How Acupuncture Is Explained: Traditional and Contemporary Views Side by Side

Four kinds of statement

People often mix very different kinds of claim when they say "how acupuncture works". This page keeps them apart:

1. A traditional explanation: how the tradition itself describes health, illness and treatment.
2. A proposed mechanism: an idea from modern researchers about what the needle might do in the body.
3. A physiological finding: something measured in animals or people, such as a chemical released or a brain-imaging change.
4. A clinical outcome: whether people's symptoms improved in a trial, and compared with what.

A finding in one layer does not prove the layer above or below it.

The traditional explanation

In East Asian medicine, acupuncture is embedded in naturalistic theories built on ideas such as yin and yang, qi, wind and dampness. A leading medical review describes these as conceptual frameworks that emphasize the reliability of ordinary human sensory awareness. It also notes that many physicians who practice acupuncture reject such prescientific notions. [1]

In this framework, qi (often translated as vital energy) is said to move along channels called meridians. The meridian system is described as consisting of 12 main channels, first set out in the Huangdi Neijing. [2] UNESCO's description of the practice says the theory holds that the body is connected by channels, and that stimulating them can promote self-regulating functions and bring health. [3]

Traditional practitioners also describe a needling sensation called deqi, literally "the arrival of vital energy". In theory it is a prerequisite for clinical effect. [4] Whether it predicts results is unsettled: a systematic review found few clinical studies, with conflicting results, and concluded the evidence was not solid enough to say. [5]

Respecting a tradition means describing it in its own terms. It does not mean treating qi or meridians as findings of anatomy. The research on whether points and meridians can be located is on do acupuncture points exist, and a fuller treatment of qi is on qi and acupuncture.

Authors differ on how much the classical writers knew of the body. One 2024 review argues the early authors had extensive anatomical and physiological knowledge, and says that offers clues to mechanisms. [6] A review of traditional Chinese medicine premises says anatomical study was relatively uncommon, and that many studies trying to find meridians anatomically do not meet current research-conduct guidelines. [2]

WHO's guidelines state that using acupuncture in modern medical care means taking it out of its traditional context and applying it as a technique, without having to reconcile traditional and modern theory. [7]

Proposed contemporary mechanisms

NCCIH says how acupuncture works is not fully understood, and describes three kinds of possible effect: on the nervous system, on local tissue, and nonspecific effects. [8] The rest of this section sorts the evidence for each.

A mechanism seen in mice does not show the same thing happens in people, or that it matters for any condition. The adenosine and vagal-adrenal papers are animal work, and the connective-tissue idea is a hypothesis.

What clinical trials add

A pooled analysis of individual data from 39 trials, covering 20,827 people with chronic musculoskeletal pain, osteoarthritis, headache or shoulder pain, found acupuncture better than sham by about 0.2 standard deviations and better than no acupuncture by about 0.5. The authors said pain relief could not be explained solely by placebo, and that factors beyond needling at the correct point locations contribute. [14]

Here is why that matters for the "how". The gap over no treatment is much larger than the gap over sham, and NCCIH says the pattern suggests nonspecific effects contribute. [8] A historical review also argues that some sham methods were chosen without checking whether they have physiological effects of their own, so sham comparisons may understate acupuncture's effects. [15] So neither "it is only placebo" nor "meridians have been proven" follows from the trials. See sham acupuncture and placebo controls.

What is established and what is not

• Promising but limited: a modest effect on some chronic pain beyond sham, with part of the benefit from expectation and context. [14][8]
• Preclinical or limited: specific chemical and nerve pathways. [10][13]
• Insufficient evidence: that deqi, or points chosen by traditional theory, improve results. [5]
• Not established: that qi or meridians exist as physical structures. [2]

For more, see how acupuncture may work: proposed mechanisms, brain imaging and neuroscience and the overview on what acupuncture is.