Spine (Phila Pa 1976). 2013 (Sep 15); 38 (20): 1723–1729
Bruce F Walker, Jeffrey J Hebert, Norman J Stomski, Brenton R Clarke,
Ross S Bowden, Barrett Losco, Simon D French
School of Health Professions
Murdoch University, Murdoch, Australia
STUDY DESIGN: Blinded parallel-group randomized controlled trial.
OBJECTIVE: Establish the frequency and severity of adverse effects from short-term usual chiropractic treatment of the spine when compared with a sham treatment group.
SUMMARY OF BACKGROUND DATA: Previous studies have demonstrated that adverse events occur during chiropractic treatment. However, as a result of design limitations in previous studies, particularly the lack of sham-controlled randomized trials, understanding of these adverse events and their relation with chiropractic treatment is suboptimal.
METHODS: We conducted a trial to examine the occurrence of adverse events resulting from chiropractic treatment. It was conducted across 12 chiropractic clinics in Perth, Western Australia. The participants comprised 183 adults, aged 20 to 85 years, with spinal pain. Ninety-two participants received individualized care consistent with the chiropractors' usual treatment approach; 91 participants received a sham intervention. Each participant received 2 treatments.
RESULTS: Completed adverse questionnaires were returned by 94.5% of the participants after appointment 1 and 91.3% after appointment 2. Thirty-three percent of the sham group and 42% of the usual care group reported at least 1 adverse event. Common adverse events were:
|increased pain||sham 29%||usual care 36%|
|muscle stiffness||sham 29%||usual care 37%|
|headache||sham 17%||usual care 9%|
The relative risk (RR) was not significant for adverse event occurrence (RR = 1.24; 95% CI: 0.85-1.81), occurrence of severe adverse events (RR = 1.9; 95% CI: 0.98-3.99), adverse event onset (RR = 0.16; 95% CI: 0.02-1.34), or adverse event duration (RR = 1.13; 95% CI: 0.59-2.18). No serious adverse events were reported.
CONCLUSION: A substantial proportion of adverse events after chiropractic treatment may result from natural history variation and nonspecific effects.
Key words: back pain, neck pain, adverse effects, chiropractic,
manipulation, spinal, randomized controlled trial, clinical trial,
placebo effects, nocebo effects.
Level of Evidence: 2 Level of Evidence: 2
From the FULL TEXT Article:
Chiropractic therapy is commonly used to manage musculoskeletal
conditions in high-income countries. [1, 2]
The occurrence of adverse events resulting from chiropractic
treatment is of considerable interest to chiropractors
and the general public. Most adverse events associated
with chiropractic treatment are mild, short lasting, and typical
of musculoskeletal condition symptoms. [3–11] However,
due to a lack of appropriately designed studies, particularly
sham-controlled trials, there are differences in views about
what constitutes a chiropractic treatment-related adverse
The occurrence of adverse events after chiropractic treatment
has been examined in 1 randomized controlled trial, 
5 prospective single-arm studies, [4–8] and 3 retrospective studies. [9–11]
These studies reported that 34% to 61% of participants
experienced at least 1 adverse event. [4–8] Most events
were benign, transient, and typically consisted of increased
pain, muscle stiffness, tiredness, headache, and radiating discomfort. [4–8] Less common events were dizziness, nausea, tinnitus,
and impaired vision. More serious adverse events associated
with chiropractic treatment, including disc injury, cauda
equina syndrome, fracture, and stroke, have been reported
but the rate has not been robustly established. [9, 11–13]
Predictors of adverse events have been identified in 4 previous
studies of chiropractic treatment. [6–8, 14] These predictors
included female sex, [6, 8] age (27–46 yr),  high-velocity manipulation
(compared with low-velocity mobilization),  first treatment
session,  medication use,  more than 1 region treated or 
only thoracic spine treated, and" treatment including cervical
rotation,  work status,  and general practitioner visit in previous
6 months.  Notably, none of the identified predictors have
been found to influence adverse events consistently across
Several limitations constrain the findings of previous studies.
