Two bodies of evidence count what happens to people after acupuncture. A 2021 systematic review by Petra Bäumler and colleagues at LMU Munich pooled 21 prospective studies covering more than 845,000 patients and 7.4 million treatments. A 2004 survey by Hugh MacPherson and colleagues followed 6,348 patients of British Acupuncture Council members by post for three months and asked them what they had noticed. Prospective means the studies recorded events as treatment went on, rather than collecting complaints afterwards. This page reports what they found, from the reactions that are common and brief to the few that need a doctor.

How often anything happens

The review found at least one adverse event, meaning any unwanted effect, in about 9 in 100 patients over a course of treatment, and 9.4 events for every 100 treatments. Half of those events were bleeding, pain or redness at the spot where a needle went in. In the patient survey, 682 of the 6,348 patients, about 11 in 100, reported at least one event over three months. The events they reported most were severe tiredness and exhaustion, pain where a needle went in, and headache.

Common: a few patients in every hundred

The review sorted events into groups and estimated how often each occurs. Six groups each affected between 1 and 5 in 100 patients over a course:

  • bleeding where a needle came out
  • pain at the needle site
  • other reactions at the needle site
  • reactions of the autonomic nervous system, the part of the nervous system that runs without conscious control; in the two studies that recorded the most of them, these were mainly reports of unusual tiredness after treatment
  • the treated symptoms getting worse
  • events the authors grouped under the central nervous system

A symptom that gets worse after treatment is sometimes called a healing crisis. The review notes that name and still counts the worsening as an adverse event, because international consensus treats an increase in the burden of a disease as one.

Uncommon: a few patients in every thousand

Nine more groups each affected between 1 and 7 in 1,000 patients over a course: symptoms of the peripheral nerves, the nerves outside the brain and spinal cord; pain away from the needle site; digestive or gynaecological symptoms; headache; symptoms of the heart and circulation; effects on movement; skin reactions across the body; emotional reactions; and sleeping problems. Events affecting breathing were rarer again, at about 4 in 10,000 patients.

Rare and serious

The review put serious adverse events at about 1 in 10,000 patients. In the studies it pooled, these included a collapsed lung, fainting, circulatory collapse and admission to hospital. The three reported most often were a collapsed lung, strong reactions of the heart and circulation including fainting, and falls or injuries, each at 1 to 3 cases per million treatments. One study the review cites recorded 2 collapsed lungs among more than 200,000 patients who had an average of 10 treatments each.

A collapsed lung, or pneumothorax, happens when air fills the space between the lung and the chest wall; after acupuncture, a needle that reaches through the chest wall can let the air in. MedlinePlus, the US National Library of Medicine’s health encyclopedia, lists its common symptoms as sharp chest or shoulder pain made worse by a deep breath or a cough, and shortness of breath. A larger one can add chest tightness, light-headedness and near fainting, a fast heart rate and bluish skin. MedlinePlus advises contacting a health care provider for any of them. AcuiQ’s point pages print a pneumothorax caution wherever a source gave one, and Reading a point warning explains where those cautions come from.

Infection at a needle site did not occur in the prospective studies the review pooled, although case reports describe it. MedlinePlus lists the signs of cellulitis, a bacterial skin infection, as redness or swelling that spreads, warm skin, pain or tenderness, and fever with chills, and advises contacting a provider right away.

What these numbers cannot tell you

The review’s authors note that the practitioners in most of the pooled studies were well trained, and give that as the likely reason several serious events known from case reports did not appear. Their rates describe treatment of that standard. The review also covers needling only: it left out acupressure, laser, point injection, moxibustion on its own, and studies limited to press needles or ear points. AcuiQ’s own entries record what a study prescribed, not what happened to the people in it, so no protocol on this site carries a safety record of its own; What AcuiQ does not do sets out that limit.