meta-analyses of efficacy and effectiveness.

Table 33

Summary of outcomes by meditation practice and by comparison group included in meta-analyses of efficacy and effectiveness.

Results from individual clinical trials of meditation practices in substance abuse are summarized in Table H3 in Appendix H.*

Analysis of Publication Bias

The lack of trials available for a meta-analysis on the effects of meditation practices in substance abuse precluded an assessment of publication bias.

Summary of the Results

Table 34 summarizes the results of the meta-analyses of the treatment effects (statistical and clinical significance) of meditation practices in hypertension and cardiovascular diseases.

Table 34. Summary of the meta-analyses of the treatment effects of meditation practices in hypertension and cardiovascular diseases (statistical and clinical significance).

Table 34

Summary of the meta-analyses of the treatment effects of meditation practices in hypertension and cardiovascular diseases (statistical and clinical significance).

Hypertension

Twenty-seven trials (24 RCTs, and 3 NRCTs) have evaluated the effects of meditation practices in hypertension. The majority of trials on hypertension have been conducted in Yoga (eight studies). The trials have been predominantly conducted in the United States in participants with a mean age of 51 years (range: 41 to 60 years). All studies were conducted in patients with a diagnosis of essential hypertension and used a parallel-group design. The majority of the trials were short- and medium-term. Comparison groups included HE, NT, WL, BF, PMR, and rest. The methodological quality of trials was low with only two trials considered high quality.

Data from 16 studies were available for direct meta-analyses. Outcomes suitable for meta-analysis included blood pressure, body weight, heart rate, total cholesterol, HDL-C, LDL-C, dietary intake, physical activity, and psychological measures such as stress, anger, and self-efficacy.

Direct meta-analyses showed that compared to HE, TM® did not produce significantly greater benefits on blood pressure (SBP and DBP), heart rate, TC, HDL-C, LDL-C, body weight, dietary intake, physical activity, measures of stress, anger, and self-efficacy. A subgroup analysis by study duration showed short-term significant improvement in SBP with TM®, but not over the long-term. When compared to PMR, TM® produced significantly greater benefits in SBP and DBP. When RR was compared to BF, RR did not produce significantly greater benefits on blood pressure (SBP and DBP). Qi Gong was significantly more effective than a WL in reducing SBP. Compared to NT, Yoga produced significant reductions in DBP, but not in SBP. As the results among trials were heterogeneous, a subgroup analysis showed that the effect of Yoga on SBP was significantly greater when compared to a control group without an adjuvant treatment. The same subgroup analysis was conducted for the outcome of DBP and the magnitude of the effect changed from significant to nonsignificant when Yoga was compared to a control group with an adjuvant treatment. When compared to HE, Yoga did not produce significantly greater benefits on SBP and DBP. Heterogeneity in this outcome suggested that short-term trials showed statistically significant benefits in blood pressure, whereas the effects decreased over time. Compared to HE, Yoga produced significant benefits in controlling stress. When compared with blood pressure checks, Zen Buddhist meditation did not produce significantly greater reduction in SBP, but did produce a significant reduction in DBP.

When Tai Chi, Yoga plus BF, and Yoga were indirectly compared with NT, they significantly reduced SBP. These three interventions were also better than HE to reduce SBP. For the outcome of DBP, Yoga plus BF and Yoga alone were the only interventions that significantly reduced DBP when compared to NT. Yoga was also better than HE. Yoga was nonsignificantly superior to TM® for the outcomes of body weight, heart rate and stress. Compared to TM®, RR significantly helped to reduce smoking.

Cardiovascular Diseases

Twenty-one trials (15 RCTs and 6 NRCTs) have evaluated the effects of meditation practices in cardiovascular diseases. The majority of trials have been conducted in Yoga (seven studies). The trials have been predominantly conducted in the United States in participants with a mean age of 63 years (range: 52 to 77 years). Clinical conditions of study populations included MI, coronary artery disease, angina, arrhythmias, peripheral occlusive disease, and congestive heart failure. All studies used a parallel-group design. The majority of the trials were medium-term. Comparison groups included exercise, no intervention group, pharmacological interventions, HE, usual care not specified, group therapy, WL, listening to music, cognitive restructuring training. The methodological quality of trials was low with only two trials considered high quality.

Data from six studies were available for direct meta-analyses. Outcomes suitable for meta-analysis included TC, LDL-C, and body weight; however, only the results from the two trials comparing the use of Yoga with exercise for the reduction of body weight could be combined. This direct meta-analysis showed that Yoga was no better than exercise at producing changes in body weight. Indirect comparisons showed that there were no significant differences in measures of anxiety between Yoga and MBSR.

Substance Abuse

Seventeen trials (13 RCTs and 4 NRCTs) have evaluated the effects of meditation practices in substance abuse. The majority of trials have been conducted on TM® (five studies). The trials have been predominantly conducted in the United States in participants with a mean age of 33 years (range: 21 to 45 years). All studies used a parallel-group design. The majority of the trials were short- and medium- term. Control groups included BF, exercise, group therapy, PMR, rest, counseling, psychotherapy, relaxation, neurotherapy, stereotaxic surgery, low frequency pulsed magnetic field therapy, and pharmacotherapy. The methodological quality of trials was low with only three trials considered high quality. Study results were not combined because the trials were too dissimilar in meditation practice, comparison group, and data for common outcomes of interest. In addition, the results of the three highest quality trials258,265,266 (Jadad score = 3/5) examining, respectively, Mindfulness meditation, RR, and Yoga are inconclusive with respect to the effectiveness of meditation pratices.

The study comparing Mindfulness meditation with usual care (NS) 258 for alcohol and cocaine abuse found little indication that Mindfulness meditation enhanced treatment outcomes for substance abuse patients. The study comparing RR with PMR and rest groups265 for alcohol abuse found generalized effects for BP, but not for the other outcome measures (anxiety, HR, and GSR). The RR and PMR groups did not exhibit increased BP as observed in control subjects. RR and PMR produced significant changes in tension. The study comparing Yoga with exercise266 for alcohol abuse found a significantly greater recovery rate for the Yoga group.

Table 32 provides a summary of the meta-analyses of the treatment effects of meditation practices in hypertension and cardiovascular diseases in terms of the statistical and clinical significance of the findings. Overall, we found that TM® had no advantages over HE to improve measures of SBP, DBP, body weight, heart rate, stress, anger, self-efficacy, cholesterol, dietary intake, and level of physical activity in hypertensive patients. Compared to PMR, TM® produced clinically and statistically significant benefits to reduce SBP. The results for DBP were of borderline clinical significance. Caution should be exerted when interpreting these results. Meta-analyses were derived from only two open label trials; therefore, performance bias and detection bias may have contributed to an overestimate of the treatment effect. RR was not shown to be superior to BF at reducing blood pressure in hypertension.

Qi Gong was superior to WL to reduce blood pressure in subjects with essential hypertension; however, the clinical significance of this finding is questionable, as the effect estimate is quite imprecise (i.e., wide confidence interval), the comparison is based on a few low-quality studies, and the appropriateness of a WL comparison group is questionable. Yoga did not produce clinically or statistically significant effects in blood pressure when compared to NT. Compared to HE, Yoga produced statistically significant changes in measures of stress. The clinical value of this change, however, is questionable (approximately a one-point reduction in measures of stress). Results were obtained from only two open label trials and this could have affected the subjective determination of outcomes. Finally, Zen Buddhist meditation was not better than blood pressure checks to reduce SBP. Although the result for DBP was clinically and statistically significant, caution should be exerted as there was some heterogeneity among the studies that contributed data for this outcome.

Yoga was no better than physical exercises to reduce body weight in patients with cardiovascular disorders. When the relative effectiveness of a variety of meditation practices was assessed using indirect meta-analysis, we found that there were no significant differences between MBSR and Yoga to control anxiety symptoms in cardiovascular patients.