In the prospective studies, the chiropractors providing
treatment also administered the questionnaires, which possibly
resulted in underreporting of adverse events. [4–8] In addition,
recall bias may also have led to an underestimation of
adverse events in the retrospective studies.  Moreover, all
previous studies lacked a sham intervention, which may have
resulted in an overestimation of adverse events as some events
could have been associated either with natural history or with
nonspecific effects. Therefore, the estimates of adverse events
resulting from the specific effect of treatment are not known.
What is known is that adverse events after chiropractic treatment
of spinal pain range from trivial to catastrophic. Given
these facts plus the limitations of previous studies, additional
research is required to develop a more accurate safety profile
of chiropractic treatment.
We assessed whether common adverse events differed
between participants who received usual chiropractic treatment
or a sham intervention for spinal pain. We also captured
information about the types, severity, onset, and duration of
MATERIALS AND METHODS
The complete protocol for this study has been published elsewhere
and it was registered with the Australian New Zealand
Clinical Trials Registry (ACTRN12611000542998). 
We conducted this parallel-group randomized controlled trial
during 3–month period between August 2012 and October
2012 in 12 Western Australia metropolitan chiropractic centers.
Participants were allocated to either
(1) a sham group including typical interaction with the practitioner or
(2) a usual care group providing individualized chiropractic treatment.
Participants and outcome assessors were blinded to
group allocation. All participants provided demographic and
clinical characteristics at entry to the trial, adverse events were
evaluated 2 days after each of the 2 treatments, and blinding
was assessed at 2–week follow-up.
Participants were recruited from newspaper advertisements.
All included participants were 18 years of age or older, English
literate, had nonspecific spinal pain (neck, midback, or low
back pain) of at least 1–week duration, and scored at least 3
on the Numerical Rating Scale (NRS) for pain  and 12 on the
Functional Rating Index. 
We excluded participants who thought that they would
be unable to tolerate any intervention potentially delivered
in usual chiropractic, including: manipulation, mobilization,
traction, soft tissue massage, and physical modalities.
We also excluded participants who had spinal pain related
to cancer or infection, spinal fracture, spondyloarthropathy,
known osteoporosis, progressive upper or lower limb weakness,
symptoms of cauda equina syndrome or other significant neurological condition, disc herniation, cardiovascular
disease, uncontrolled hypertension, cognitive impairment,
blood coagulation disorder, previous spinal surgery, previous
history of stroke or transient ischémie attacks, pacemaker or
other electrical device implanted, and a current compensation
Each participant was assigned to either a sham group or a
usual chiropractic group, whereupon 2 treatments were delivered
with approximately 1 week between treatments. The
chiropractors delivering either the sham or usual chiropractic
treatment attended a training session that provided instruction
about the trial and how to undertake their respective
treatments. To be eligible, all chiropractors need to practice
within the Western Australian metropolitan region and were
required to be registered with the Chiropractic Board of
the practitioners in this group comprised 4
registered chiropractors who administered at each visit (a) detuned
ultrasound ; (b) an Activator instrument,  a handheld
device that delivers a low impulse wound to lowest output
and administered on the back randomly through a tongue
depressor to disperse any remaining force; and (c) a randomly
placed hand on the spine while ultrasound was administered
to give a "hands-on" experience.
the practitioners in this group consisted of 8
registered chiropractors who provided individualized chiropractic
care consistent with their usual treatment approach.
The only condition that may have influenced the chiropractors'
usual treatment approach was a request to adhere to
Australian imaging guidelines. 
A statistician used a random number generator to create a
permuted block randomization list with variable block sizes
of 8 to 12. The group assignment was placed in sequentially
numbered, opaque, sealed, envelopes. Staff not administering
baseline or outcomes measures opened the envelope and allocated
participants to the groups.
Informed Consent and Blinding
Murdoch University's Human Research Committee granted
ethics approval for the study (2011/109). All participants provided
written informed consent. Research staff administering
baseline measures and outcome measures were also blinded
to group allocation.
