BACKGROUND. Spinal manipulation is a noninvasive manual procedure applied to specific body tissues with therapeutic intent. Although spinal manipulation is commonly used in children, there is limited understanding of the pediatric risk estimates.
OBJECTIVE. Our goal was to systematically identify and synthesize available data on adverse events associated with pediatric spinal manipulation.
METHODS. A comprehensive search was performed of 8 major electronic databases (eg, Medline, AMED, MANTIS) from inception to June 2004 irrespective of language. Reports were included if they (1) were a primary investigation of spinal manipulation (eg, observation studies, controlled trials, surveys), (2) included a study population of children who were aged 18 years or younger, and (3) reported data on adverse events. Data were summarized to demonstrate the nature and severity of adverse events that may result rather than their incidence.
RESULTS. Thirteen studies (2 randomized trials, 11 observational reports) were identified for inclusion. We identified 14 cases of direct adverse events involving neurologic or musculoskeletal events. Nine cases involved serious adverse events (eg, subarachnoidal hemorrhage, paraplegia), 2 involved moderately adverse events that required medical attention (eg, severe headache), and 3 involved minor adverse events (eg, midback soreness). Another 20 cases of indirect adverse events involved delayed diagnosis (eg, diabetes, neuroblastoma) and/or inappropriate provision of spinal manipulation for serious medical conditions (ie, meningitis, rhabdomyosarcoma).
CONCLUSIONS. Serious adverse events may be associated with pediatric spinal manipulation; neither causation nor incidence rates can be inferred from observational data. Conduct of a prospective population-based active surveillance study is required to properly assess the possibility of rare, yet serious, adverse events as a result of spinal manipulation on pediatric patients.
Spinal manipulation is a noninvasive manual procedure applied to specific body tissues with therapeutic intent. A variety of different care providers (eg, physiotherapists, massage therapists, physicians, osteopaths, naturopaths) may perform manipulation as part of their practice, but it is most frequently performed by chiropractors.1 The procedure is delivered by hand and can vary in velocity, amplitude, duration, frequency, location on the body, and direction of force. According to the American Chiropractic Association, chiropractic procedures specifically include the adjustment and manipulation of the articulations and adjacent tissues of the human body, particularly of the spinal column.2 Spinal manipulation, the mainstay of chiropractic care, “emphasizes the inherent recuperative power of the body to heal itself without the use of drugs or surgery.”3
Chiropractic medicine is the most common complementary and alternative medicine (CAM) practice used by children.4,5 Children made an estimated 30 million visits to US chiropractors in 1997.6 In Canada, a recent survey of 1804 participants in a university hospital pediatric emergency department in Toronto, Ontario, indicated that 31% of children sought chiropractic care.7 In addition, a cross-sectional survey of a random sample of 140 Canadian chiropractors (57% response rate) revealed that all the respondents also treated children (0–18 years old), and 13% of all visits over the preceding month involved children and youth.8 Although, chiropractors are trained to treat neuromuscular problems that stem from a mechanical disability within the body,9 children have been found to visit chiropractors for a variety of reasons including health promotion, musculoskeletal problems, asthma, otitis media, allergies, infantile colic, tonsillitis, attention-deficit/hyperactivity disorder, and enuresis.6,9–14
Spinal manipulation has been suggested as a potential cause of cerebrovascular accidents (eg, stroke) through mechanical injury to the vertebral artery. In adults, concerns regarding serious adverse events related to spinal manipulation have led to a series of studies examining risk.15–17 For example, a review of the literature on complications of spinal manipulation, which evaluated case reports, surveys, and review articles, identified 295 complications, yielding estimates of vertebrobasilar accidents from 1 in 20000 patients to 1 per 1 million cervical manipulations, and cauda equina syndrome to be <1 per 1 million treatments.17–19 Data from 4 prospective investigations of 2058 adults who received chiropractic spinal manipulation indicated that 30% to 55% reported a minor adverse event (eg, local discomfort, additional pain, stiffness, headache, fatigue, fainting).20–23
Despite the fact that spinal manipulation is widely used on children, pediatric safety data are virtually nonexistent. Consequently, some pediatricians believe that the use of spinal manipulation on children is dangerous and advise against its use,24–26 whereas other health care practitioners and many parents continue with this practice. There is an urgent need to quantify the risk associated with spinal manipulation in children.
