Gua sha scrapes the skin with a smooth edge until small bleeds appear under it. Arya Nielsen and colleagues, in a 2007 study of skin blood flow, define it as therapeutic surface frictioning that intentionally raises transitory petechiae (pinpoint bleeds) and ecchymosis (bruising), and list its other names: cao gio in Vietnamese, coining, scraping and spooning. The red marks are called sha. A 2010 systematic review by Lee Myeong Soo and colleagues at the Korea Institute of Oriental Medicine describes it in Chinese medicine as a way to relieve blood stagnation at the body surface, used for pain and for colds, flu and other respiratory problems.
How it is done
The Lee review lists the tools: a Chinese soup spoon, an edge-worn coin, a slice of water-buffalo horn, a cow rib, honed jade, or a metal cap with a smooth round lip, used over skin wetted with oil or water. A 2023 commentary by Nielsen, Marsha Handel, Jennifer Stone and Lee defines the stroke as unidirectional press-stroking of a lubricated area, and says that no flame or alcohol is involved and that oil, if warmed at all, is warmed in a container of water. A 2014 case report by Tsai Keng-Kuang and Wang Chih-Hung describes repeated stroking with pressure, using a spoon or coin. The neck and back trials in the table below gave one or two treatments; a 2017 Chinese trial in perimenopausal women by Meng and colleagues gave 15-minute sessions once a week for 8 weeks.
The marks
Nielsen and colleagues’ 2023 commentary describes sha as blood that has leaked out of vessels into the tissue under the skin, fading in 3 to 4 days; a 2012 pilot trial by Lauche and colleagues gives 2 to 5 days. The same commentary lists what Western clinicians have called the marks in print, among them injury, abuse, pseudo-battery, burns, purpura and hematoma, and treats each as a mischaracterisation of an intended effect.
Nielsen and colleagues wrote in 2007 that no study had yet measured the physiological effects of gua sha. At the University Hospital of Essen in Germany they scraped an area of the back in 11 healthy volunteers from the nursing and medical staff and measured surface blood flow by laser Doppler imaging. Flow at the treated area rose fourfold for the first 7.5 minutes and stayed above baseline for the whole 25 minutes measured. Every volunteer reported less muscle pain at once, at the treated site and at a related distant site, and the authors found that the distant relief was not explained by more blood flow there; they concluded that the pain-relieving mechanism was unidentified.
A second line of work concerns haem oxygenase-1 (HO-1), an enzyme that protects cells against oxidative stress. Kwong and colleagues, in a 2009 imaging protocol in genetically modified mice, recorded its upregulation after scraping and proposed that it responds to the haemoglobin products released into the tissue. In people, a 2011 case study by Chan and colleagues reported raised plasma HO-1 and lower liver enzymes 48 hours after scraping in one carrier of active hepatitis B; their 2013 follow-up in 4 inactive carriers and 9 people without hepatitis B found no change in HO-1, liver enzymes or cytokines. Yuen and colleagues’ pilot trial, in the table below, recorded falling HO-1 and TNF-α levels in saliva after gua sha. The 2023 commentary cites the mouse work for its statement that sha upregulates HO-1.