Topic IV. Evidence on the Role of Effect Modifiers for the Practice of Meditation

We aimed to identify the role of effect modifiers (e.g., patient and meditation characteristics) as moderators of the treatment effect measured in clinical trials of meditation practices in hypertension, other cardiovascular diseases, and substance abuse. The small number of trials per comparison and the limited data from primary studies precluded meta-regression analyses using RCT-level covariates to assess the role of specific effect modifiers for the practice of meditation. We were also unable to conduct subgroup analyses to explore differences among subgroups of patients as the trials failed to report results by the effect modifiers being considered (i.e., characteristics of the practice or patients). Therefore, we will describe the findings from the individual studies that reported data on the role of effect modifiers.

Hypertension

Of 27 trials that examined the effect of meditation practices for hypertension, only seven trials203,205,206,209,220222 conducted a subgroup analysis or a multiple regression analysis to explore the role of a variety of effect modifiers. A summary of the analysis is provided in Table 35.

Table 35. Summary of the analyses of effect modifiers for achieving benefits from meditation practice for hypertension.

Table 35

Summary of the analyses of effect modifiers for achieving benefits from meditation practice for hypertension.

Four studies205,206,220,221 conducted an analysis of the role of effect modifiers on health outcomes resulting from the practice of TM®. They used multiple regression models205,206,221 or subgroup analyses220,222 by a variety of effect modifiers such as age,205,206,206 gender, antihypertensive medication use,206,220,222 income,205 education,205 and smoking206 One study209 conducted a subgroup analysis by age, gender, severity of hypertension, duration of disease, and medication use for the effects of RR. Another study203 conducted a subgroup analysis by severity of hypertension and duration of the disease on the effects of mantra meditation and relaxation techniques. Finally, one study on a technique modeled after TM® conducted subgroup analyses of medication use,222 and marital status.222

All the trials were likely to have conducted post hoc analyses as the analyses were not reported as part of the plan of analysis in the Methods sections of the studies. It is unknown whether authors of the trials decided to selectively report on the variables that showed a statistically significant positive effect.

None of the trials that provided data on effect modifiers of meditation practices for hypertension analyzed the effect of the dose of practice necessary to achieve health outcomes. Neither the role of the direction of attention during meditation nor the rhythmic aspects of the practice were explored in the studies. The trials did not provide data on how ethnicity predicts health outcomes resulting from the practice of meditation. The role of individual variables to predict success in the process of meditation (expressed as adherence or acceptance) was not explored in the trials of meditation practice and hypertension.

Cardiovascular Diseases

Of 21 trials on the effects of meditation practice on cardiovascular (CV) diseases, only two trials234,239 conducted subgroup or multiple regression analyses to explore the role of effect modifiers on achieving potential benefits of meditation practice. A summary of the analysis is provided in Table 36.

Table 36. Summary of the analyses of effect modifiers for achieving benefits from meditation practice for cardiovascular diseases.

Table 36

Summary of the analyses of effect modifiers for achieving benefits from meditation practice for cardiovascular diseases.

Using a multiple regression model, one trial234 explored whether age, education, medication use, and diet restrictions were predictors of the effectiveness of RR in patients with CV diseases. Another trial239 conducted a subgroup analysis by type of condition, (i.e., patients with angina versus patients with risk factors) of the effect of an intervention that combined Yoga and dietary changes. Both trials likely conducted post hoc analyses as they were not reported as part of a plan of analysis in the Methods sections of the studies. It is unknown whether there is an outcome selection bias in the reporting of variables that were included in the analysis. None of the trials explored the effect of the dose practice necessary to achieve health outcomes, the role of direction of attention during meditation, or the rhythmic aspects of the practice. The trials did not provide data on whether ethnicity or other individual variables affect associated health outcomes or whether these variables can be used to predict the successful practice of meditation.

Substance Abuse

Of 17 trials investigating the effect of meditation practices on substance abuse disorders, only 4 trials264,267,268,274 conducted subgroup or multiple regression analyses to explore the role of a variety of effect modifiers on achieving potential benefits of meditation practices. A summary of the analysis of effect modifiers is provided in Table 37.

Table 37. Summary of the analysis of effect modifiers for achieving benefits from meditation practice for substance abuse.

Table 37

Summary of the analysis of effect modifiers for achieving benefits from meditation practice for substance abuse.

The trials did not report on effect of variables such as age, gender, or ethnicity. One trial on the effect of RR that incorporated PMR and cognitive restructuring268 conducted a subgroup analysis by level of drinking and level of social support received. The effect of other patient characteristics on the outcomes achieved after practicing meditation were not reported in the studies. One study267 conducted a subgroup analysis by regularity of practice of TM®. A third trial, on a meditation practice not further specified,274 conducted a subgroup analysis by participation in Alcoholic Anonymous groups. All the trials seemed to use exploratory post hoc analyses that were intended to be hypothesis generating. It is unknown whether authors of the trials selectively reported the variables that showed a statistically significant positive effect. The fourth trial264 conducted a subgroup analysis on differences in outcomes between high compliers and noncompliers.

None of the trials that provided data on effect modifiers of meditation practices for substance abuse analyzed the effect of the dose of practice necessary to achieve health outcomes, the role of direction of attention during meditation, or the rhythmic aspects of the practice.

Summary of the Results

The role of effect modifiers such as characteristics of the practice or patient characteristics has so far been neglected in primary research on the effects of meditation practices. Therefore, we were unable to use a linear meta-regression procedure to explore any interactions between patient characteristics or characteristics of the practice and the magnitude of the overall effect of meditation practices for hypertension, cardiovascular diseases, and substance abuse. Individual studies (seven trials on hypertension, two on cardiovascular diseases, and four on substance abuse) conducted subgroup or multiple regression analysis; however, no analyses were reported a priori in the “Methods” sections of the studies. No conclusions on the role of effect modifiers can be drawn from the analysis of the individual studies. Individual patient data are required to appropriately examine this issue.

Topic V. Evidence on the Physiological and Neuropsychological Effects of Meditation Practices

Three hundred and eleven intervention studies provided evidence on 1,323 measures of the physiological and neuropsychological effects of meditation practices. Physiological outcomes only were reported in 253 studies, cognitive and neuropsychological outcomes only in 34 studies, and both physiological and neuropsychological outcomes were reported in 24 studies. The main characteristics and methodological quality of the studies included in topic V are summarized in Tables I1 to I3 in Appendix I.*

General Characteristics

Of the 311 studies providing data for this topic, 54 percent (n = 167) were RCTs, 21 percent (n = 65) were NRCTs, and 25 percent (n = 79) used a before-and-after design. The studies that examined the physiological and neuropsychological effects of meditation practices were composed of 110 trials on Yoga, 47 on TM®, 38 on Tai Chi, 34 on RR, 17 on mantra meditation not further described, 15 on Qi Gong, 12 on MBSR, 10 on Zen Buddhist meditation, 9 on meditation practices not further described, 8 on MM, 4 on CSM, 3 on Acem meditation, and 2 each on MBCT and Vipassana meditation. The studies were published between 1956 and 2005 (median year of publication, 1995; IQR, 1986 to 2002). Most of the studies (88 percent, n = 274) were published as journal articles. Seven percent (n = 22) were theses or dissertations, four percent (n = 13) were abstracts, and one percent (n = 2) were published as research letters. Fifty percent of the studies were conducted in North America (n = 155), followed by Asia (34 percent, n = 106), Europe (11 percent, n = 35), Australasia (three percent, n = 10), and other regions (two percent, n = 5).

Overall Methodological Quality

Randomized controlled trials. The methodological quality of the RCTs was analyzed by the individual components of the Jadad scale. Overall, the methodological quality of the 167 RCTs was poor (median Jadad score 2/5; IQR, 1 to 2). Thirteen percent (n = 21) of the RCTs were considered high quality (i.e., Jadad score of 3 or more). Only one study168 obtained a score of 4 and no study obtained a perfect score of 5. The remaining 146 RCTs had a high risk of bias.

We found that only 32 (19 percent) of the studies described the randomization procedure. Of these 32 studies, 24 described an adequate procedure to randomize study participants to treatment groups, and 8 described inadequate or unreliable methods of randomization that might have introduced imbalances between group characteristics and jeopardized the estimates of the overall treatment effect.

