Fever

Fever is a body temperature above the normal range, and in the studies on this page it has two different origins. Most of the recent trials concern fever in labour: women who have an epidural for pain relief run a higher temperature than those who do not, a condition called epidural-related maternal fever. The background of a 2026 trial by Tang and colleagues puts its incidence at 15 to 30% of women given labour analgesia. The older studies concern fever from infections such as the common cold, in adults and in children. The two bodies of research share almost no methods and are reported separately below.

Does acupuncture help fever?

No Cochrane review covers acupuncture, acupressure or bloodletting for fever, and the reviews that do exist answer narrower questions. A 2025 network meta-analysis of 34 trials ranked electroacupuncture and magnetic beads taped to the ear among the three most effective ways of preventing fever after an epidural in labour, without grading certainty in its abstract. The four acupoint trials in labour found for this page compared the method with no added treatment or, in one, with electrodes that delivered no current. For fever from infection, a 2025 review of massage in children rated its pooled evidence moderate in certainty, and the needling trials compared acupuncture with a drug.

A search of PubMed and Europe PMC in November 2026 found no review or randomised trial of bloodletting, the pricking of points such as the fingertips or the apex of the ear to release a few drops of blood, for fever in people; a PubMed search found no study of acupuncture for rheumatic fever. The nearest Cochrane review, of acupuncture for mumps in children, a febrile viral illness, found no trial that met its criteria in its 2015 update by He and colleagues.

Fever in labour after an epidural

The network meta-analysis by Yu and colleagues at West China Second University Hospital, Sichuan University, published in the American Journal of Obstetrics and Gynecology in 2025, searched four databases to May 2024 and included 34 randomised trials with 8,235 women. Thirty-three tested prevention and one treatment. A network meta-analysis compares treatments that were never tested against each other in the same trial by linking them through shared comparators. The authors assessed each trial with the Cochrane risk-of-bias tool and ranked the interventions by a statistic called the surface under the cumulative ranking curve. Electroacupuncture, magnetic bead ear therapy, changes to how the epidural was given and preventive steroids each lowered the odds of fever, with odds ratios between 0.29 and 0.64 that the abstract does not separate by intervention. Electroacupuncture, the ear beads and steroids ranked as the top three. The abstract does not say how many of the 34 trials tested an acupoint method, which ones, or how certain the evidence was. Its search ended before the Tang trial was published.

The Chen trial, by Chen and colleagues at the Tianhe District Maternal and Child Health Care Hospital in Guangzhou, published in Chinese in 2021, is the only one of the four with a sham control. The control group wore electrodes on the same hand and elbow points with no current. The stimulated group also used less analgesic and had fewer chills, and the groups did not differ in pain scores, length of labour, mode of birth or the babies' Apgar scores.

The three ear trials all used Shenmen, internal genitals, endocrine and sympathetic, written TF4, TF2, CO18 and AH6a in the Chinese ear-point notation the first Wen paper uses; Ear acupuncture describes the method. Their control groups had the epidural and nothing in its place, so neither women nor staff were blinded. The two Wen papers come from the same anaesthesiology department at the First Affiliated Hospital of Nanchang University, by the same four authors, in the same year. The Zhongguo Zhen Jiu paper randomised 160 women, 80 to each group, and lost 12 and 15. The Zhen Ci Yan Jiu paper gives 180 women, 90 to each group, in its English abstract and 79 and 81 in its Chinese one, and adds placental pathology, reporting less inflammation of the membranes with the beads. Neither abstract mentions the other, and they may report some of the same women. The Tang trial, at the First Affiliated Hospital of Guizhou University of Traditional Chinese Medicine, set postpartum sleep and fatigue as its primary outcomes. It counted fever as a tympanic temperature of 37.5 °C or more that lasted over 30 minutes, kept the outcome assessors blind to group, and lists the absence of blinding and of a sham procedure among its limitations. Labour pain reports the evidence on acupuncture for pain relief in labour itself.

