Evidence Review · Complementary & Alternative Medicine

Fatalities After CAM: What the Peer-Reviewed Evidence Shows

A summary of Edzard Ernst's 2011 overview of deaths linked to acupuncture, EDTA-chelation therapy, and chiropractic — and why the true toll is almost certainly higher than what gets published.

Source: Ernst E. "Fatalities after CAM: an overview." Br J Gen Pract. 2011;61(587):404–405.

Most CAM practitioners and patients assume these treatments carry little to no risk. The peer-reviewed record says otherwise. Edzard Ernst — former professor of complementary medicine at the University of Exeter, and a contributor with this site's operator to Healing, Hype or Harm? — compiled the results of three of his own systematic reviews into a single 2011 overview published in the British Journal of General Practice. The numbers below come directly from that overview.

A systematic review identified 86 documented deaths following acupuncture. Many of the original case reports were incomplete, so causality could not always be established with confidence. The most common causes of death were pneumothorax (a punctured lung) and puncture of the heart, a major blood vessel, the central nervous system, or the liver, along with various largely bacterial infections. Most reports originated in China, Japan, and other Asian countries, with additional cases from Germany, Norway, and the United States. Ernst notes that pneumothorax is also the single most common serious non-fatal complication of acupuncture — and that every death in the review would likely have been avoidable with proper technique and anatomical knowledge.

Some CAM practitioners promote intravenous EDTA-chelation therapy for cardiovascular conditions such as claudication, despite it being biologically implausible and not demonstrably effective for that use (its legitimate, conventional use is for severe heavy-metal poisoning). A systematic review of deaths reported since 1990 found nine fatalities, with causes including myocardial infarction, cardiac arrest, and arrhythmia. Non-fatal adverse effects from IV chelation continue to be reported regularly. Because the benefits of "alternative" chelation for arteriosclerotic disease are unproven, Ernst concludes the risk-benefit balance does not support its use for that purpose.

A separate systematic review found 26 published case reports of death following chiropractic treatment, plus additional fatalities that were never formally published in the medical literature. The typical mechanism was dissection of a vertebral artery following high-velocity manipulation of the upper spine — the artery tears, a clot forms, and the patient suffers a stroke. A survey of UK neurologists found that vascular accidents following neck manipulation were essentially never reported through formal channels, suggesting the published case count is a significant undercount of the true toll.

Ernst's overview is explicit that acupuncture, chelation, and chiropractic are only examples. His review cites individually documented fatalities and serious adverse events connected to herbal medicine, oxygen-ozone therapy, massage, osteopathy, restrictive alternative diets, high-dose vitamin supplements, aromatherapy oils, homeopathic remedies, hypnotherapy, meditation, and yoga — each with its own published case report in the medical literature.

He also flags a category of harm that doesn't show up in a fatality count at all: indirect risk. This includes bad medical advice given online, aggressive sales tactics used to sell treatments, patients replacing an effective conventional treatment with an ineffective alternative one, and patients delaying effective treatment while they try CAM first. That last pattern — treatment delay while pursuing alternative approaches for a serious illness — is exactly what happened in this site's own Brian Clement/Hippocrates Health Institute investigation.

Ernst's closing point is a risk-benefit one, not simply a risk one: proponents often argue these deaths are rare relative to the volume of treatments given, and safer than many conventional interventions. That may be true — but risk has to be weighed against benefit, not viewed alone. When a treatment's benefit is small, uncertain, or absent, even a small risk tips the balance against it.