The majority of RCTs (97 percent, n = 162) did not use double blinding to conceal the identity of the interventions. Four studies (two percent) were reported as double-blind trials. Finally, 52 percent (n = 86) of the RCTs provided a description of withdrawals and dropouts from the study.

Adequate concealment of treatment allocation was reported in five percent (n = 8) of the RCTs and was reported but considered inadequate in one percent (n = 2). The remaining RCTs (94 percent, n = 157) failed to describe how they concealed the allocation of subjects to the interventions under study. Finally, the source of funding was disclosed in 46 percent (n = 76) of the RCTs. A summary of the methodological quality of RCTs is presented in Table 38.

Table 38. Methodological quality of RCTs on the physiological and neuropsychological effects of meditation practices.

Table 38

Methodological quality of RCTs on the physiological and neuropsychological effects of meditation practices.

Nonrandomized controlled trials. Overall, the quality of the 65 NRCTs was low (median modified-Jadad score 0/3; IQR, 0 to 1). Thirty-seven percent of the studies (n = 24) received one point out of three for the individual components of the Jadad scale, in all cases for a description of withdrawals or dropouts. The remaining 63 percent did not receive any points. No studies described themselves as double blind. Finally, the source of funding was reported in 32 percent (n = 21) of the NRCTs. A summary of the methodological quality of NRCTs is presented in Table 39.

Table 39. Methodological quality of NRCTs on the physiological and neuropsychological effects of meditation practices.

Table 39

Methodological quality of NRCTs on the physiological and neuropsychological effects of meditation practices.

Before-and-after studies. The quality of the 79 before-and-after studies was low. Only four percent (n = 3) of studies contained a sample population that could be considered representative of the target population. The blinding of outcome assessors to the intervention and assessment was described in 3 percent (n = 2) of studies, the number of study withdrawals in 15 percent (n = 12) and reasons for study withdrawals in 8 percent (n = 6). However, 94 percent of studies (n = 74) reported using the same method of outcome assessment for the pre- and post-intervention periods. Funding source was disclosed in 32 percent (n = 25) of studies. A summary of the methodological quality of before-and-after studies is provided in Table 40.

Table 40. Methodological quality of before-and-after studies on the physiological and neuropsychological effects of meditation practices.

Table 40

Methodological quality of before-and-after studies on the physiological and neuropsychological effects of meditation practices.

Outcome Measures

The ten most commonly reported physiological outcome measures were (1) cardiovascular functioning such as heart rate or blood pressure (169 studies), (2) pulmonary functionary FEV1 and FVC (67 studies), (3) periferal nerve tests such as skin conductance (40 studies), (4) adrenocortical functioning such as cortisol and adrenaline levels (26 studies), (5) lipoprotein levels (25 studies), (6) EMG (23 studies), (7) carbohydrate metabolism such as glucose and insulin levels (18 studies), (8) brain electrophysiology such as EEG patterns (17 studies), (9) metabolic product levels such as lactic acid level (16 studies), and (10) CNS hormone and blood composition (11 studies each).

The ten most commonly reported cognitive/neuropsychological outcomes were measures of (1) attention (19 studies), (2) memory (12 studies), (3) perception (12 studies), (4) other cognitive measures such as overall cognitive functioning (11 studies), (5) reasoning (10 studies), (6) sensorimotor functioning (10 studies), (7) language (7 studies), (8) creativity (4 studies), (9) intelligence (4 studies), and (10) spatial ability (4 studies). Table I6 in Appendix I contains the complete list of reported outcome measures and their associated studies.*

Results of Quantitative Analysis

To summarize the results of the physiological and neuropsychological effects of meditation practices, we combined study results when two or more studies agreed on the type of meditation practice, comparison group, outcomes assessed, and had usable outcome data. Table 41 summarizes the type of meditation practice, comparison group, and outcomes that were available for direct meta-analyses. Meta-analyses of physiological and neuropsychological outcomes on populations with hypertension, cardiovascular diseases or substance abuse have been reported in topic III.

Table 41. Summary of outcomes by meditation practice by comparison group by population included in meta-analyses of physiological and neuropsychological effects of meditation practices.

Table 41

Summary of outcomes by meditation practice by comparison group by population included in meta-analyses of physiological and neuropsychological effects of meditation practices.

Outcomes on the physiological and neuropsychological effects of meditation practices for which data could be combined into a direct meta-analysis were provided by 53 unique studies for a total of 15 comparisons examining five meditation techniques: TM®, RR, Yoga, Tai Chi, and Qi Gong. The remaining 258 studies were not suitable for direct meta-analysis because no more than one study was available for pooling. Yoga interventions provided the most studies for comparison (28 studies), followed by TM® (10 studies), Tai Chi (7 studies), RR (6 studies), and Qi Gong (2 studies). The trials were published between 1974 and 2005 (median year of publication, 1993; IQR, 1989 to 2001).

Of the 53 intervention studies included for meta-analysis, 20 used an RCT design, 8 used an NRCT design, and 25 were before-and-after studies. The majority of studies (n = 43) examined outcomes in healthy populations (athletes, college and university students, workers, military, prisoners, and elderly). The remaining studies examined individuals with hypertension (6 studies) and type II DM (4 studies). The main characteristics and conclusions of the individual studies included in the meta-analyses are provided in Appendix J.*

Methodological Quality of Included Studies

Intervention studies. The median Jadad score for the 20 RCTs was 2/5 (IQR, 1 to 2) (Table 40). No trials were described as double blind and no studies were considered to have employed adequate concealment of treatment allocation. A description of withdrawals and dropouts was provided in 10 of the trials.278287 Only five RCTs reported the source of funding.281,282,285,286,288

The median Jadad score for the eight NRCTs was 1/3 (IQR, 0 to 1). No NRCTs described themselves as double blind (blinding of participant and outcome assessor). Five trials provided a description of withdrawals and dropouts.289293 Three NRCTs reported the source of funding180,181,293 (Table 42).

Table 42. Methodological quality of RCTs and NRCTs included in meta-analyses for physiological and neuropsychological effects of meditation practices.

Table 42

Methodological quality of RCTs and NRCTs included in meta-analyses for physiological and neuropsychological effects of meditation practices.

Before-and-after studies. The overall methodological quality of the 25 included before-and-after studies was low (Table 43). Only one study294 was considered to have a study population representative of the population of interest. Twenty-five studies employed the same method of outcome assessment for pre and post periods; no studies reported blinding of outcome assessors. Two studies295,296 provided a description of withdrawals or dropouts; no studies provided reasons for withdrawals. Nine studies reported their source of funding.294,296303

Table 43. Methodological quality of before-and-after studies included in meta-analyses for physiological and neuropsychological effects of meditation practices.

Table 43

Methodological quality of before-and-after studies included in meta-analyses for physiological and neuropsychological effects of meditation practices.

Transcendental Meditation®

Ten studies assessing the physiological and neuropsychological effects of TM® were identified for meta-analysis: three RCTs,279,282,324 four NRCTs,289,291293 and three before-and-after studies.295,311,319 Among the controlled studies, five studies compared TM® versus NT,279,282,291293 and two compared TM® versus WL.289,309

TM® versus NT

Blood pressure. Three studies282,291,292 totaling 132 participants (TM® = 67, NT = 65) provided data for a meta-analysis on the effects of TM® on blood pressure in healthy populations (Figure 28). The combined estimate of changes in SBP (mm Hg) showed a small, but nonsignificant improvement (reduction) in favor of NT (WMD = 0.93; 95% CI, -9.53 to 11.39). There was evidence of high heterogeneity among the studies regarding the mean change in SBP (p = 0.04, I2 = 69.7 percent). The studies differed in duration with two studies being long-term282,291 (10 and 12 months, respectively) and the remaining study292 being short-term (3 months). A subgroup analysis indicated that for the long-term studies there was a nonsignificant improvement (reduction) in SBP favoring TM® (WMD = -5.24, 95% CI, -12.85, 2.37); for the short-term study, there was a statistically significant improvement favoring NT (Figure 28).

Figure 28. Meta-analysis of the effect of TM® versus NT on SBP.

Figure

Figure 28. Meta-analysis of the effect of TM® versus NT on SBP.

