Table 5 Estimates for the incidence of serious adverse events following spinal manipulative therapy

2014

Todd A. J. et al. [98]

From a SR: 1 SAE in 250 million pediatric visits.

From a discussion paper: 0 SAEs reported in >30,000 treatments by medical manipulators.

2010

Carnes D. et al. [88]

From a pCohort: 14 cases of ‘unbearably severe side effects’ in 4712 treatments (0.13%). Upper risk rate for ‘serious adverse events’ of approximately 0.01% (3/28,109 consultations).

Their estimation from all pCohorts: Upper 95% CI incidence risk rate of major adverse events of 0.007% (0/42,451) after treatment or 0.01% (0/22,833) per patient.

From RCTs: No ‘major adverse events’ in the 31 RCTs (which included 2281 participants who received manual therapy and 2779 who received other therapies). Upper incidence rate of major adverse events of ~0.13% (0/2301) after manual therapy treatment.

2009

Gouveia L. O. et al. [99]

Their own synthesis (based on surveys): Between 5 strokes in 100,000 manipulations to 1.46 SAEs in 10,000,000 manipulations and 2.68 deaths in 10,000,000 manipulations.

2008

Miley M. L. et al. [127]

From a CC (which they consider the best available estimate): Approximately 1.3 cases of VAD or occlusion attributable to CMT would be observed within 1 week of manipulative therapy for every 100,000 persons <45 years of age receiving CMT.

From reviews: Published estimates of the incidence of VAD and stroke after range from 1 in 5.8 million to 1 in 5000.

2007

Chou R. & Huffman L. H. [141, 143]

From SRs: <1 SAE per 1 million patient visits.

2006

Snelling N. J. [135]

From a SR: 1 additional disc herniation or CES in 3.7 million manipulations (in pts, with lumbar disc herniation).

2004

Oliphant D. [115]

From their own estimation: <1 worsening LDH or CES in 3,72 million manipulations (in pts. with lumbar disc herniation), 1 worsening lumbar disc herniation or CES in 1,78 million manipulations (including manipulations under anesthesia; in pts. with lumbar disc herniation).

From other reviews: 1 CES in 128 million manipulations (given the quality score 84%), 1 CES in 100 million manipulations (given the quality score 86%), <1 (CES or herniation) in 1 million manipulations (given the quality score 74%), 1 LDH or CES in 2,789,709 manipulations (1 LDH in 8,369,129 manipulations, and 1 CES in 4,184,564 manipulations) (given the quality score 32%).

From a retrospective study: ‘They stated they were 95% confident that the risk of complication of manipulation for patients with back pain and sciatica was between 0% and 5%.’

From a prospective study: ‘A prospective evaluation of 2000 patients attending a chiropractic college clinic failed to reveal even one major complication’, ‘1000 new patients and 4700 treatments and found no permanent complications’.

From surveys: 1 minor or transient complication but no serious or permanent complications in 38,137 lumbar spinal manipulations.

From pooling the prospective and retrospective studies together: 0 major, serious, or permanent complications in >2100 patients (>13,100 treatments). 0 complications in 117 patients diagnosed as having LDH (>2000 spinal manipulation of probable disc herniations).

2002

Gross A. R. et al. [142]

From SRs: 1 serious complication in 20,000 to 5 serious complications in 10,000,000 cervical spine manipulations (rated as low accuracy and level V evidence), 1 stroke from cervical manipulation in 100,000 (0.001%).

From a survey: 1 CVA in 228,050 manipulations, 1 CVA in 1.3 million, 5 CVA in one million.

2002

Stevinson C. & Ernst E. [124]

Their own summarisation: ‘Estimates of the incidence of serious complications range from 1 per 2 million manipulations to 1 per 400,000’.

From reviews and a letter: 1 SAE per 1–2 million treatments.

From surveys: 1 sleight neurologic complication per 40,000 manipulations, 1 severe complication per 400,000 manipulations, 1 stroke per 1,300,000 treatments of cervical SMT.

From insurance claim data referred to in a SR: 1 stroke per 2 million manipulations.

From a CC: 1.3 VBA within 1 week of treatment in 100,000 pts <45 years receiving chiropractic treatment.

1996

Assendelft W. J. J. et al. [131]

Their own conclusion (partly based on the articles not appearing in their result section): From 1 VBA in 20,000 patients to 1 VBA in 1 million cervical manipulations. <1 CES in 1 million treatments.

From a SR: No complications in 1500 patients treated with manipulation in clinical trials.

From surveys: 1 slight neurological complication in 40,000 cases, 1 important complication in 400,000 manipulative procedures, 1 VBA in 228,050 manipulations, <5 strokes in 100.000 patients receiving neck manipulations.

1996

Hurwitz E. L. et al. [150]

From their own estimation: 5–10 VBA or other complications (spinal cord compression, vertebral fracture, tracheal rupture, diaphragm paralysis, internal carotid hematoma, cardiac arrest) in 10,000,000 manipulations, 3–6 major impairment (paralysis, neurologic deficit, other permanent functional impairment) in 10,000,000 manipulations, <3 deaths in 10,000,000 manipulations.

From surveys: 1 serious complication in 400,000 to >1 million manipulations, 1 CVA accident in 3.85 million cervical spine manipulations.

They compare the incidence rates with NSAID consumption (0.39–3.2 serious gastrointestinal event in 1000 subjects) and cervical spine surgery (15.6 neurologic complications (spinal cord or nerve root injury, recurrent laryngeal nerve palsy, dural leak, and injury to cervical sympathetic nerve trunk (Horner's syndrome)) in 1000 surgeries and 6.9 deaths in 1000 surgeries.

1995

Dabbs V. & Lauretti W. J. [114]

Their own summarisation: 0.5–2 strokes in one million cervical manipulations performed, 1 serious vascular complication in 100.000 patients who undergo a course of treatment (10–15 sessions of cervical manipulation over the course of a year) with cervical manipulation, or 0.001%, 1 death in 400.000 pts. treated, or an ‘overall death rate of 0.0025% per course of treatment for patients with neck pain who are treated with cervical manipulation.’

They compare this with a risk of 0.4% for getting serious gastrointestinal ulcers requiring hospitalization because of NSAID use, and a risk of 0.04% for death from gastrointestinal bleeding caused by NSAID use.

Their own calculation based on insurance company data: <1 stroke in 2 million cervical manipulations.

From surveys: 1 serious complication in 400.000 cervical manipulations (no reported deaths), 1 complication in 518.000 manipulations, 1 stroke in 500.000 cervical manipulations, no serious incidence in >500.000 manipulations, 2–3 ‘more-or-less serious incidents’ in one million treatments.

From reports: no vertebral artery injury or stroke in 5 million cervical manipulations, no significant complications in 168.000 cervical manipulations.

From a review: 1–2 strokes in one million manipulations.

1992

Shekelle P. G. et al.[92]

Their own estimation: <1 case of CES in 100 million lumbar spinal manipulations.