Our primary outcome was adverse event occurrence. We
inquired about occurrence of adverse events by using a questionnaire
that was informed by previous research. [4–8] The
occurrence of an adverse event was assessed by an item stating
"Did you experience any new unwelcome symptoms or
an increase of your presenting symptoms during the first
48 hours (2 d) after treatment?"(yes/no). Further details
about any adverse events were obtained by 4 open and 4
close-ended questions about increased pain, muscular stiffness,
headache, and radiating discomfort. Each question
inquired about the intensity (11–point NRS), onset (5 categorical
responses ranging from <10 min to >24 hr), and duration
(5 categorical responses ranging from <1 hr to >2 d).
Participants completed the adverse event questionnaire 2 days
after each appointment and returned it by postal mail. The
period between the appointment and completion of the questionnaire
was selected to allow sufficient time for the manifestation
of adverse events.
Our a priori sample size estimation was based on detecting a
20% difference in the occurrence of adverse events between
the 2 treatment groups. Assuming an a level of 0.05 and
a 2–tailed hypothesis, recruiting 180 participants (90 per
group) would provide 80% power to detect a difference of
at least this magnitude. We thought a 20% difference to
be a conservative estimate, because previous studies have
shown that about half of those who receive chiropractic
treatment experience adverse events, [4–8] whereas in a previous
study, the de-tuiied ultrasound adverse event rate was
less than 10%. 
All data were reported descriptively. We classified adverse
event intensity as NRS score of 1 to 3 = mild; NRS score
of 4 to 6 = moderate; and NRS score of 7 to 10 = severe.
Missing data were handled with multiple imputation.  We
undertook intention to treat and available case analyses. We
calculated relative risk (RR) statistics with 95% confidence
intervals (CIs) for the following outcomes: adverse event
occurrence, severe adverse event occurrence (severe or not
severe), adverse event duration (>24 hr or <24 hr), and time
to onset of adverse event (>24 hr or <24 hr). When more
than 1 adverse event was reported, the most intense adverse
event was identified by the highest NRS score and used for all
analyses. Blinding was evaluated by using the Bang Index. 
We screened 272 potential participants, of whom 198 satisfied
selection criteria. Between August to September 2012, 183
participants were randomized to either a sham intervention
group (n = 91) or a usual care group (n = 92). Participant
flow through the study is displayed in Figure 1.
At baseline, there were no important differences in demographic
details or clinical characteristics, apart from the higher
percentage of females in the usual care group (42% compared
with 31%). Participant baseline demographic details and clinical
characteristics are displayed in Table 1.
The vast majority of patients (98%) had experienced spinal
pain for more than 3 months. Three quarters had experienced
spinal pain for more than 5 years (75% in sham group;
73% in usual care). The overwhelming majority indicated
that it had been more than 1 year since last experiencing a
4–week pain-free period (89% in sham group; 98% in usual
care group), and more than two-thirds indicated that it had
been more than 1 year since their last 1–week pain-free period
(71% in sham group; 68% in usual care group).
In total, 8 chiropractors (5 males) delivered usual care and
4 chiropractors (2 males) provided the sham intervention.
Chiropractors had on average 12.6 (SD, 2.3) years' clinical
experience, whereas chiropractors in the sham group had
on average 3.6 (SD, 1.1) years' clinical experience. About 3
quarters of the chiropractors had obtained their qualifications
from Australian universities (8/11). All chiropractors were
registered and practiced full time.
Type of Therapies Used in Usual Care Group
Details about therapies used are displayed in Table 2.
Flow of the patients
through each stage of the study.
Demographics and Clinical Characteristics*
Types of Therapies Used
Across Both Appointments
Adverse Events: Types, Severity, Onset, and Duration
In total, 33% of the sham group and 42% of the usual care
group reported at least 1 adverse event after either appointment.