The combined estimate of changes in DBP (mm Hg) indicated a small, but nonsignificant improvement (reduction) in favor of TM® (WMD = -1.63, 95% CI, -8.01 to 4.75) (Figure 29). There was evidence of high heterogeneity among the studies regarding the mean change in DBP (p = 0.04, I2 = 68.8 percent). As noted above, the studies differed in duration. A subgroup analysis indicated a statistically and clinically significant reduction in DBP in favor of TM® (WMD = -5.19, 95% CI, -10.24 to -0.13) in the long-term studies (Figure 29).

Figure 29. Meta-analysis of the effect of TM® versus NT on DBP.

Figure

Figure 29. Meta-analysis of the effect of TM® versus NT on DBP.

Cholesterol level. Three studies282,291,292 totaling 132 participants (TM® = 67, Nt = 65) provided data on the effects of TM® on LDL-C levels in healthy populations. The combined estimate of changes in LDL-C level (mg/dL) indicated a small, nonsignificant improvement (reduction) in favor of TM® (WMD = -15.08; 95% CI, -29.03 to -1.14). The results for the trials were homogeneous (p = 0.44, I2 = 0 percent). However, because of the difference in duration of the studies noted above (two were long-term282,291 and one was short-term292), we conducted a subgroup analysis by duration of study (Figure 30). The long-term studies indicated a statistically significant improvement (reduction) in favor of TM® (WMD = -23.94; 95% CI, -43.87 to -4.00).

Figure 30. Meta-analysis of the effect of TM® versus NT on cholesterol level.

Figure

Figure 30. Meta-analysis of the effect of TM® versus NT on cholesterol level.

Verbal fluency. Two studies279,293 totaling 117 participants (TM® = 66, NT = 51) provided data on the effects of TM® on verbal creativity in healthy populations (Figure 31). The combined estimate of changes in measures of verbal fluency showed a large, significant improvement (increase) in favor of TM® (SMD = -0.74; 95% CI, -1.12 to -0.36). The results of the combined studies were homogeneous (p = 0.73, I2 = 0 percent).

Figure 31. Meta-analysis of the effect of TM® versus NT on verbal fluency.

Figure

Figure 31. Meta-analysis of the effect of TM® versus NT on verbal fluency.

TM® (no control)

Blood pressure. Three before-and-after studies295,311,319 totaling 58 participants provided data on the effect of TM® on blood pressure (mm Hg) in hypertensive populations (Figure 32). The combined estimate of changes in SBP indicated a statistically and clinically significant improvement (reduction) favoring TM® (change from baseline = -10.95; 95% CI, -17.52 to -4.39). There was substantial heterogeneity in the study results (p = 0.16; I2 = 64.1 percent).

Figure 32. Meta-analysis of the effect of TM® (no control) on blood pressure.

Figure

Figure 32. Meta-analysis of the effect of TM® (no control) on blood pressure.

The combined estimate of changes in DBP also indicated a statistically and clinically significant improvement (reduction) favoring TM® (change from baseline = -6.86; 95% CI, -10.54 to -3.19). There was moderate heterogeneity in the study results for DBP (p = 0.16; I2 = 46.3 percent).

All three studies were of low methodological quality; moreover, the potential biases inherent in the before-and-after design may be responsible for the variability of results. Similar interventions, durations (not reported by Benson311), and study populations were used in the three studies. Though all three studies examined hypertensive patients, the baseline measures suggest that the DBP of participants in the Benson311 study (mean DBP 94 ± 9 mm Hg) was lower upon entrance to the trial than the other two studies (minimum 90 mm Hg).

TM® versus WL

Heart rate. Two studies289,309 totaling 70 participants (TM® = 41, WL = 29) provided data on the effects of TM® on heart rate (bpm) in healthy populations (Figure 33). The combined estimate of changes in heart rate showed small, significant improvement (reduction) favoring TM® (WMD = -5.94; 95% CI, -11.54 to -0.35). The trial results were homogeneous (p = 0.73, I2 = 0 percent).

Figure 33. Meta-analysis of the effect of TM® versus WL on heart rate.

Figure

Figure 33. Meta-analysis of the effect of TM® versus WL on heart rate.

Blood pressure. The same two studies289,309 provided data on the effects of TM® on blood pressure in healthy populations (Figure 34). The combined estimate of changes in SBP (mm Hg) showed a small, significant improvement (reduction) favoring TM® (WMD = -8.74; 95% CI, -17.47 to -0.02). There was moderate heterogeneity in the study results (p = 0.15; I2 = 52.6 percent). It is unclear what clinical differences among the study participants are responsible for the heterogeneity of this outcome.

Figure 34. Meta-analysis of the effect of TM® versus WL on blood pressure.

Figure

Figure 34. Meta-analysis of the effect of TM® versus WL on blood pressure.

The combined estimate of changes in DBP (mm Hg) also showed a small, significant improvement (reduction) favoring TM® (WMD = -3.61; 95% CI, -6.62 to -0.59). There was little heterogeneity in the study results (p = 0.31; I2 = 4.6 percent).

Relaxation Response

Six studies191,283,284,288,304,306b assessing the effect of RR on physiological and neuropsychological outcomes were identified for meta-analysis: five RCTs191,283,284,288,304 and one NRCT.306b Two studies compared RR versus BF191,306b and four compared RR versus rest.283,284,288,304

RR versus BF

Muscle tension. Two studies306b,191 totaling 48 participants (RR = 24, BF = 24) provided data on the effect of RR on muscle tension (Figure 35). The combined results of changes in muscle tension (microvolts) indicated a small, nonsignificant change favoring RR (WMD = -1.28; 95% CI, -3.23 to 0.68). There was little heterogeneity in the study results (p = 0.29; I2 = 11.7 percent).

Figure 35. Meta-analysis of the effect of RR versus BF on muscle tension.

Figure

Figure 35. Meta-analysis of the effect of RR versus BF on muscle tension.

RR versus rest

Heart rate. Three trials283,288,304 totaling 99 participants (RR = 45, rest = 44) provided data on the effect of RR on heart rate in healthy populations (Figure 36). The combined estimate of changes in heart rate (bpm) showed a significant improvement (reduction) favoring rest (WMD = 2.56; 95% CI, 1.32 to 3.80). The study results were homogeneous (p = 0.70, I2 = 0 percent).

Figure 36. Meta-analysis of the effect of RR versus rest on heart rate.

Figure

Figure 36. Meta-analysis of the effect of RR versus rest on heart rate.

Blood pressure. Two studies283,284 totaling 109 participants (RR = 45, rest = 44) provided data on the effect of RR on blood pressure in healthy populations (Figure 37). The combined estimate of changes in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) favoring RR (WMD = -5.67; 95% CI, -12.76 to 1.42). There was evidence of moderate heterogeneity between the study results (p = 0.23; I2 = 31.6 percent). The combined estimates for DBP showed a small, nonsignificant improvement favoring RR (WMD = -6.98; 95% CI, -16.05 to 2.09). There was evidence of considerable heterogeneity between the study results (p = 0.11; I2 = 60.5 percent). Both studies were short-term and similar in participant characteristics (proportion of males to females, healthy) and intervention. The most likely source of heterogeneity is study design (RCT284 versus NRCT283).

Figure 37. Meta-analysis of the effect of RR versus rest on blood pressure.

Figure

Figure 37. Meta-analysis of the effect of RR versus rest on blood pressure.

Yoga

Twenty-eight studies assessing the physiological and neuropsychological effect of Yoga were identified for meta-analysis: 8 RCTs,140,177,204,278,280,281,305,308 2 NRCTs,180,181 and 18 before-and-after studies.83,294,297,298,300303,310,312315,318,320323 Four trials168,204,305,325 compared Yoga versus NT, two trials280,281 compared Yoga versus exercise, two177,181 compared Yoga versus free breathing (FB), two278,308 compared Yoga versus medication, and two140,180 compared Yoga (unilateral left nostril breathing [ULNB]) versus another Yoga intervention (unilateral right nostril breathing [URNB]).

