Adverse Events of Massage Therapy in Pain-Related Conditions: A Systematic Review

Pain-related massage, important in traditional Eastern medicine, is increasingly used in the Western world. So the widening acceptance demands continual safety assessment. This review is an evaluation of the frequency and severity of adverse events (AEs) reported mainly for pain-related massage between 2003 and 2013. Relevant all-languages reports in 6 databases were identified and assessed by two coauthors. During the 11-year period, 40 reports of 138 AEs were associated with massage. Author, year of publication, country of occurrence, participant related (age, sex) or number of patients affected, the details of manual therapy, and clinician type were extracted. Disc herniation, soft tissue trauma, neurologic compromise, spinal cord injury, dissection of the vertebral arteries, and others were the main complications of massage. Spinal manipulation in massage has repeatedly been associated with serious AEs especially. Clearly, massage therapies are not totally devoid of risks. But the incidence of such events is low.


Introduction
Massage, as any systematic form of touch or manipulation performed on the soft tissues of the body to provide comfort and promote health [1][2][3], has become popular in the United States and the rest of the world in recent decades. It has also been recommended by the Chartered Society of Physiotherapy for the management of various pain-related conditions, especially those of musculoskeletal origin [4], such as neck pain, low back pain, headache, and migraine [5][6][7][8]. This is supported by numerous systematic reviews of a large number of randomized controlled trials (RCTs) [9][10][11][12]. Between 2002 and 2007, the 1-year prevalence of use of massage by the US adult population increased from 5% (10.05 million) to 8.3% (18.07 million), and massage belongs to one of the most popular complementary and alternative medicine (CAM) therapies in the USA [13]. The increased use brings attention to the safety and quality of the modality.
A number of large surveys on the safety of massage have been conducted. Most reported incidents have been fairly minor, and incidence rates were low. For example, from surveys and review articles, the risk of a serious irreversible complication (e.g., stroke) for cervical manipulations has been reported to vary from one adverse event in 3020 to one in 1,000,000 manipulations, and another review of the articles on complications of spinal manipulation, which identified 295 complications, yielded estimates of vertebrobasilar accidents from one in 20 000 patients to one per 1,000,000 cervical manipulations and cauda equina syndrome to be less than one per 1,000,000 treatments [14][15][16]. The authors of these studies concluded that serious AEs seem to be rare and massage is generally a safe intervention. So this systematic review seeks to evaluate all published data (between 2003 and 2013) about adverse effects of massage therapy. We specifically hope to help the clinician feel comfortable and informed in conversations with their patients regarding the appropriate,  safe, and effective use of massage, not only in pain-related conditions.

Search Strategy.
We searched 6 databases in an attempt to locate all existing case reports (irrespective of language of publication) with original data on AEs following any type of massage therapy published between January 2003 and June 2013 in electronic form. PubMed including MEDLINE, EMBASE, The Cochrane Library (via Wiley), CNKI, CQVIP, and Wanfang digital databases were searched. Search terms were "massage, manual therapy, tuina, and chiropractic. " These terms were combined with "safe, safety, adverse event, adverse reaction, side effects, complications, and risk. "

Inclusion and Exclusion Criteria.
Only original case reports of complications or AEs of massage, manual therapy, and tuina published from January 2003 to June 2013 were included in this review. All those clinical study designs should be published in peer-reviewed journals, and like conference proceedings, cross-sectional and other descriptive designs and narrative reviews were excluded. Two coauthors independently screened the titles and abstracts of all papers found from the initial search. Disagreements between the two authors were resolved through discussion.
We excluded multiple inclusions and analyses of the same AEs as well as irrelevant studies. An irrelevant study was defined as a non-case report, such as a review, commentary, or clinical trial. Treatments not typically carried out by a massage therapist were also excluded, such as cardiac massage, prostatic massage, or carotid sinus massage. Adverse events related to massage oils, for example, allergies to aromatherapy oils or to the use of ice in conjunction with massage, were also excluded. All articles were evaluated and validated by one of the authors according to inclusion criteria.

