Adverse Events Associated with Yoga: A Systematic Review of Published Case Reports and Case Series

While yoga is gaining increased popularity in North America and Europe, its safety has been questioned in the lay press. The aim of this systematic review was to assess published case reports and case series on adverse events associated with yoga. Medline/Pubmed, Scopus, CAMBase, IndMed and the Cases Database were screened through February 2013; and 35 case reports and 2 case series reporting a total of 76 cases were included. Ten cases had medical preconditions, mainly glaucoma and osteopenia. Pranayama, hatha yoga, and Bikram yoga were the most common yoga practices; headstand, shoulder stand, lotus position, and forceful breathing were the most common yoga postures and breathing techniques cited. Twenty-seven adverse events (35.5%) affected the musculoskeletal system; 14 (18.4%) the nervous system; and 9 (11.8%) the eyes. Fifteen cases (19.7%) reached full recovery; 9 cases (11.3%) partial recovery; 1 case (1.3%) no recovery; and 1 case (1.3%) died. As any other physical or mental practice, yoga should be practiced carefully under the guidance of a qualified instructor. Beginners should avoid extreme practices such as headstand, lotus position and forceful breathing. Individuals with medical preconditions should work with their physician and yoga teacher to appropriately adapt postures; patients with glaucoma should avoid inversions and patients with compromised bone should avoid forceful yoga practices.


Introduction
Yoga is rooted in Indian philosophy and has been a part of traditional Indian spiritual practice for around 3000 years [1]. While the goal of yoga has been described as uniting mind, body, and spirit, it has now become a popular means to promote physical and mental well-being [1,2]. While yoga traditionally also comprises advice for ethical lifestyle and spiritual practice [1][2][3][4], it is most often associated with physical postures (asanas), breathing techniques (pranayama), and meditation (dyana) in North America and Europe [2]. These more physically-oriented yoga forms are gaining increased popularity as a therapeutic practice: in 2008, about 15% of the American adult population reported practicing yoga or being at least strongly interested in it [5]. Of those who were already practicing yoga, about half started practicing explicitly to improve their health status, resulting in more than 13 million people practicing yoga for health reasons [6,7]. It has been estimated that about 30 million people are regularly practicing yoga worldwide [8]. Yoga has also been recognized as medical therapy: about 14 million Americans (6.1% of the population) reported that yoga was recommended to them by a physician or other therapist [5].
While yoga has often been regarded as beneficial and without harm, this view has been challenged in recent years. Mainly based on anecdotal evidence, the safety of yoga has been questioned in a number of lay-press articles [9][10][11]. In particular, a recent New York Times article by William J.
Broad has listed a number of alarming cases of yoga-associated injuries [11]. As these publications seem to have led to a general uncertainty among yoga practitioners and those interested in starting practice [12], it is important to systematically assess the safety of yoga. Therefore, this review aims to assess published case reports and case series on yoga-associated adverse events in order to analyze a) which adverse events were most often reported, b) which yoga forms and specific practices were most often associated with adverse events, and c) which persons (e.g. those which specific preconditions) were most often reported to be affected.

Eligibility Criteria
Original English or German language case reports and case series were eligible if they were published in a peer-reviewed journal and reported on yoga-associated adverse events in healthy humans or human patients. Non-case reports such as clinical trials, reviews, basic research, or commentaries were excluded. A specific practice was regarded as 'yoga' if a) it was explicitly labeled as yoga by the authors, b) it was labeled with the name of a specific yoga practice, and/or c) the described practice clearly resembled typical yoga practices. Adverse events were classified as yoga-associated if they appeared in temporal connection with yoga practice and/or a causal relationship was assumed by the authors of the report.

Data Extraction and Management
For case reports, data were extracted on time of publication, country of origin, age and gender of the case, the specific yoga practice and yoga posture or breathing techniques, and the experience of the practitioner. Data on the reported adverse event, its treatment and clinical outcome were also extracted. For case series, the time of publication, origin, number of cases, the cases age and gender, the specific yoga practice and yoga posture or breathing techniques, the reported adverse event, its treatment and clinical outcome were collected.

Literature Search
The literature search revealed a total of 517 non-duplicate records of which 469 were excluded because they did not report on yoga practices, were not case reports or case series or did not report adverse events. Out of 48 full-texts assessed for eligibility, 11 articles were excluded because they were not on yoga [13][14][15][16][17][18][19][20], were not case reports or case series [21][22][23], or were double publications on the same case [24]. Finally, 35 case reports  and 2 case series reporting on a total of 76 unique cases were included [60,61] (Figure 1).