Yoga (no control)

Heart rate. Seven before-and-after studies83,302,310,313,320,322,323 (17, 10, 12, 50, 25, 40, and 18 participants, respectively) provided data on the effect of Yoga (no control) on heart rate (bpm) in healthy populations (Figure 38). The substantial heterogeneity among the study results (p < 0.00001; I2 = 95.9 percent) precluded reporting a combined estimate. Four of the seven studies indicated significant improvements (reduction) after practicing Yoga,302,310,313,323 whereas three did not favor the intervention.83,320,322

Figure 38. Meta-analysis of the effect of Yoga (no control) on heart rate.

Figure

Figure 38. Meta-analysis of the effect of Yoga (no control) on heart rate.

Possible sources for the observed heterogeneity were explored. Two studies were very short-term and reported study duration in number of sessions (one to six).83,322 The remaining five studies had a 3-month study period.302,310,313,320,323 The studies also differed in the frequency and length of intervention sessions: two studies reported sessions of less than 1 hour,83,322 two studies reported 1-hour sessions,313,320 two studies reported sessions of 4 hours,302,323 and one study310 did not report session length. Five studies were considered to have used composite interventions302,310,313,320,323 composed of some combination of postures, breathing techniques, cleansing practices, meditation, and lectures. Two studies were considered to have used single interventions; however, the two studies employed different techniques (breathing exercises322 and Raja meditation83).

The age range of participants also varied, with the mean ages ranging from 25313 to 35 years.322 Three studies included only men,83,313,323 one study included only women,310 two studies included both men and women in almost equal proportion,302,320 and one study failed to report the gender of participants.322

Two studies294,303 totaling 33 participants provided data on the effect of Yoga on heart rate in hypertensive populations (Figure 39). The combined estimate of changes in heart rate (bpm) showed a small, significant improvement (reduction) favoring Yoga (change from baseline = -6.79; 95% CI, -9.97 to -3.60). There was evidence of moderate heterogeneity among the study results (p = 0.21, I2 = 35.6 percent). The possible sources of heterogeneity were explored. While the subjects in both studies were similar in mean age, the Vijayalakshmi study303 did not include women, whereas Damodaran294 had a male to female ratio of 1:4. The two studies also differed in quality with Damodaran294 having a study population considered representative of the population of interest, while Vijayalakshmi303 had a nonrepresentative study population.

Figure 39. Meta-analysis of the effect of Yoga (no control) on heart rate in hypertensive populations.

Figure

Figure 39. Meta-analysis of the effect of Yoga (no control) on heart rate in hypertensive populations.

Blood pressure. Five studies302,310,313,322,323 totaling 201 participants provided data on the effect of Yoga on blood pressure in healthy populations (Figure 40). The combined estimate of changes in SBP (mm Hg) indicated a small, significant improvement (reduction) favoring Yoga (change from baseline = -8.05; 95% CI, -14.01 to -2.09). There was evidence of substantial heterogeneity among the study results (p = 0.00001; I2 = 89.1 percent).

Figure 40. Meta-analysis of the effect of Yoga (no control) on blood pressure.

Figure

Figure 40. Meta-analysis of the effect of Yoga (no control) on blood pressure.

The combined estimate of changes in DBP (mm Hg) also indicated a small, significant improvement (reduction) favoring Yoga (change from baseline = -6.22; 95% CI, -7.73 to -4.70). The study results were homogeneous (p = 0.52; I2 = 0 percent).

The discrepancy between the measures of heterogeneity found for SBP and DBP is possibly accounted for by the difference in baseline measures. The participants in the Schmidt302 study had a combined SBP baseline noticeably higher (9 mm Hg from the next highest) than the those of the other studies, and it is unclear what clinical differences may be responsible for this difference in baseline measures. Other than baseline measures, the studies were comparable in study design, duration, and other participant characteristics. The baseline measures of DBP were similar across all groups.

Respiratory rate. Three studies83,310,323 with 17, 40, and 18 participants respectively provided data on the effect of Yoga on respiratory rate in healthy populations (Figure 41). The heterogeneity among study results was substantial (p = 0.0001; I2 = 85.5 percent) and precluded combining the studies. The Telles study83 differed from the other two studies in duration and the type of yogic practice used. Anantharaman323 and Telles323 were short-term (3-month) studies using postures and breathing exercises. In contrast, Telles83 employed Raja yoga meditation (seated meditation with a fixed gaze) that lasted three sessions (approximately 1 week).

Figure 41. Meta-analysis of the effect of Yoga (no control) on respiratory rate.

Figure

Figure 41. Meta-analysis of the effect of Yoga (no control) on respiratory rate.

Galvanic skin resistance. Two studies83,323 totaling 58 participants provided data on the effect of Yoga on galvanic skin resistance in healthy populations (Figure 42). The combined estimate of changes in skin resistance (kilohms) indicated a nonsignificant difference favoring the “before Yoga” period (change from baseline = 3.12, 95% CI, -12.15 to 18.40). There was evidence of moderate heterogeneity between the study results (p = 0.24, I2 = 27.2 percent).

Figure 42. Meta-analysis of the effect of Yoga (no control) on galvanic skin resistance.

Figure

Figure 42. Meta-analysis of the effect of Yoga (no control) on galvanic skin resistance.

The possible sources of heterogeneity in the outcome of galvanic skin resistance were differences in the intervention, comparison period, and duration of study. One study323 used a multicomponent intervention that consisted of yogic postures, breathing exercises, meditation, cleansing exercises, and lectures. The comparison period preceded the learning of any yogic techniques, but was not fully described. The outcome measurements were taken at the end of a 3-month period. The second study83 used a seated meditation technique in which participants fixed their gaze on a light and thought positive thoughts about a universal force. The nonmeditation period involved sitting quietly without targeted thinking; the outcome measures were assessed the day after the baseline measures.

Fasting blood glucose (type II DM). Two studies,298,321 with 149 and 24 participants respectively, provided data on the effect of Yoga on levels of fasting blood glucose in populations with type II DM (Figure 43). The heterogeneity of the combined study results was too high (p = 0.001; I2 = 90.6 percent) to report an overall estimate. While both studies employed Yoga postures and breathing techniques, Jain298 employed two breathing techniques called “kapalbhati” (also described as a cleansing practice and a milder form of bhastrika119) and “ujjayi,” and a variety of postures and cleansing practices. Singh321 employed “bhastrika pranayama”(a breathing exercise) and postures, and did not use cleansing practices. In addition, while the Singh study used sessions of about 30 minutes duration, the daily sessions in the Jain study lasted 2.5 hours (1.5 hours in the morning and 1 hour in the evening).

Figure 43. Meta-analysis of the effect of Yoga (no control) on fasting blood glucose (type II DM).

Figure

Figure 43. Meta-analysis of the effect of Yoga (no control) on fasting blood glucose (type II DM).

Fasting blood glucose (healthy). Two studies301,313 totaling 30 participants provided data on the effect of Yoga on levels of fasting blood glucose in healthy populations (Figure 44). The study of Manjunatha301 provided data on the effect of four different sets of two asana techniques. For each of the sets, we pooled the results with the results from the Joseph313 study. Each of the combined estimates of change in blood glucose level (mg/dL) showed a small, nonsignificant improvement (reduction) favoring Yoga (change from baseline ranged from -3.64 [95% CI, -7.92 to 0.64] to -3.81 [95% CI, -7.97 to 0.35]). There was no evidence of heterogeneity for any of the pooled results (p-values range from 0.37 to 0.47; I2 = 0 percent).

Figure 44. Meta-analysis of the effect of Yoga (no control) on fasting blood glucose.

Figure

Figure 44. Meta-analysis of the effect of Yoga (no control) on fasting blood glucose.

Breath holding time. Four studies300,312,314,323 provided data on the effect of Yoga on breath holding time (seconds) in healthy populations. Three studies300,312,314 totaling 112 participants examined breath holding time after inspiration The combined results of changes indicated a large improvement (increase) after practicing Yoga (change from baseline = -18.85; 95% CI, -22.64 to -15.05). The study results were homogeneous (p = 0.49; I2 = 0 percent).

