Chronic cough

The European Respiratory Society (ERS) 2020 guideline counts a cough as chronic once it has lasted eight weeks in an adult or four weeks in a child. A 2025 review lists its common causes as upper airway cough syndrome, from the nose and sinuses, cough-variant asthma, gastro-oesophageal reflux and respiratory infections, and notes that one person can have more than one. Cough-variant asthma was first described, the ERS guideline records, as asthma in which cough is the only symptom. The trials measure cough on symptom scores rated separately for day and night, and quality of life on the Leicester Cough Questionnaire (LCQ), 19 questions in which a higher score is better. Many also report a total effective rate, the share of patients each trial judged improved by its own criteria.

No Cochrane review

Cochrane has published no review of acupuncture, acupressure or acupoint application for cough. Its reviews of other treatments contain no acupuncture trial: the 2019 review by Slinger and colleagues of speech and language therapy for unexplained chronic cough found two trials with 162 adults, and the 2015 review by Molassiotis and colleagues of cough in cancer included 17 studies of brachytherapy, laser or photodynamic therapy and drugs, and drew no practice recommendations from them.

Later reviews

The first three reviews below compare acupuncture or a related method with drugs, or add it to drugs and compare the result with the same drugs alone. None of their trials used sham acupuncture, a control procedure made to look like the real one (Sham acupuncture covers the kinds in use), so none can separate the needling from the attention that comes with it. The fourth, of cough in lung cancer, takes in the sham-controlled trial of electrical stimulation described below. A mean difference (MD) is the gap between groups in the units of the scale; I² is the share of the variation between trials that chance does not explain.

The review by Lee and colleagues, from the Korea Institute of Oriental Medicine and Kyung Hee University, is the only one to take chronic cough whatever its cause. Twelve tested acupoint herbal patching, a herbal medicine patch stuck on the points; eight tested needle acupuncture, three moxibustion (heat from burning moxa held over the points), and the rest patching with acupressure, hot compresses, bloodletting or herbal injection at points. All the control groups had conventional treatment except one, which had none. The authors judged every trial at high risk of bias because neither patients nor practitioners were blinded, rated the certainty of the evidence generally moderate, and noted that no trial counted coughs. None of its trials was restricted to cough after an infection. The trials most often used BL13 (Feishu), in 25 of them, then GV14 (Dazhui), CV17 (Danzhong) and Dingchuan (XB01B, which the review writes EX-B1).

Choi’s review, from the same institute, rated its findings moderate certainty for the cough scores and the LCQ and low for the total effective rate, downgrading for unclear blinding and small trials. Xiong’s review, from Chengdu University of Traditional Chinese Medicine, found that none of its 11 trials described how anyone was blinded or how allocation was concealed, and called their quality low. A 2022 review of acupoint application for cough-variant asthma, 13 trials with 1,237 people, was retracted by the journal in March 2024 and is not used here. Asthma reports the trials in asthma generally, Gastro-oesophageal reflux those in reflux itself, and COPD those in chronic obstructive pulmonary disease.

Trials against sham

In a 2024 trial in Shandong, Wang and colleagues randomised 110 people having keyhole lung cancer surgery to transcutaneous electrical acupoint stimulation (TEAS), current passed through pads on the skin, or to the same pads with no current. Stimulation ran for 30 minutes before anaesthesia and daily for four days afterwards at BL13, BL20 (Pishu), BL23 (Shenshu), LI04 (Hegu), LU07 (Lieque) and KD03 (Taixi). The trial reported lower cough scores with TEAS from the first day to three months after surgery, and LCQ scores higher at one and three months.

In Harbin, Zhang and colleagues gave 100 people sedated for flexible bronchoscopy, a camera passed into the airways, 30 minutes of TEAS at LI04, PC06 (Neiguan) and LU07 beforehand, or the same electrodes with the stimulator off. Patients and the bronchoscopist did not know which they had. The bronchoscopist scored cough lower in the TEAS group, which also needed less lidocaine and fentanyl; patients’ own discomfort scores did not differ.

Added to usual care, without a sham

A 2024 trial at Southern Medical University in Guangzhou randomised 130 people with lung cancer to routine nursing or routine nursing with seven days of acupuncture at LU07, LU09 (Taiyuan), BL13 and BL20. The trial reported cough 0.32 points better on a symptom inventory with acupuncture, and no difference in sputum or breathlessness. In children, a Guangzhou trial published in 2026 added two weeks of acupoint application to bronchodilator and inhaled steroid treatment in 50 of 100 children with chronic cough, and reported lower daytime and nighttime cough scores than in the 50 who had the drugs alone, with no adverse reactions.

At a cancer hospital in Mumbai, Solanki and colleagues tested acupressure against the cough that an intravenous dose of fentanyl, an opioid painkiller, can set off within seconds. Among 225 women, those given finger pressure at KD27 (Shufu) on both sides, below the inner ends of the collarbones, coughed in 1.3% of cases, against 12.7% with undiluted fentanyl alone and 6.8% with diluted fentanyl; the severity of cough did not differ. There was no sham pressure.

What guidelines say

The ERS guideline for adults and children, published in January 2020, does not mention acupuncture. For refractory chronic cough in adults it recommends a trial of low-dose morphine and suggests a trial of gabapentin or pregabalin, and it suggests cough control therapy, physiotherapy and speech and language therapy, as a conditional recommendation on moderate-quality evidence. The American College of Chest Physicians (CHEST) guideline on unexplained chronic cough, published in January 2016, does not mention it either; it suggests gabapentin and multimodality speech pathology therapy.

The Chinese Thoracic Society’s English-language guideline, published in 2018 from a literature search that ended in June 2015, added a section on traditional Chinese medicine. It covers herbal prescriptions by syndrome, calls their evidence level low, and does not mention acupuncture. The society’s 2021 revision, published in Chinese in January 2022, was drafted with experts in traditional Chinese medicine among other fields, according to its abstract; its full text could not be checked for acupuncture. NICE’s NG120, published on 7 February 2019, covers antibiotics and self-care for acute cough in England and does not mention acupuncture.

What the reviews could not settle

The reviews could not say whether any effect comes from the points stimulated. Only Cao’s review took in a sham-controlled trial, and it pooled acupuncture with rehabilitation, exercise and other non-drug treatments; Lee’s found no trial that counted coughs, so its results rest on symptom scores and questionnaires collected without blinding. They could not say how large an effect is: the 0.85-point fall in Lee’s pooled cough score comes from three trials whose I² of 86% shows them disagreeing. Every trial in Lee’s and Choi’s reviews was run in China, against drugs that differed from one trial to the next, and the sham-controlled trials above treated cough around surgery and bronchoscopy rather than a cough of months.

AcuiQ’s chronic cough, cough-variant asthma, postoperative cough and cough pages list the protocols individual studies prescribed, each with its citation.