The types, onset, and duration of adverse events are displayed
in Tables 3 to 5. Most participants who experienced
an adverse event reported more than 1 event (71% in sham
group; 77% in usual care group). In total, 198 adverse
events were reported (92 in sham group; 106 in usual care
group). Common adverse events were increased pain (29%
in sham group; 36% in usual care group), muscle stiffness
(29% in sham group; 37% in usual care group), headache
(17% in sham group; 9% in usual care group), and
radiating discomfort (15% in sham group; 15% in usual
care group). Less common adverse events, each of which
accounted for less than 5% of adverse events in the respective
groups, included dizziness, muscle spasm, fatigue,
sleeplessness, and joint swelling. The RR for experiencing
an adverse event was not significant (RR = 1.24; 95% CI:
Across both appointments, adverse event intensity was
most comrrionly moderate in the sham group (50%; n =
46/92) and either moderate (37%; n = 39/106) or severe
(37%; n = 39/106) in the usual care group. The rate of severe
adverse events was not different between the groups (RR =
1.9; 95% CI: 0.98–3.99).
Across both appointments, 79% of the adverse events
reported in the sham group (73/92) and 84% in the usual care
group (89/106) occurred within 24 hours. The RR for adverse
event onset was not significant (RR = 0.16; 95% CI: 0.02–1.34).
For duration and across both appointments, 51% of the
adverse events in the sham group (47/92) and 41% in the
usual care group persisted for less than 24 hours (44/106).
The RR for the duration of adverse events was not significant
(RR = 1.13; 95% CI: 0.59–2.18).
The intention to treat and available case analyses provided
consistent results with 1 exception. Regarding the occurrence
of serious adverse events, the RR estimates were 2.02 (95%
CI: 1.01–4.07) in the available case analysis and 1.97 (95%
CI: 0.98–4.0) in the intention-to-treat analysis.
The proportion of participants who identified the assigned
treatment was 67% for the sham group and 85% for the
usual care group. Bang Index values showed that 25% of the
sham group (95% CI: 10%^0%) and 61% of the usual care
group (95% CI: 48%–74%) guessed correctly beyond what
would be expected by chance. 
This was the first study to use a sham-controlled design
to examine adverse events after chiropractic treatment. A
substantial proportion of adverse events experienced during
chiropractic care for spinal pain may be the result of
natural symptom ñuctuation or from nonspecific effects.
Adverse events were common in both the usual chiropractic
care and sham groups, but no important differences were
seen between the groups and no serious adverse events were
reported. However, although very similar, the estimates of
severe adverse event risk arising from the intention to treat
and available case analyses resulted in conflicting conclusions.
Although the intention-to-treat approach was our primary
analysis, we cannot rule out the possibility of increased risk
of severe adverse event occurrence with chiropractic treatment
compared with sham therapy. The adverse event rate
reported by the usual care group in this study was consistent
with rates reported by previous studies (42% compared
with 34%–61 %). [4–8] Moreover, the finding that most adverse
events were benign and transitory is also consistent with other
chiropractic studies. [4–8] The proportion of adverse events in
these previous studies due to other effects such as natural
history or nonspecific effects remains indeterminable because
none of the studies used a sham arm. However, the results of
our study suggest that many adverse events experienced after
chiropractic treatment result from either natural history variation
or nonspecific effects. Some may view these results as
evidence that chiropractic treatment is essentially an entirely
benign intervention, but it more likely reflects that our study
was underpowered to detect a statistically significant difference
Studies of interventions other than manual therapies have
also associated nonspecific effects with adverse events. [24–26]
Interestingly, some studies demonstrated that the adverse
events reported by participants in either the placebo or the
sham arm mirror the adverse events in the active intervention
arm. [27–31] This association has been attributed in part to
the effect of patient expectancy, which typically depends on
details about possible adverse events conveyed through information
sheets, consent forms, or the investigators' behavior. [24–31]
Other studies have shown that a strong aversion to
experience an adverse event, coupled with a sense of helplessness
about avoiding it, may evoke negative emotions and subsequent
reductions in beneficial nonspecific effects.  It then
seems likely that an expectation of adverse events coupled
with not wanting to experience adverse events may promote
nonspecific effects that contribute to adverse events. 