Four studies300,312,314,323 totaling 152 participants examined breath holding after exhalation. The combined results of changes indicated a large improvement (increase) after practicing Yoga (change from baseline = -14.53; 95% CI, -16.82 to -12.24). The study results were homogeneous (p = 0.87, I2 = 0 percent) (Figure 45).

Figure 45. Meta-analysis of the effect of Yoga (no control) on breath holding time after inspiration and expiration.

Figure

Figure 45. Meta-analysis of the effect of Yoga (no control) on breath holding time after inspiration and expiration.

Auditory reaction time. Two trials,300,318 with 27 and 41 participants respectively, provided data on the effect of Yoga on auditory reaction time (milliseconds) (Figure 46). Though both studies found statistically significant results favoring Yoga, the results of the studies were too heterogeneous to report as a combined estimate (p = 0.0001; I2 = 94.1 percent). Possible sources of heterogeneity include characteristics of study participants and duration of the intervention. Madanmohan300 included men only with an age range from 18 to 21 years. Malathi318 included men only with an age range from 30 to 45 years. Both studies were short-term; however, the Madanmohan300 study had a duration of 12 weeks compared to six weeks for the Malathi study.318 Finally, participants in the Madanmohan300 study practiced Yoga for 30 minutes per day; those in the Malathi318 study practiced 1 hour per day.

Figure 46. Meta-analysis of the effect of Yoga (no control) on auditory reaction time.

Figure

Figure 46. Meta-analysis of the effect of Yoga (no control) on auditory reaction time.

Visual reaction time. Two studies300,318 totaling 110 participants provided data on the effect of Yoga on visual reaction time (Figure 47). The combined estimate of change in visual reaction time (milliseconds) indicated a small, significant improvement (reduction) favoring Yoga (change from baseline = -36.06; 95% CI, -53.65 to -18.57). There was evidence of substantial heterogeneity between the study results (p = 0.03, I2 = 79.4 percent). As noted in the previous section, the observed heterogeneity is possibly accounted for by differences in participant characteristics, study duration, and duration of practice.

Figure 47. Meta-analysis of the effect of Yoga (no control) on visual reaction time.

Figure

Figure 47. Meta-analysis of the effect of Yoga (no control) on visual reaction time.

Intraocular pressure. Two studies297,315 totaling 67 participants provided data on the effect of Yoga (unilateral nostril breathing [UNB]) on intraocular pressure in healthy populations (Figure 48). The two studies assessed ipsi- and contralateral eye and nostril combinations. We did not pool the results of the studies because the outcomes were measured under different conditions, which may have resulted in the observed heterogeneity. Chen297 took before and after measures while the study participants were at rest. Kocer315 took baseline measures while participants were resting; the “after” measures were taken during exercise. Kocer315 reported a statistically significant change favoring UNB for all eye/nostril combinations. In contrast, Chen297 reported a nonsignificant change favoring no UNB for left nostril breathing. For right nostril breathing, the results favored UNB; however, only the right nostril/right eye combination was statistically significant.

Figure 48. Meta-analysis of the effect of Yoga (no control) on intraocular pressure.

Figure

Figure 48. Meta-analysis of the effect of Yoga (no control) on intraocular pressure.

Yoga versus exercise

Heart rate. Two trials280,281 totaling 91 participants provided data on the effect of Yoga on heart rate in healthy populations (Figure 49). The combined estimate of changes in heart rate (bpm) indicated a small, nonsignificant improvement (reduction) favoring Yoga (WMD = -1.39; 95% CI, -8.24, 5.47). There was evidence of high heterogeneity between the study results (p = 0.10, I2 = 62.6%). The heterogeneity is possibly accounted for by the difference in study duration. Blumenthal280 was a short-term study (6 weeks) that found statistically nonsignificant results favoring exercise and Bowman281 was a long-term study (14 months) that found statistically nonsignificant results favoring Yoga.

Figure 49. Meta-analysis of the effect of Yoga versus exercise on heart rate.

Figure

Figure 49. Meta-analysis of the effect of Yoga versus exercise on heart rate.

Oxygen consumption. Two studies280,281 totaling 91 patients provided data on the effect of Yoga on oxygen consumption in healthy populations (Figure 50). The combined estimate of changes in VO2 max (ml/kg/min) indicated a small, nonsignificant improvement (increase) favoring exercise (WMD = 1.91; 95% CI, -0.48 to 4.31). The study results were homogeneous (p = 0.81, I2 = 0%).

Figure 50. Meta-analysis of the effect of Yoga versus exercise on oxygen consumption (VO2 max).

Figure

Figure 50. Meta-analysis of the effect of Yoga versus exercise on oxygen consumption (VO2 max).

Yoga versus free breathing

Verbal ability. Two studies177,181 totaling 104 participants (Yoga = 52, free breathing = 52) provided data on the effect of Yoga, specifically, unilateral left and right nostril breathing (ULNB, URNB), on verbal ability based on a consonant-vowel matching task (Figure 51). For ULNB, the combined estimate of changes in verbal ability (score) indicated a small, nonsignificant improvement (increase) favoring free breathing (WMD = 0.26; 95% CI, -1.96 to 2.47). There was evidence of moderate heterogeneity between the study results (p = 0.18; I2 = 44.8 percent).

Figure 51. Meta-analysis of the effect of Yoga (ULNB) versus free breathing on verbal ability.

Figure

Figure 51. Meta-analysis of the effect of Yoga (ULNB) versus free breathing on verbal ability.

For URNB, the combined estimate of changes in verbal ability (score) indicated a small, nonsignificant improvement (increase) favoring free breathing (WMD = 1.50; 95% CI, -1.37 to 4.37). There was evidence of substantial heterogeneity between the study results (p = 0.10; I2 = 62.2 percent). A possible source of heterogeneity may be study design. Block177 used an RCT design while Sanders181 employed an NRCT design and found statistically nonsignificant results favoring Yoga. Both studies were of low methodological quality. Lack of reporting prevented further exploration of sources of heterogeneity related to participant characteristics, and study duration.

Spatial ability. Two studies177,181 totaling 104 participants (Yoga = 52, free breathing = 52) provided data on the effect of Yoga, specifically (ULNB and URNB) on measures of spatial ability (Figure 52). For ULNB, the combined measures of spatial ability (score) indicated a nonsignificant improvement (increase) favoring free breathing (SMD = 0.05; 95% CI, -0.34 to 0.43). The study results were homogeneous (p = 0.95; I2 = 0 percent).

Figure 52. Meta-analysis of the effect of Yoga (ULNB and URNB) versus free breathing on spatial ability.

Figure

Figure 52. Meta-analysis of the effect of Yoga (ULNB and URNB) versus free breathing on spatial ability.

For URNB, the combined estimate of change in measures of spatial ability indicated a nonsignificant improvement (increase) favoring free breathing (SMD = 0.24; 95% CI, -0.15 to 0.63). The study results were homogeneous (p = 0.53; I2 = 0 percent).

Yoga versus NT

Blood pressure. Two studies204,305 totaling 79 participants (Yoga = 40, NT = 39) provided data on the effect of Yoga (Shavasana) on blood pressure in healthy populations (Figure 53). The combined estimate of change in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) favoring Yoga (WMD = -8.10; 95% CI, -16.94 to 0.74). The study results were homogeneous (p = 0.88; I2 = 0 percent).

Figure 53. Meta-analysis of the effect of Yoga (shavasana) versus NT on blood pressure.

Figure

Figure 53. Meta-analysis of the effect of Yoga (shavasana) versus NT on blood pressure.

The combined estimate of changes in DBP (mm Hg) also indicated a small, nonsignificant improvement (reduction) favoring Yoga (WMD = -6.09; 95% CI, -16.83 to 4.64). The study results were homogeneous (p = 0.98; I2 = 0 percent).

Yoga versus medication

Fasting blood glucose. Two studies,278,308 with 154 (Yoga = 82, medication = 72) and 21 (Yoga = 11, medication = 10) participants respectively, provided data on the effect of Yoga on fasting blood glucose in populations with type II DM (Figure 54). The heterogeneity between the results was too high to report a combined result (p < 0.00001; I2 = 96.2 percent). The results of Agrawal278 indicated a large, significant improvement (reduction) favoring Yoga (mean difference = -32.07; 95% CI, -43.57 to -20.57). The results of Monro308 showed a smaller, but still significant improvement (reduction) favoring Yoga (mean difference = -1.80; 95% CI, -3.26 to -0.34).