Reported Cases
Characteristics of the included case reports and case series are shown in   Taiwan [58], South Korea [30], and Australia [50]. The first included report was published in 1969, the number of reports published each year gradually increased until 2012 ( Figure 2). Of the 76 cases, 66 had no preconditions that were associated with the adverse events while 9 case reports described an aggravation of existing preconditions, i.e. 3 cases of glaucoma [25,35,38], 3 cases of osteopenia [61], and 1 case each of asthma [55], psychosis [45], and affective disorder [59]. One case had a congenital hyperelasticity of connective tissue which might have facilitated the occurence of adverse events [34]. Fifty-one cases were female, 25 male; the mean age was 44.23 years.
The yoga practice that was most often associated with reported adverse events was Pranayama or yoga breathing with 4 reported cases [40,41,54,55], followed by Hatha yoga (an umbrella term for physical yoga practices) [25,31,48] and Bikram yoga [43,45,51] with 3 cases each. Siddha yoga meditation [47,56] and Vinyasa yoga (a yoga practice that involves flowing sequences of yoga postures synchronized to the breath) [28] were practiced in 2 and 1 cases, respectively. The other case reports or case series did not report the specific yoga practice.

Discussion
This systematic review included 76 unique cases of yogaassociated adverse events. Most adverse events affected the musculoskeletal, nervous, or visual system. More than half of the cases for which clinical outcomes were reported reached full recovery, 1 case did not recover at all, and 1 case died. Headstand was by far the most often cited yoga posture; and Pranayama and  Bikram yoga were the yoga practices that were most often associated with adverse events. Incidence rates of adverse events associated with yoga are best estimated from large prospective surveys of practitioners. However, these data are rare. In a small survey in 110 Finnish Ashtanga Vinyasa Yoga practitioners, 62% of respondents reported at least one yoga-related musculoskeletal injury, mainly sprains and strains [62]. About half of those reported full recovery, the other half partial recovery. Ashtanga Vinyasa Yoga is a physically demanding yoga style that uses standardized sequences of physical yoga postures with synchronized breathing [62]. More recently, in a large national survey, 78.7% of about 2500 Australian yoga practitioners indicated that they had never been injured during yoga [63]. The remaining practitioners mainly reported minor injuries. 4.6% of respondents had been injured in the past 12 months; 3.4% reported injuries that occurred under supervision. In accordance with the present systematic review, the postures that were most commonly associated with injuries were headstand, shoulder stand and variations of the lotus pose [63]. A survey in more than 1300 mainly North American yoga teachers and therapists found that respondents considered injuries of the spine, shoulders, or joints the most common; many respondents regarded yoga as generally safe and associated adverse events with excessive effort, inadequate teacher training, and unknown medical preconditions [64]. Systematic reviews on clinical trials on yoga interventions generally found insufficient reporting of safety data [65][66][67][68]. However, if adverse events were reported, they could mostly be classified as non-serious [65][66][67].
Out of 76 cases in the present review, 1 fatality was reported [33]. However, the practice described was ''voluntary mouth-tomouth Yoga breathing exercises'', which can hardly be characterized as a typical yoga practice. This practice is not described in any standard handbook of yoga practices [1,69]. Moreover, postmortem toxicological studies revealed significant levels of longacting barbiturates that can be argued to be at least partially responsible for her death. Another case report reported a neuropathy being caused by falling asleep in a seated forward bend due to opioids and tricyclic antidepressants [57]. As yoga requires awareness and concentration [70,71], it is recommended that practitioners abstain from using alcohol or recreational drugs during practice in order to avoid adverse events.
Several of the reported adverse events occurred in yoga teachers [27,32,45], who can be assumed to practice more intensely and more often than non-teachers. The yoga postures that were most often associated with adverse events were headstand, shoulder stand, postures that required putting 1 or both feet behind the head, and variations of the lotus position. All these postures can be considered advanced postures that should normally not be practiced by beginners or individuals with medical preconditions [1]. So-called inversions like headstand and shoulder stand are often regarded as a special category of yoga postures that should be practiced only by experienced practitioners, with extreme care.   [1,72]. Two of the 3 cases who had practiced shoulder stand [36,52] and 8 of the 10 cases that had practiced head stand [25,32,35,37,38,46,52,53] reported orbital adverse events, mainly glaucomatous symptoms. It has been reported that headstand induces a twofold increase in intraocular pressure [73]. However, intraocular pressure returned to baseline values immediately after headstand and no association of regular yoga practice with chronically increased intraocular pressure was found [73]. Therefore, beginners should be exceedingly cautious with inversions, which may be contraindicated for individuals with a history or positive family history of glaucoma. Voluntary vomiting is a common Kriya or cleansing technique in traditional yoga [69]. It is however very rarely practiced in North America or Europe [2]. As a case of intermittent reflux symptoms [27] and another one of dental erosion [48] -both of which originated from India -can be assumed to be directly related to regular vomiting, and the postulated cleansing properties of the practice are not in accordance with biomedical science, this practice should be discouraged in general.
Further, 4 adverse events were associated with yoga breathing, or pranayama. While gentle forms of yoga breathing, such as the relaxed abdominal breath, may be appropriate for beginners, extreme forms that involve holding or forcing the breath are considered an advanced yoga practice that should not be done by those new to yoga. [1,22,74]. None of the respective case reports stated the length of practice of the affected individual [40,41,54,55]. Yoga practitioners should be advised to be careful when practicing pranayama and perhaps not start practicing forceful techniques such as Kapalabathi, i.e. a practice that resembles hyperventilation, before they have gained a considerable body control and have mastered easier breathing techniques [1,74]. People with medical conditions should consult their physician regarding the appropriateness of extreme breathing techniques.
Bikram yoga is a modern yoga style that includes traditional Hatha yoga practices in a room heated to 105uF with a humidity of 40% [75]. Bikram yoga is a very intense physical yoga practice that includes forceful exercise and competition [75]. At least 1 of the 3 Bikram yoga-associated adverse events, a hyponatriaemia due to excessive fluid replacement after intensive sweating [51], can be directly related to the specific conditions in Bikram yoga and cannot be transferred to other yoga styles. The extreme heat and intensity of the Bikram yoga practice may make this style of yoga inappropriate for older adults and people with medical conditions.
The majority of cases were female and the number of reports published each year gradually increased from 1969 to 2012. These findings reflect general characteristics of yoga practitioners. About 75% of all yoga users are female [76,77] and yoga is gaining increased popularity over time: in 1994, about 5 million American Adults practiced yoga [78], by 2002, more than 10 million [7], and by 2007, more than 13 million [6]. Most cases included in this review originated from the USA. While there are no reliable data on prevalence of yoga use outside the USA, this might reflect a presumable higher prevalence of yoga use in the USA compared to most other countries worldwide [79].
There are several limitations in this review. Only case reports and case series that were published in peer-reviewed journals were included to ensure a certain quality of assessment and reporting. However, cases that were published in grey literature might have enhanced the findings of the review. Moreover, the quality of reporting in the included case reports and case series generally was low. Only few reports described the specific yoga form practiced or the practice experience of the case. Even more critically, for about 2 thirds of reported cases, no information on clinical outcomes was provided. This makes it hard to estimate the number of nonrecovered or only partially recovered cases; information that is crucial for assessing the safety of yoga. Case reports and case series are anecdotal by nature. Therefore, this systematic review is unable to estimate the total number or frequency of adverse events associated with yoga.

Conclusions
As any other physical or mental practice, yoga is not without risk. However, given the large number of practitioners worldwide [6][7][8], only relatively few serious adverse events have been reported in healthy individuals. Therefore, there is no need to discourage yoga practice for healthy people. It has however been stressed that yoga should not be practiced as a competition and that yoga teachers and practitioners should never push themselves (or their students) to their limits [9]. Beginners should avoid advanced postures such as headstand or lotus position and advanced breathing techniques such as Kapalabathi. Practices like voluntary vomiting should perhaps be avoided completely.
As yoga has been shown to be beneficial for a variety of conditions [65,66,68], it can also be recommended to patients with physical or mental ailments, as long as it is appropriately adapted to their needs and abilities and performed under the guidance of an experienced and medically trained yoga teacher. Especially, patients with glaucoma should avoid inversions and patients with compromised bone and other musculoskeletal disorders should avoid forceful or competitive yoga forms. Yoga should not be practiced while under the influence of psychoactive drugs.

Supporting Information
Checklist S1 PRISMA Checklist.