The systematic review is published in the July issue of the Journal of the Royal Society of Medicine.
'Even allowing for an extraordinarily high level of under-reporting, spinal manipulation has been associated with about 600 serious adverse events,' said lead researcher Edzard Ernst, Professor of Complementary Medicine at Peninsula Medical School, Universities of Exeter and Plymouth.
He added: 'In addition, it causes non-serious adverse effects in about 50 percent of all patients who use it. If any drug were linked to such rates of harm, I somehow doubt that it would still be on the market.'
The author reviewed case reports, retrospective case series, prospective case series, case-controlled studies and surveys and found evidence of adverse events which ranged from less serious problems - such as low level pain - to serious or even life-threatening events. These included disc herniation, bone fractures, spinal cord injury and stroke.
The author suggests that patients rarely receive enough information about the associated risks of spinal manipulation to allow them to give informed consent.
Comment below and tell us what you think.