Figure 54. Meta-analysis of the effect of Yoga versus medication on fasting blood glucose.

Figure

Figure 54. Meta-analysis of the effect of Yoga versus medication on fasting blood glucose.

Possible sources of heterogeneity are differences in the proportion of males and females included in the studies and the complexity of the interventions. Agrawal278 had a proportion of males to females of approximately 1:1 and Monro308 had a proportion of approximately 2:1. In addition, the Agrawal study278 used a complex intervention consisting of Yoga asanas, diet modification, aerobic exercise, and HE, whereas the Monro study308 used a standard set of postural, breathing, and relaxation exercises.

Yoga (ULNB) versus Yoga (URNB)

Heart rate. Two studies140,180 totaling 74 participants (ULNB = 37, URNB = 37) provided data on the effect of Yoga (ULNB) on heart rate in healthy populations (Figure 55). The combined estimate of changes in heart rate (bpm) showed a small, nonsignificant improvement (reduction) favoring ULNB (WMD = -2.12; 95% CI, -4.41, 0.17). The study results were homogeneous (p = 0.43; I2 = 0 percent)

Figure 55. Meta-analysis of the effect of Yoga (ULNB) versus URNB on heart rate.

Figure

Figure 55. Meta-analysis of the effect of Yoga (ULNB) versus URNB on heart rate.

Tai Chi

Seven studies assessing the physiological and neuropsychological effect of Tai Chi were identified for meta-analysis: four RCTs,285287,307 one NRCT,290 and two before-and-after studies.296,317 Four studies compared Tai Chi to NT,285,286,290,326 and two compared Tai Chi versus exercise.287,307

Tai Chi versus NT

Heart rate. Two studies,285,290 with 28 (Tai Chi = 18, NT = 10) and 20 (Tai Chi = 10, NT = 10) participants respectively, provided data on the effect of Tai Chi on resting heart rate in healthy, elderly populations (Figure 56). The heterogeneity in the study results was too high to report a combined result (p = 0.002; I2 = 89.2 percent). The results of Chen290 indicated a small, nonsignificant change favoring Tai Chi (mean difference = -0.70; 95% CI, -1.50 to 0.09). In contrast, Sun285 showed a larger, significant improvement (reduction) favoring NT (mean difference = 1.26; 95% CI, 0.28 to 2.24).

Figure 56. Meta-analysis of the effect of Tai Chi versus NT on heart rate.

Figure

Figure 56. Meta-analysis of the effect of Tai Chi versus NT on heart rate.

The opposite direction of effect between the two studies is possibly a result of the study design, frequency of practice, or complexity of the intervention. Chen290 used an NRCT design, while Sun285 conducted an RCT. In addition, the participants in the Chen study practiced a Tai Chi program of 24 forms for 1 hour twice weekly. The participants in the Sun study285 practiced 2 hours once a week and incorporated HE and stress management techniques in addition to a Tai Chi program.

Blood pressure. Three studies285,286,290 with 28 (Tai Chi = 18, NT = 10), 20 (Tai Chi = 10, NT = 10), and 34 (Tai Chi = 17, NT = 17) participants respectively provided data on the effect of Tai Chi on blood pressure in healthy populations (Figure 57). The results were too heterogeneous to report as a combined estimate (p = 0.001, I2 = 94.5%). The results of the Chen290 study showed a small, nonsignificant improvement (reduction) favoring Tai Chi (mean difference = -0.69; 95% CI, -11.72 to 10.34). Sun285 indicated a moderate, significant improvement (reduction) favoring NT (mean difference = 5.20; 95% CI, 3.73 to 6.67), and the results of Thornton286 showed a moderate, significant improvement (reduction) favoring Tai Chi (mean difference = -7.70; 95% CI, -11.65 to -3.75).

Figure 57. Meta-analysis of the effect of Tai Chi versus NT on blood pressure.

Figure

Figure 57. Meta-analysis of the effect of Tai Chi versus NT on blood pressure.

The heterogeneity in results for changes in DBP (mm Hg) also precluded reporting a combined estimate (p = 0.00001; I2 = 98%). Chen290 reported a small, nonsignificant improvement (reduction) favoring Tai Chi (mean difference = 0.44; 95% CI, -5.03 to 5.91). Sun285 reported a moderate, significant improvement (reduction) favoring NT (mean difference = 5.20; 95% CI, 4.23 to 6.17), and the results of Thornton286 showed a moderate, significant improvement (reduction) favoring Tai Chi (mean difference = -8.50; 95% CI, -11.00 to -6.00).

The possible sources of heterogeneity are the age of the study participants and the frequency of practice. While Thornton286 used healthy volunteers between the ages of 33 to 55 years (mean age 48 years), Sun285 employed healthy elderly participants over 60 years of age. The Chen290 study used participants between the ages of 50 and 74 years. The Sun285 and Thornton286 studies used a similar frequency and form of Tai Chi intervention (two to three times per week, 108 forms), while the Chen290 study employed only one Tai Chi session per week and did not describe the number of forms or Tai Chi style used.

Tai Chi versus exercise

Blood pressure. Two RCTs287,307 totaling 110 participants (Tai Chi = 55, exercise = 55) provided data on the effect of Tai Chi on blood pressure in healthy populations (Figure 58). The combined estimate of changes in SBP (mm Hg) showed a moderate, nonsignificant improvement (reduction) favoring exercise (WMD = 1.79; 95% CI, -0.82 to 4.41). There was evidence of low heterogeneity between the studies regarding the mean change in SBP (p = 0.28; I2 = 14.7 percent).

Figure 58. Meta-analysis of the effect of Tai Chi versus exercise on blood pressure.

Figure

Figure 58. Meta-analysis of the effect of Tai Chi versus exercise on blood pressure.

The combined estimate of changes in DBP (mm Hg) showed a small, significant improvement (reduction) favoring exercise (WMD = 0.83; 95% CI, 0.18 to 1.48). There was evidence of low heterogeneity between the studies regarding the mean change in DBP (p = 0.32; I2 = 0.8 percent). The heterogeneity is possibly accounted for by the difference in the ages of the participants, and the frequency and duration of the intervention. Jin307 used healthy volunteers with a mean age of 36 years; Young287 used healthy elderly participants with a mean age of 67 years. The Jin study employed two 1-hour Tai Chi sessions, whereas the Young study lasted 12 weeks and had four 30-minute sessions of Tai Chi per week.

Tai Chi (no control)

Heart rate. Two studies296,317 totaling 74 participants provided data on the effect of Tai Chi on heart rate in healthy populations (Figure 59). The combined estimate of changes in heart rate (bpm) indicated a small, nonsignificant improvement (reduction) favoring Tai Chi (change from baseline = -2.34; 95% CI, -5.29 to 0.60). There was evidence of heterogeneity between the studies (p = 0.09, I2 = 65 percent). A possible source of heterogeneity is the difference in the age of study participants. The mean age of participants in Jones296 was 53 ± 10 years; the mean age of participants in Liu317 was 22 ± 3 years. Liu317 did not report the frequency or complexity of the intervention used, so the studies could not be compared for these variables.

Figure 59. Meta-analysis of the effect of Tai Chi (no control) on heart rate.

Figure

Figure 59. Meta-analysis of the effect of Tai Chi (no control) on heart rate.

Blood pressure. Two studies296,317 totaling 74 participants provided data on the effect of Tai Chi on blood pressure in healthy populations (Figure 60). The combined estimate of changes in SBP (mm Hg) indicated a small, nonsignificant improvement (reduction) favoring Tai Chi (change from baseline = -3.35; 95% CI, -7.55 to 0.85). There was evidence of heterogeneity between the studies regarding change from baseline (p = 0.12; I2 = 57.7 percent).

Figure 60. Meta-analysis of the effect of Tai Chi (no control) on blood pressure.

Figure

Figure 60. Meta-analysis of the effect of Tai Chi (no control) on blood pressure.