A comprehensive search was developed by a clinical librarian in collaboration with content experts to identify all relevant reports regardless of publication status. The following electronic databases were searched: Central (second quarter, 2004), Medline (1966–2004), PubMed (1966–2004), Embase (1988–2004), CINAHL (1982–2004), AltHealthWatch (1990–2004), MANTIS (1900–2005), and ICL (1985–2004) from inception to June 2004 irrespective of language. For a copy of our search strategy, please contact the corresponding author. The primary authors of relevant articles and experts in the area of spinal manipulation were contacted for information on additional studies. In addition, reference lists of relevant articles were examined.
Two reviewers (Dr Johnston and Ashish Mahajan, BSc) independently reviewed the titles and abstracts (where available) of all articles generated from the electronic and grayliterature search. The full manuscripts of reports relevant to adverse events associated with spinal-manipulation criteria were retrieved. Independent reviewers (Dr Johnston, Ms Cramer, and Denise Adams, BSc) assessed the full articles of each potentially relevant study by applying the following predetermined set of eligibility criteria: (1) the study was a primary investigation/report (ie, case reports, case series, case control, randomized, controlled trials, and survey or surveillance studies); (2) part or all of the study population was 18 years or younger and; (3) adverse events were reported. We searched for adverse events related to delivery of spinal manipulation (eg, pain, weakness, disability), and noted adverse events related to delayed or missed diagnoses when they were described. Inclusion was not limited by the condition studied, provider of the spinal-manipulation intervention (eg, chiropractor, osteopath, physiotherapist, physician), or the comparison intervention.
Independent reviewers (Dr Johnston, Ms Cramer, and Denise Adams, BSc) used a structured data-extraction form to independently extract data. The key data extracted were type of publication/report (eg, case report, randomized, controlled trial), participant characteristics (eg, age and gender), previous medical diagnosis/diagnoses, type of manipulation (eg, light fingertip pressure, mobilization, flexion, rotation, high-velocity/low-amplitude), location of manipulation (eg, cervical), schedule of manipulation, outcome, and timing of the adverse event in relation to therapy. Any discrepancies between reviewers were discussed and resolved by referring to the original report and, if necessary, consultation with a third reviewer.
Adverse events were classified by using the following categories: severe (indicating hospitalization, permanent disability, mortality), moderate (transient disability, involving seeking medical care but not hospitalization), minor (self-limited, did not require additional medical care), and delayed diagnosis or treatment (moderate to severe adverse event [as defined above] not directly related to the administration of a spinal manipulation but as a result of delayed diagnosis or treatment of a medical condition [eg, meningitis]). Adverse events were tabulated by using descriptive statistics. A priori, we planned to summarize adverse events derived from randomized, controlled trials using risk differences (with corresponding 95% confidence intervals) because events were considered rare; we planned to summarize data (and conduct subgroup analyses if possible) according to severity and type of adverse event, timing of the adverse event, age group (birth to 5 years, 6–13 years, 14–18 years), type and schedule of manipulation, location of manipulation, and type of practitioner.
Electronic database searches identified a total of 13916 articles for consideration. After screening, 164 potentially relevant articles were identified for full review. An additional 68 potentially relevant articles were identified from review of the reference lists and by contacting authors of included reports and experts in field. Thirteen studies met inclusion criteria (10 English, 2 French, 1 German): 2 randomized, controlled trials, 4 case series, and 7 case reports (Fig 1). There were 212 articles that were excluded: 77 were not primary studies, 70 did not involve pediatric patients, 29 did not involve spinal manipulation, 32 did not report an adverse event, and 4 were irretrievable. A full list of excluded articles is available from the corresponding author.
We identified 14 cases of direct adverse events (Table 1), 27–36 9 of which were classified as serious and resulted in hospitalization, permanent disability, or death.27–33 In 10 of 14 cases, the adverse events occurred within 24 hours of spinal manipulation.27,29,31–33,35–37 We identified an additional 20 cases of delayed diagnosis and/or inappropriate provision of chiropractic care (type of spinal manipulation unspecified in all cases) that resulted in indirect adverse events.38–41 Seven cases involved delayed treatment of cancer (eg, diabetes, aggressive osteosarcoma, metastatic neuroblastoma).40,41 Two cases involved delayed treatment for meningitis, and 1 case involved delayed treatment for embryonal rhabdomyosarcoma.38 Each of the latter 3 cases resulted in death38 (Table 2).
We were unable to combine data or conduct a priori subgroups analyses because of methodologic heterogeneity between trials.35,36 Below, we describe the 2 clinical trials and the nature and severity of adverse events related to pediatric spinal manipulation. The remaining direct and indirect adverse events are described in Tables 1 and 2.