Careful consideration should be given to how the information
from this study is presented to patients. As numerous
studies have shown, disclosing information about the
risks of adverse events increases the likelihood of adverse
events occurring. [24–31] Conversely, framing information about
adverse events in positive terms (noting that most patients
did not experience an adverse event), rather than negative
terms (detailing the minority who experienced an adverse
event), can lead to a lower adverse event rate. [33, 34] In light of
this, we recommend that a form of words be developed for
chiropractic patients that accurately reflects the results of our
study about potential common adverse events without unnecessarily
Conducting this study under typical clinical conditions
enhances the external validity of our findings. However, this
study was powered to detect a 20% difference in adverse
events; therefore, we were underpowered to detect the magnitude
of between-group differences observed in our sample of
participants. In terms of internal validity, we did not measure
either anxiety or depression, and it should be noted that it
is possible that differing levels of anxiety or depression may
have influenced the participants' experience of adverse events
in either group. However, there were no important differences
between groups in the psychosocial characteristics we
assessed at baseline including the Pain Catastrophizing Scale
and Fear-Avoidance Beliefs Questionnaire. Nevertheless,
it should be noted that it is possible that differing levels of
anxiety or depression may have influenced the participants'
experience of adverse events in either group.
We endeavored to blind study participants to group allocation
but were unsuccessful. This was probably due to inherent
difficulties in finding an adequate sham intervention to use in a
chiropractic trial or indeed any type of randomized controlled
trial.  The lack of blinding success may have influenced the
reporting of adverse events. In particular, the adverse event
rate may have been underreported in the sham group as the
majority thought they were receiving an inactive intervention.
Finally, the chiropractors providing usual care were more
experienced than chiropractors delivering the sham intervention.
However, experience is unlikely to affect an inert sham
intervention such as that delivered in this trial. It may be
the case that the more experienced chiropractors had better
interpersonal skills, which may have reduced the number of
adverse events attributable to nonspecific effects.
Additional studies of larger and more diverse populations
are warranted. Such studies should be powered to detect the
magnitude of between-group differences observed in this trial
and include a wait list arm to account for natural variation in
Adverse events resulting from chiropractic are
Most adverse events resulting from chiropractic
are benign and transitory.
A substantial proportion of adverse events resulting from chiropractic
seem to be due to nonspecific effects.
The authors gratefully acknowledge the assistance of Professor
Charlotte Leboeuf-Yde in reviewing the draft article, Christine
Losco for assistance with recruitment and data collection, and
the participating chiropractors.
Walker BF, Müller R, Grant WD.
Low back pain in Australian adults. Health provider utilization and care seeking.
J Manipulative Physiol Ther 2004;27:327-35.
Lambeek LC, van Tulder MW, Swinkels IC, et al.
The trend in total cost of back pain in The Netherlands in the period 2002 to 2007.
Hurwitz EL, Morgenstern H, Vassilaki M, et al.
Adverse reactions to chiropractic treatment and their effects on satisfaction and clinical outcomes among patients enrolled in the UCLA Neck Pain Study.
J Manipulative Physiol Ther 2004;27:16-25.
Barrett AJ, Breen AC.
Adverse effects of spinal manipulation.
J Royal Soc Med 2000;93:258-9.
Cagnie B, Vinck E, Beernaert A, et al.
How common are side effects of spinal manipulation and can these side effects be predicted?
Man Ther 2004;9:151-6.
Leboeuf-Yde C, Hennius B, Rudberg E, et al.
Side effects of chiropractic treatment: a prospective study.
J Manipulative Physiol Ther 1997;20:511-5.
Rubinstein S, Leboeuf-Yde C, Knol D, et al.
Predictors of adverse events following chiropractic care for patients with neck pain.
J Manipulative Physiol Ther 2008;31:94-103.
Senstad O, Leboeuf-Yde C, Borchgrevink C.
Frequency and characteristics of side effects of spinal manipulative therapy.
A report on the occurrence of cerebral vascular accidents in chiropractic practice.
J Can Chiropr Assoc 1993;37:104-6.
Klougart N, Leboeuf-Yde C, Rasmussen LR.
Safety in chiropractic practice. Part II: treatment to the upper neck and the rate of cerebrovascular incidents.
J Manipulative Physiol Ther 1996;19:563-9.
Rezai M, Cote P, Cassidy JD, et al.
The association between prevalent neck pain and health-related quality of life: a cross-sectional analysis.