The combined estimate of changes in DBP (mm Hg) also showed a small, nonsignificant improvement (reduction) favoring Tai Chi (change from baseline = -0.93; 95% CI, -3.31 to 1.44). The study results were homogeneous (p = 0.41, I2 = 0 percent). Though there were differences between the studies regarding the age of participants (mean ages respectively 53 ± 10 years and 22 ± 3 years) and proportion of males to females (1:5 and 1:1), it is unlikely that these are the sources of the discrepancy in the measures of heterogeneity for SBP and DBP results. As a result, it is unclear what clinical differences may be responsible for the discrepancy.

Qi Gong

Two before-and-after studies that assessed the effect of Qi Gong on physiological outcomes were identified.191,299

Qi Gong (no control)

Heart rate. Two studies191,299 totaling 29 participants assessed the effect of Qi Gong on heart rate in healthy populations (Figure 61). The combined estimate in changes in heart rate (bpm) indicated a small, nonsignificant improvement (reduction) after practicing Qi Gong (change from baseline = -1.21; 95% CI, -6.18 to 3.76). The study results were homogeneous (p = 0.79, I2 = 0 percent).

Figure 61. Meta-analysis of the effect of Qi Gong (no control) on heart rate.

Figure

Figure 61. Meta-analysis of the effect of Qi Gong (no control) on heart rate.

Summary of the Results

Overall, 311 studies evaluated the physiological and neuropsychological effects of meditation practices. The majority of studies used an RCT design (54 percent) and Yoga was the most common intervention (35 percent) that was studied. The overall methodological quality of all studies was low with only one study168 considered high quality. The majority of studies were of short and medium duration.

Data from 53 studies (20 RCTs, 8 NRCTs, 25 before-and-after) were considered for direct meta-analysis to provide an evaluation of the effects of TM®, RR, Yoga, Tai Chi, and Qi Gong. The intervention groups were compared variously against BF, exercise, free breathing, medication, NT, WL, and UNB. Outcomes suitable for meta-analysis included blood pressure (SBP and DBP), heart rate, total cholesterol, respiratory rate, fasting blood glucose, and galvanic skin resistance. The majority of studies used healthy participants (45 studies) as the comparison group; people with type II DM and with essential hypertension comprised the only other study populations (six studies and four studies, respectively). All results below apply to healthy populations unless otherwise indicated.

Transcendental Meditation®

Direct meta-analysis showed that compared to NT, TM® did not produce significantly greater benefits on blood pressure (SBP and DBP). However, there was significant improvement in LDL-C levels and verbal creativity with TM®. When compared to WL, TM® produced significantly greater reduction in SBP and DBP. Before-and-after studies on TM® for patients with essential hypertension indicated a statistically significant reduction in SBP and DBP after practicing TM®. The heterogeneity present for the comparisons evaluating blood pressure changes and cortisol levels suggests that there were important clinical differences among the studies; however, the small number of studies precluded subgroup analyses.

Relaxation Response

The results of meta-analysis showed that compared to BF, RR did not produce significantly greater reduction in muscle tension. When RR was compared to a condition of rest, the rest group showed a significantly greater reduction in heart rate.

Yoga

When compared to NT, Yoga did not show a significantly greater benefit in lowering SBP or DBP. When compared to exercise, Yoga did not significantly lower heart rate or increase oxygen consumption. Compared to URNB, ULNB showed no significantly greater benefit in reducing heart rate.

When compared to free breathing, Yoga (UNB) showed no statistically or clinically significant benefit in improving verbal or spatial ability test scores. Finally, when examined using a before-and-after design, practicing Yoga did not demonstrate a significant benefit decreasing heart rate. There was also no significant benefit for Yoga in increasing galvanic skin resistance, reduction of intraocular pressure, and reduction of fasting blood glucose in healthy populations. There was varied heterogeneity among studies combined for heart rate, respiratory rate, galvanic skin resistance, and intraocular pressure, suggesting important clinical differences among the studies.

Before-and-after studies showed a significantly greater benefit after practicing Yoga in reducing heart rate in hypertensive populations. In healthy populations, practicing Yoga demonstrated a significant benefit in reducing DBP. There was also indication that Yoga has significantly greater benefit in increasing breath holding time after inspiration and expiration, in decreasing visual reaction time, and in the reduction of intraocular pressure (two of four outcomes). There was varied heterogeneity between studies combined for heart rate in hypertensive patients, SBP, and fasting blood glucose in patients with type II DM, suggesting important clinical differences between the studies. The heterogeneity present in the results examining respiratory rate, auditory reaction time, intraocular pressure (two of four outcomes), and fasting blood glucose prevented calculating an overall estimate of effect and suggested important clinical differences between the studies.

Tai Chi

The results of studies that compared Tai Chi to NT were too heterogeneous to provide combined estimates for the effect of Tai Chi on heart rate and blood pressure. In addition, the small number of studies precluded a subgroup analysis.

When compared to exercise, Tai Chi showed no significantly greater reduction in SBP, but did indicate a significant benefit in the reduction of DBP. In before-and-after studies, there was no significant reduction in heart rate, SBP and DBP after practicing Tai Chi than before. Substantial heterogeneity was also present in this comparison and, as with NT, a lack of studies prevented a subgroup analysis.

Qi Gong

Qi Gong did not significantly reduce heart rate in elderly populations, nor did it significantly reduce SBP and DBP in healthy populations.

There were 22 outcome measures on the physiological and neuropsychological effects of meditation practices for which a combined estimate could be produced with little or no statistical heterogeneity. The comparisons, overall effect estimate, and statistical and clinical significance of each is outcome is summarized in Table 44.

Table 44. Summary of statistical and clinical significance of physiological outcomes examined in clinical studies on meditation practices.

Table 44

Summary of statistical and clinical significance of physiological outcomes examined in clinical studies on meditation practices.

Statistically and clinically significant changes in healthy participants were produced by TM® for heart rate, DBP, and LDL-C (TM® versus NT), and for SBP and DBP (TM® versus WL). The increase in verbal creativity (SMD = 0.74; TM® versus WL) is also statistically significant, but it is unlikely that this change would be clinically meaningful. In contrast, the change in SBP (TM® versus NT) was not statistically significant; however, the effect estimate suggests a clinically meaningful reduction of 5.24 mm Hg. When compared to rest, RR was more effective at reducing heart rate. However, though statistically significant, the change suggested by the overall effect estimate is unlikely to be clinically meaningful.

Statistically and clinically meaningful changes were produced in healthy participants by the practice of Yoga for breath holding (18 and 14 breaths/minute) and DBP (-6.22 mm Hg). There was a significant reduction in heart rate in hypertensive patients (-7 bpm); however, the clinical significance of this change depends on the baseline measures of the population for which the intervention is being considered. All the changes described above were observed in studies using a before-and-after design, a design that is unable to control for a host of extraneous variables that may bias the study results (e.g., temporal trends, regression to the mean, and sensitivity to design parameters) and potentially overestimate the effect of the intervention. Therefore, any causal claim about the effect of the intervention should be considered in light of these methodological shortcomings and caution should be exercised when interpreting these results.

The overall effects of Yoga based on RCTs and NRCTs indicate that this practice does not produce significant changes in healthy populations in oxygen consumption, spatial ability, SBP, DBP, or fasting glucose.

When Tai Chi was compared to exercise, there was a statistically significant reduction in DBP; however, the change was not clinically significant (0.83 mm Hg). No significant change was observed in SBP. Before-and-after studies on Tai Chi did not indicate a clinically significant change in either SDP or DBP; however, these results should be interpreted in light of the stronger evidence available from the two combined RCTs. Likewise, no statistically or clinically significant changes in heart rate were produced by the practice of Qi Gong. Nevertheless, this result is based on before-and-after studies and the result should be considered carefully in light of the methodological difficulties described previously.

Finally, the low methodological quality of all the studies included in the meta-analysis is an additional cause for interpreting all the results described here with caution.

Footnotes

*

Appendixes and evidence tables cited in this report are provided electronically at http://www​.ahrq.gov/clinic/tp/medittp​.htm