Two trials were included.35,36 The first trial randomly assigned 171 children to 1 of 2 active treatment groups; both groups received spinal manipulation (the only discernable difference was that 1 group waited 2 weeks to start therapy).35 Two adverse events were reported, both of moderate severity (ie, required medical attention). One case involved the onset of severe headaches and stiff neck after cervical manipulation, which improved gradually over the next 2 weeks with additional soft tissue therapy. Neither the parent nor the child could recall any previous symptoms involving serious headaches or stiff neck. A second case involved the onset of acute lumbar pain postmanipulation, which also resolved.35 The second study was a randomized, controlled trial of 20 children (9 received treatment and 11 served as controls) that reported 2 minor adverse events involving 1 case of midback soreness that resolved after a few days and 1 case of irritability for a short period posttreatment.36 The authors concluded that the patients tolerated the treatments well with only minimal, self-limiting adverse effects.36
To our knowledge, we report the first systematic review of adverse events related to pediatric spinal manipulation irrespective of language. We identified 14 adverse events in 10 reports, 9 of which were serious and resulted in hospitalization, permanent disability, or death.27,29–34 An additional 20 cases of delayed diagnosis and/or inappropriate provision of spinal manipulation resulted in indirect adverse events.38,40,41 Case reports and case series can be interpreted as spontaneous reporting or “passive” surveillance. Although they are useful to demonstrate the type and nature of adverse events, these reports do not provide information on the incidence of adverse events because of the lack of data regarding the total number of manipulations provided (ie, denominator data).
Spontaneous reporting of adverse events is well known to underestimate risk. An example of the limitations of passive surveillance was documented by a British survey of neurologists that was meant to ascertain cases of serious neurologic complications occurring 24 hours post–cervical manipulation, in which 24 respondents identified 35 such cases over the previous year; none of these had been reported previously.42 Lessons from spontaneous reports of adverse drug reactions suggest that <10% of serious adverse events are reported.43 The serious concerns regarding both the quantity and quality of these spontaneous reports limit assessment of causation. Given the large numbers of children who have received spinal manipulation during the decades assessed by our search strategy, adverse events resulting from spinal manipulation are either remarkably rare or underreported.
A number of risk factors may predispose a child to an adverse event as a result of spinal manipulative procedures, including immaturity of the spine, rotational manipulation of the cervical spine, and high-velocity spinal manipulations.44–47 We found that all 9 serious adverse events (eg, death resulting from subarachnoid hemorrhage, paraplegia, etc) occurred in children under 13 years of age.30,32,34 In a case series, Ragoet32 presented 3 cases of dislocated atlas as a result of pediatric spinal manipulation. Evidence suggests that there is a strong correlation between severity of injury to the spinal cord and the immaturity of the spine44 and that the atlas (cervical vertebra 1 [C1]) and dens of the axis (C2) of children are more vulnerable to trauma than those of adults.45 Although 5 of our serious adverse-event reports did not specify the type of spinal manipulation used, 2 of the 4 that reported serious adverse events specified that the practitioner used rapid and/or strong rotational maneuvers.27,31 The majority of complications attributed to spinal manipulative therapy have occurred as a result of rotational manipulation of the cervical spine.46 In addition, high-velocity manipulations of the spine have the potential for serious complications resulting from diagnostic error/inadequate patient assessment.47 Although the authors did not clearly specify the type of spinal manipulation provided, 2 of the severe adverse-event reports identified underlying risk factors (spinal cord astrocytoma, congenital occipitalization) that may have predisposed the child to the subsequent serious adverse event (ie, quadriplegia, unsteady gait).29,34 An error in the diagnosis of any number of preexisting conditions such as arteritis, arthritic and cardiac conditions, clotting abnormalities, meningitis, or vertebrobasilar insufficiency may predispose children to neurologic and/or vertebrobasilar complications.45,48–50
A major challenge in proving or refuting causation between pediatric spinal manipulation and serious adverse events is the lack of sufficient randomized trials. Unfortunately, like randomized trials of conventional treatment, many randomized trials of CAM fail to adequately evaluate for potential adverse events.51,52 In particular, the 2 trials included in this review failed to adequately describe most of the 10 recommendations on reporting harms-related issues suggested by the Consolidated Standards of Reporting Trials (CONSORT) statement.35,36,53 Moreover, developing risk estimates for rare events requires population-based sampling. For this reason, our review was not limited to randomized, controlled trials but assessed all reported primary medical literature including observational studies and reports. It is concerning that modeling from reports of adverse drug reactions suggests that more than 1 to 3 spontaneous case reports of rare or uncommon adverse events is unlikely to be coincidental.54,55
The only other previous review of this topic was narrative, not systematic.56 The authors neglected to include a number of potentially important databases (eg, Central, Embase, ICL) and did not search for non-English reports.57–59 Although they concluded that the risk of neurologic and/or vertebrobasilar complications from chiropractic manipulation was 1 in 250 million pediatric visits,56 we feel that this estimate is inaccurate and likely underestimates risk. Numerator data were derived from incomplete assessment of cases identified in the medical literature, and denominator data were based on an estimated number of chiropractic visits made by children in the United States. Although our search strategy was more comprehensive, we did not feel comfortable creating risk estimates with an uncertain denominator. We urge the development of an active surveillance model to prospectively gather data about the quantity and quality of adverse events so that risk estimates can be made with greater precision.