Eur Spine J 2009;18:371-81.
Coulter ID, Hurwitz EL, Adams H.
The appropriateness of manipulation and mobilization of the cervical spine.
Hebert JJ, Stomski N, French S, et al.
Serious adverse events and spinal manipulative therapy of the low back region: a systematic review of cases.
J Manipulative Physiol Ther 2013
Hurwitz EL, Morgenstern H, Vassilaki M, et al.
Frequency and clinical predictors of adverse reactions to chiropractic care in the UCLA neck pain study.
Spine (Phila Pa 1976) 2005;30: 1477-84.
Recall bias can be a threat to retrospective and prospective
Intern J Epidemiol 2006. doi:10.5580/2732.
Walker B, Losco B, Clarke B, et al.
Outcomes of usual chiropractic, harm &c efficacy, the ouch study: study protocol for a randomized controlled trial.
Katz J, Melzack R.
Measurement of pain.
Surg Clin North Am 2010;79:231-52.
Feise RJ, Michael Menke J.
Functional rating index: a new valid and reliable instrument to measure the magnitude of clinical change in spinal conditions.
Spine (Phila Pa 1976) 2001;26:78-86; discussion 87.
Costa LO, Maher CG, Latimer J, et al.
Motor control exercise for chronic low back pain: a randomized placebo-controlled trial.
Phys Ther 2009;89:1275-86.
http://www.activator.com. Accessed July 6, 2011.
Australian Acute Musculoskeletal Pain Guidelines Group.
Evidence-Based Management of Acute Musculoskeletal Pain.
Canberra: National Library of Australia; 2003.
Multiple Imputation for Nonresponse in Surveys.
New York, NY: John Wiley & Sons; 1997.
Bang N, Ni L, Davis C.
Assessment of blinding in clinical trials.
Control Clin Trials 2004;25:143-56.
Amanzio M, Corazzini LL, Vase L, et al.
A systematic review of adverse events in placebo groups of anti-migraine clinical trials.
Barsky AJ, Saintfort R, Rogers MP, et al.
Nonspecific medication side effects and the nocebo phenomenon.
JAMA 2002;287: 622-7.
Colloca L, Miller FG.
The nocebo effect and its relevance for clinical practice.
Psychosom Med 2011;73:598-603.
Myers MG, Cairns JA, Singer J.
The consent form as a possible cause of side effects.
Clin Pharmacol Ther 1987;42:250-3.
Lancman ME, Asconape JJ, Craven WJ, et al.
Predictive value of induction of psychogenic seizures by suggestion.
Ann Neurol 1994;3S:359-61.
Flaten MA, Simonsen T, Olsen H.
Drug-related information generates placebo and nocebo responses that modify the drug response.
Psychosom Med 1999;61:250-5.
Luparello TJ, Leist N, Lourie CH, et al.
The interaction of psychologic stimuli and pharmacologie agents on airway reactivity in asthmatic subjects.
Psychosom Med 1970;32:509-13.
Mondaini N, Gontero P, Giubilei G, et al.
Finasteride 5 mg and sexual side effects: how many of these are related to a nocebo phenomenon?
J Sex Med 2007;4:1708-12.
Vase L, Robinson ME, Verne GN, et al.
Increased placebo analgesia over time in irritable bowel syndrome (IBS) patients is associated with desire and expectation but not endogenous opioid mechanisms.
Miller FG, Colloca L.
The placebo phenomenon and medical ethics: rethinking the relationship between informed consent and risk-benefit assessment.
Theor Med Bioeth 2011; 32: 219-43.
Edwards A, Elwyn G, Covey J, et al.
Presenting risk information—a review of the effects of "framing" and other manipulations on patient outcomes.
J Health Commun 2001;6:61-82.
Boutron I, Estellat C, Guittet L, et al.
Methods of blinding in reports of randomized controlled trials assessing pharmacologie treatments: a systematic review.
PLoS Med 2006;3:e425.
Return to the ADVERSE EVENTS Page
Return to the CHIROPRACTIC AND CHRONIC NECK PAIN Page