Data Extraction.
Electronic database searches identified a total of 3282 articles for consideration. After screening, 126 potentially relevant articles were identified for full review, and 40 studies met inclusion criteria finally. There were 86 articles that were excluded for being unrelated to AEs or for having no details reported ( Figure 1). A full list of excluded articles is available from the corresponding author. When provided, we extracted author, year of publication, country of occurrence, participant related information (age, sex) or number of patients affected, the details of manual therapy, and clinician type that might have contributed to the AE, the reported AE, and its outcome. The data were extracted by two independent coauthors (P. Y. and NY. G.) and double checked to ensure matching and disagreements were resolved by consensus. Since there are no widely accepted criteria for judging the quality of AEs reports and the current studies' objective of describing case details, we did not assess the risk of bias on the included studies.

Results
The search strategy located 33 articles reporting a total of 43 case reports (in which the patients' age and/or sex were given) ( Table 1), and a total of 7 reports containing 95 AEs in case series associated with massage were identified ( Table 2). Most cases were reported from Asia especially in China ( = 24, 60% of total) and Europe (12, 30%), with few cases from the USA (3, 7.5%) and Australia (1, 2.5%), and more than half of the reported patients were female. There are 153 signs or symptoms of AEs in total, and the most common problems included disc herniation (

Discussion
Our primary objective in reviewing the case reports of AEs associated with massage has been to identify individual cases and outbreaks of AEs then to analyze their possible causes, in order to minimize the massage AEs in future and enhance the practice safety within the profession. Of the 138 cases involving the AEs following massage in 40 references (Tables  1 and 2), spinal manipulation has repeatedly been reported with serious AEs especially. Collectively, these data suggest that massage is associated with frequent, mild, and transient AEs, but sometimes it may also be indeed associated with serious complications which can lead to permanent disability or even death. Although important details of most cases are poorly reported or frequently missing, these results have clear clinical and research related implications comparatively. The true risk of injury due to spinal manipulation is still not known. Yet causal inferences may be not completely reasonable. Vascular accidents may happen spontaneously or could be caused by factors other than massage. The real serious incidence of AEs has been estimated to be ranging from 5 strokes in 100,000 manipulations to 1.46 case series in 10 million manipulations, and a rate of 2.68 deaths in 10 million manipulations has been reported [57][58][59]. The insurance industry claims [60] data support a risk of stroke as 1 per 2 million manipulations. 99% of all chiropractors practicing in Denmark completed a survey; they estimated that one case of cerebrovascular accident occurred for every 1.3 million cervical treatment sessions. The occurrence increased to 1 in every 900,000 treatment sessions for upper cervical manipulations, 8 Evidence-Based Complementary and Alternative Medicine and they noted that techniques using rotational thrusts were overrepresented in the frequency of injury.
A temporal relationship is insufficient to establish causality, and recall bias can further obscure the truth. Moreover, denominators are rarely available. Smaller randomized controlled trials (RCTs) are unlikely to detect rare AEs, and better reporting of AEs is required, obviously. Therefore Senstad et al. [61][62][63] reported the data from 3 prospective investigations of 1778 adults who received chiropractic spinal manipulation indicated that 30% to 55% reported a minor adverse event. The most common were local discomfort (53% to 60%), radiating discomfort (10% to 23%), headache (10% to 12%), tiredness (11%), or nausea; dizziness, hot skin, or "other" reactions are uncommonly reported (<5% of reactions). And of the reported reactions, reactions were mild or moderate in 85% to 90% of patients. 64% of reactions appeared within 4 hours of treatment, and 74% to 83% had disappeared within 24 hours. Interestingly, reactions are most commonly reported by women and (for both genders) at the beginning of the treatment series. Patients with long-lasting problems are more likely to report treatment reactions, and patients with no prior experience of chiropractic care do not report more reactions than patients previously treated by chiropractors. Then Cagnie et al. [58] recruited 465 new patients treated with spinal manipulation by 59 physiotherapists (Belgian). All patients were asked to complete a questionnaire about AEs subsequently. 