What reviews and trials found
| Source | What it covered | What it found |
|---|---|---|
| Lee and colleagues, Chinese Medicine, 2010 | Musculoskeletal pain: 5 randomised and 2 non-randomised controlled trials, against acupuncture, no treatment, herbal injection, massage or electric current | Two randomised trials favoured gua sha, but both were of poor quality; the authors judged the evidence insufficient to show an effect on pain |
| Braun and colleagues, Pain Medicine, 2011 | Chronic neck pain: 48 outpatients, one gua sha treatment against a local thermal heat pad, followed for 7 days; open (unblinded) design | Less pain at 1 week with gua sha (difference −29.9 mm on a 100 mm scale, 95% CI −43.3 to −16.6), with better neck disability and quality-of-life scores |
| Lauche and colleagues, American Journal of Chinese Medicine, 2012 | Chronic neck or low back pain: 40 people, one treatment against a waiting list, 39 analysed at 7 days | Less pain than the waiting list in both neck and back patients; pressure pain sensitivity improved in the neck but not the low back; no adverse events reported |
| Yuen and colleagues, Complementary Therapies in Medicine, 2017 | Chronic low back pain in older people: crossover pilot against a hot pack | No significant difference between the two on any outcome over 1 week |
| Saha and colleagues, Complementary Therapies in Clinical Practice, 2019 | Chronic low back pain: 50 people, two treatments against a waiting list | Lower pain intensity and better self-rated health after treatment; no difference in function or the other measures; no serious adverse events |
| Meng and colleagues, Menopause, 2017 | Perimenopausal syndrome: 80 women in Nanjing, gua sha added to conventional treatment against conventional treatment alone; 75 completed | Larger fall in symptom score (modified Kupperman Index, difference 4.86 points) and in quality-of-life score; no difference in oestrogen, FSH or LH |
| Ren and colleagues, Complementary Therapies in Clinical Practice, 2018 | Perimenopausal syndrome: 6 trials, most of low methodological quality; 5 added gua sha to conventional drugs | Lower Kupperman score with gua sha added (mean difference −4.57, 95% CI −5.37 to −3.77), with high heterogeneity (I² 86%) |
| Dandan and colleagues, Journal of Traditional Chinese Medicine, 2024 | Knee osteoarthritis: 12 trials, 997 people, against other treatment or with the same added treatment in both groups | Lower pain (mean difference −2.11 on a visual analogue scale, 95% CI −3.79 to −0.44, I² 99%) and better function scores; certainty low by GRADE |
Braun’s and Saha’s trials were both run at Kliniken Essen-Mitte in Germany, and Lauche’s shares authors with both. None of the five trials in the table used a sham scraping control; Yuen and colleagues give the lack of a valid sham as the reason for their design, which compared gua sha with a hot pack. Sham acupuncture sets out why the choice of control matters. The Chinese reviews mostly add scraping to drugs and compare with the drugs alone. Neck pain and Low back pain report the wider acupuncture evidence for each condition.
Adverse events and case reports
The marks have been read as abuse. Aprile and colleagues, in a 2015 forensic case report, describe a 43-year-old Chinese woman admitted to hospital for abdominal pain whose back carried an unusual pattern of reddish bruises, attributed to gua sha after investigation; they advise clinicians and forensic pathologists to consider press-stroking when an adult presents with scratch-like marks or bruises. The 2023 commentary by Nielsen and colleagues says immigrant families have been accused of child abuse, and some prosecuted, because Western doctors misread the marks. It also traces a burn case published in 2000 and still cited as a complication of coining: the photographs show round unburned patches about 3 inches across, which the authors identify as the places where cups had been set, and they conclude the burns came from negligent fire cupping. Cupping covers that technique.
The case reports that do attribute harm to scraping each describe one patient. Tsai and Wang describe a 23-year-old man who had gua sha on his neck for a sore throat that had lasted 2 days and came to an emergency department with difficulty speaking and a swollen epiglottis narrowing the airway; his voice recovered within 2 hours of intravenous steroids, and the authors attribute the swelling to a contusion of the soft tissue beneath, from inappropriate application. Park and colleagues, in a 2021 case report from Korea, describe a healthy 35-year-old woman who developed painful nodules and pustules on her arms and legs after gua sha for body contouring; cultures grew Mycobacterium massiliense, an environmental bacterium, and she recovered after drainage and the antibiotic clarithromycin. The authors call it the first such infection reported after gua sha.
On infection control, Nielsen, Kligler and Koll published safety protocols for gua sha and cupping in 2012. They note that blood and other potentially infectious material can sometimes be drawn through the skin, contaminating instruments and exposing the next person to bloodborne infection, that neither the literature nor the acupuncture profession’s national standards addressed this sufficiently, and they apply hospital standards to both techniques.
What the sources cannot tell you
The 2010 review found the evidence insufficient, and the trials published since are small, short and without a sham control, and Braun’s, Lauche’s and Yuen’s followed people for a week. The perimenopause and knee reviews pool mostly low-quality trials with high heterogeneity. The blood-flow finding comes from 11 volunteers, and the HO-1 findings in people from one case study, a 13-person follow-up and a pilot trial, which point in different directions. How often scraping causes harm is unknown: the trials that reported on safety found nothing serious, and the harms literature is single case reports, some of them, by Nielsen and colleagues’ account, misattributed. A protocol entry on AcuiQ names the modality the study used; Reading a protocol explains each field.