It is difficult to know the cause for the identified indirect adverse events (eg, delayed diagnosis and/or inappropriate provision of spinal manipulation). We postulate that this is related to lack of sufficient pediatric training for CAM providers. We have recently collected survey data on the knowledge, attitudes, and behavior of chiropractors and osteopaths with regards to children in their practice. Of 287 respondents, 71% graduated between 1992 and 2002, 78% of the respondents identified 1 semester or less of formal pediatric education during their training, 72% received minimal or no pediatric clinical training, and 93% recommended increased pediatric training in their schools.60 Collaborating with experts in pediatric education toward developing a standardized pediatric curriculum for CAM providers may offer a way forward. Such collaboration should involve the development of guidelines for medical referrals, joint integrative care between physicians and CAM providers,61 and the development of a scope of practice for pediatric chiropractic and osteopathic care. Despite what some have advised,24,26 many children continue to visit chiropractors, and many chiropractors continue to treat children.8 We believe collaboration of this nature would result in improved patient safety.
Our study has several limitations. First, we uncovered mostly case reports. Many of the cases contained limited data, and assessment of the validity of case reports is generally insufficient to reach conclusions regarding causality.52,54 However, to exclude the case reports would have severely biased our results because it would not have allowed for the identification of potentially uncommon and unexpected adverse events,54,62,63 which may differ from those detected in clinical trials.64 Second, our search strategy did not allow for systematic identification of indirect adverse events (eg, delayed or missed diagnoses); therefore, these indirect adverse events are likely underestimated. Concerns regarding chiropractic care have been raised about advice given regarding childhood immunizations, frequency of radiographs, recommendations regarding dietary supplements, and lack of familiarity with serious childhood conditions resulting in delayed diagnosis of a serious medical condition.6 These concerns were incompletely captured by our review. Finally, we excluded cases that were only identified in medicolegal proceedings or the lay press.65–67 Our search strategy was not designed to identify such reports, and to include these cases would create a false impression that they were the only ones reported in the potentially vast non–health-specific, non–peer-reviewed resources.
Spinal manipulation is common among children, and although serious adverse events have been identified, their true incidence remains unknown. Randomized, controlled trials will likely reveal common minor adverse events,20,35,36 and these events must be better reported. Prospective population-based studies are needed to identify the incidence of serious rare adverse events associated with spinal manipulation. Patient safety demands a greater collaboration between the medical community and other health care professionals, particularly chiropractors, such that we can investigate and report harms related to spinal manipulation together. In the interim, clinicians should query parents and children about CAM usage and caution families that although serious adverse events may be rare, a range of adverse events or delay in appropriate treatment may be associated with the use of spinal manipulation in children.
Dr Vohra is an Alberta Heritage Foundation for Medical Research Population Health Investigator and recipient of a Canadian Institute of Health Research New Investigator Award. Dr Johnston holds a Sick Kids Foundation Duncan L. Gordon Fellowship.
We thank Ellen T. Crumley for conducting the electronic searches; Ashish Mahajan and Denise Adams for assistance with screening and data extraction; Linda Slater and Dominic Allain for assistance with translation; and Don Spady for constructive feedback.
- Accepted July 20, 2006.
- Address correspondence to Sunita Vohra, MD, FRCPC, MSc, CARE Program, Associate Professor of Pediatrics, Stollery Children's Hospital, University of Alberta, Edmonton, Alberta, Canada T6G 2E1. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
Dr Vohra developed the study question, designed the study, oversaw its execution, and participated in data analysis and manuscript preparation; Dr Johnston coordinated the study, lead relevance screening, data-extraction, and data analysis, and drafted and revised the manuscript; Ms Cramer contributed to protocol development and study design and participated in relevance screening, data extraction, and manuscript preparation; Dr Humphreys provided expertise in chiropractic practice, spinal manipulation, and chiropractic literature, reviewed the search strategies, and participated in developing the data-extraction form and data analysis; and all the authors read and approved the final manuscript.