61% of the patients reported at least one AE, most of which were mild and transient, like headache (20%), stiffness (20%), local discomfort (15%), radiating discomfort (12%), and fatigue (12%). 61% of the problems had started within 4 hours after manipulation; 64% had resolved within 24 hours. No complications with long-lasting consequences were reported. Hurwitz et al. [64] reported the AEs documented in a 280-patient RCT which compared spinal manipulation with spinal mobilization as treatments for neck pain. 30% reported at least one AE. Patients receiving spinal manipulation were more likely to experience AEs than mobilization. The most frequently noted AEs were increase of pain, headache, tiredness, and radiating pain. 80% of the AEs began within 24 hours after treatment and were mild or of medium severity. No serious complications were noted. The three prospective case series above corroborate the results from several earlier studies [65] showing that mild to moderate AEs occur in a large proportion of patients receiving spinal manipulation, but these AEs are transient and nonserious. And recently, 767 patients were randomized to one of three treatment arms in a new study [66], to investigate differences in occurrence of adverse events between three different combinations of manual treatment techniques used by manual therapists (i.e., chiropractors, naprapaths, osteopaths, physicians, and physiotherapists) for patients seeking care for back and/or neck pain. And adverse events were measured with a questionnaire after each return visit and categorized into five levels. As a result, the most common adverse events were soreness in muscles, increased pain, and stiffness. The most frequent level of adverse event in this study was short minor lasting less than 24 hours and was rated less than or equal to three on the numeric rating scale regarding severity. No serious adverse events were reported.
Clearly, we should differentiate between various approaches. The above cases suggest that massage by nonprofessional and forceful techniques is often associated with AEs. In 8 cases the practitioners are massage therapists (5.8% of total) and 33 are chiropractors (23.9%), while in the other cases (70.3%) they are unregistered or even healthcare professionals only. So it might be unfair to assess the AEs of spinal manipulation as practiced by well-trained chiropractors alongside that associated with the untrained. Obviously from above, a variety of different care providers like physiotherapists, massage therapists, physicians, and osteopaths may perform a manipulation as part of their practice, but it should be most frequently performed by chiropractors [67]. Certainly skill and experience are important, and it is relevant to differentiate between different professions. But on the other hand, skill is a quality not easily controlled and some therapists are more skilled than others. Moreover, this review is aimed at evaluating the AEs of an intervention (massage) and not that of a profession (massage therapist/chiropractic). That is why in this review we show the implicated practitioners are not only chiropractors but also physicians, physiotherapists, "bonesetters, " and general medical practitioners.
This systematic review has several limitations. Even though the search strategy was deemed thorough, some relevant published articles might have been missed. It is possible that not all cases were identified in our searches. Although this paper has resulted in a few papers to review, it still had its strengths including the thorough search of the literature to help reduce bias in the review. We searched multiple relevant electronic databases and used two coauthors to determine articles for inclusion in the review and to evaluate the literature. But because of the inherent nature of case reports and other anecdotal reports, it is impossible to make inferences regarding cause and effect. Therefore, it is not known whether the serious AEs in cases identified in this review were caused by massage and whether the association between therapy and event was accidental or not. So the safety in massage is still far from being achieved. Further investigations are urgent to assess definite conclusions regarding this issue. In the meantime, it should be necessary to establish a system of risk alert for guaranteed surveillance on this type of CAM and safe practice guidelines are required and could continue to be enforced.

Conclusions
In conclusion, although serious AEs associated with massage in general and pain-related massage in particular are few, massage therapies are not totally devoid of risks. Spinal manipulation in massage has repeatedly been associated with serious AEs especially. But the incidence of such events is probably low. Adequate regulation could further minimize the risks. So we recommend that not only adequate training in biomedical knowledge for practitioners, such as anatomy and microbiology, but also safe practice guidelines are required and should continue to be enforced in order to minimize massage AEs.