Fever from colds and other infections

Two randomised trials from Tianjin compared needling at GV14 (Dazhui) with an injected antipyretic drug, a fever-lowering medicine the abstracts call antondine or antongding. Both are in Chinese and share an author, Zhao JG. In the larger, from 2007, Xiao and colleagues at the No. 464 Hospital of the People's Liberation Army gave 133 people electroacupuncture and 128 the injection. Temperatures were lower in the acupuncture group at every observation point over 24 hours, and the temperature began to fall after about 1.4 hours against 3.4 with the drug. The authors classed 27.8% of the acupuncture group as cured against 10.9%, and counted fever as abated in 75.9% against 55.5%. In the 2006 trial by Ling and Zhao, at Tianjin University of Traditional Chinese Medicine, 45 people had manual acupuncture at GV14 and 41 the injection, and the authors counted 9 and 2 cured after 24 hours. Neither trial had a sham arm.

A 2014 case series by Kwon and colleagues in a South Korean military medical service room treated eight soldiers with a fever of 38 °C or more from a viral upper respiratory infection, needling GV14, GB20 (Fengchi) and TE8 alongside points from a newer method called equilibrium acupuncture. Temperature fell in all eight after one 20-minute session, and the report had no comparison group.

A 2022 retrospective cohort study by Li and colleagues at Xiyuan Hospital, Beijing, drew on the records of an online clinic platform that promotes point application, the taping of herbal pastes to points. Among people with fever diagnosed as a common cold in late 2020, it matched 1,229 who had point application with 423 who did not, and the fever disappeared within 2 days 1.82 times as often in the application group. The pastes went most often on CV08 (Shenque) at the navel and on GV14, and the median age of the people the authors counted as responding was 2. The study was not randomised: propensity score matching balanced the groups on age, sex and other factors in the records. Point application describes the method.

Fever in children

The 2025 review by Liu and colleagues at Beijing University of Chinese Medicine, in BMC Pediatrics, searched to November 2023 and included 15 randomised trials with 1,661 children. It pooled two kinds of massage: tuina, the Chinese therapeutic massage that works along channels and points, and tepid water massage, sponging the body with lukewarm water. The controls were antipyretic drugs. Massage alone did not differ clearly from the drugs on temperature after 120 minutes, and massage added to drugs lowered temperature more than drugs alone. In three trials with 400 children, tuina alone left fewer children febrile than the drugs did (risk ratio 0.41, 95% confidence interval 0.26 to 0.65). The authors rated their two main pooled outcomes moderate in certainty, marked down for inconsistency between trials, and reported no adverse reactions in the tuina trials. None of the trials compared tuina with a sham or with no treatment. Tuina describes the method.

What guidelines say

NICE's guideline Fever in under 5s: assessment and initial management (NG143), published on 7 November 2019 and last updated on 26 November 2021, is current. Its recommendations do not mention acupuncture, massage or any complementary therapy. They advise against tepid sponging, the method in several of the trials in the Liu review, and against using antipyretic drugs with the sole aim of lowering a child's temperature; paracetamol or ibuprofen is for a child who appears distressed. The American Academy of Pediatrics' 2011 clinical report on fever and antipyretic use in children makes the same distinction: its abstract describes fever as a mechanism that helps fight infection and sets the child's comfort, not a normal temperature, as the goal of treatment. Its abstract does not mention acupuncture or other complementary therapies.

What the evidence could not settle

For fever in labour, only the Chen trial compared stimulation with a sham, and its abstract gives no fever rates; the three ear trials compared beads with no added treatment, and two of them may share participants. The network meta-analysis that ranked the methods does not say in its abstract how many acupoint trials it drew on or how certain its rankings are. For fever from infection, the GV14 trials measured how fast temperature fell against an injected drug, an outcome the NICE and American Academy of Pediatrics guidance do not treat as the aim, and their abstracts do not define the cured and effective rates they report. No study found tested bloodletting for fever in people.

AcuiQ's fever, high fever and epidural-related fever pages list the protocols individual studies prescribed, each with its citation.