- ↵American Chiropractic Association. About us. Available at: www.amerchiro.org/level2_css.cfm?T1ID=10&T2ID=117. Accessed April 24, 2006
- ↵Association of Chiropractic Colleges. Chiropractic paradigm. Available at: www.chirocolleges.org/paradigm_scopet.html. Accessed April 24, 2006
- ↵Barnes PM, Powell-Griner E, McFann K, Nahin RL. Complementary and alternative medicine use among adults: United States, 2002. Adv Data.2004;(343):1–19
- ↵Spigelblatt L, Liane-Ammara G, Pless B, Guyver A. The use of alternative medicine by children. Pediatrics.1994;94 :811– 814
- ↵Goldman RD, Vohra S. Complementary and alternative medicine use by children visiting a pediatric emergency department. Can J Clin Pharmacol.2004;11 :e245– e256
- ↵Lewit K. Manipulative Therapy and Rehabilitation of the Locomotor System. 2nd ed. Oxford, United Kingdom: Butterworth-Heinemen; 1996
- Rothwell DM, Bondy SJ, Williams SJ. Chiropractic manipulation and stroke: a population-based case-control study. Stroke.2001;32 :1054– 1060
- Haynes MJ. Stroke following cervical manipulation in Perth. Chiropr J Aust.1994;24 :42– 46
- Waimer S. Pediatric chiropractic. J Paediatr Child Health.2002;7 :114– 115
- ↵Departments of Pediatrics of Pediatric Hospitals in Canada. Children and chiropractors. Can J Pediatr.1995;2 :5– 6
- ↵Jocobi G, Riepert TH, Kieslich M, Bohl J. Case of death during physical therapy according to Vojta [in German]. Z Physiother.2001;53 :573– 576
- ↵National Institute on Developmental Delays. Vojta therapy: history. Available at: www.nidd.org/therapy_vojta.htm. Accessed November 26, 2005
- ↵Zimmerman AW, Kumar AJ, Gadoth N, Hodges FJ 3rd. Traumatic vertebrobasilar occlusive disease in childhood. Neurology.1978;28 :185– 188
- ↵Held JP. Dangers of cervical manipulation in neurology [in French]. Ann Med Phys (Lille).1966:251– 259
- ↵Smith RL. At Your Own Risk: The Case Against Chiropractic. New York, NY: Simon & Schuster; 1969
- ↵Stevenson C, Honan W, Cooke B, Ernst E. Neurological complications of cervical spine manipulation. J R Soc Med.2001;94 :107– 110
- ↵Reid DC, Robinson BE. Contra-indications and precautions to spinal joint manipulations: a review. Can J Rehabil.1988;2 :71– 78
- ↵Greenman PE. Principles of manipulation of the cervical spine. J Manual Med.1991;6 :106– 113
- ↵Terret AGJ, Kleynhans AM. Complications from manipulation of the low back. Chiropr J Aust.1992;22 :129– 140
- ↵Ganesan V, Chong WK, Cox TC, Chawda SJ, Prengler M, Kirkham FJ. Posterior circulation stroke in childhood: risk factors and recurrence. Neurology.2002;59 :1552– 1556
- ↵Pistolese RA. Risk assessment of neurological and/or vertebrobasilar complications in the pediatric chiropractic patient. J Vertebral Subluxation Res.1998;2 :77– 85
- Pham B, Klassen TP, Lawson ML, Moher D. Language of publication restrictions in systematic reviews gave different results depending on whether the intervention was conventional or complementary [published correction appears in J Clin Epidemiol. 2006;59:216]. J Clin Epidemiol.2005;58 :769– 776
- ↵Greenhalgh T, Peacock R. Effectiveness and efficiency of search methods in systematic reviews of complex evidence: audit of primary sources. BMJ.2005;331 :1064– 1065
- ↵Adams D, Amernic H, Humphreys K, et al. A survey of complementary and alternative medicine practitioners' knowledge, attitudes, and behavior regarding children in their practice. Paper presented at: Pediatric Academic Society Meeting; San Francisco, CA; April 29–May 2, 2006
- ↵Stricker BH, Psaty BM. Detection, verification, and quantification of adverse drug reactions. BMJ.2004;329 :44– 47
- ↵Dalgaard JB. Medico-legal aspects of manipulatory treatment in Denmark. J Manual Med.1991;6 :114– 116
- Vogel B. Wendell v Greenback: Chiropractic malpractice—negligent spinal adjustment and herniation of disc at C6-7: Defense verdict. Sacramento, CA: Sacramento County Superior Court; 1998
- ↵Smith TK. A new market: chiropractors seeking to expand practices take aim at children. Wall Street Journal (Eastern Edition).1993:A1
- Copyright © 2007 by the American Academy of Pediatrics