Xu 2012.

Methods	Parallel‐group randomized trial, conducted in China. Study dates not reported.
Participants	130 adults, ASA physical status I ‐ III, undergoing infratentorial craniotomy. Excluded: those with previous experiences with acupuncture, nausea or vomiting within 24 h before surgery, preoperative use of antiemetics (except dexamethasone), cardiac pacemaker, cardioverter, or defibrillator, pregnant or breastfeeding at time of surgery, obese (BMI > 35), mental retardation or psychiatric illness.
Interventions	Group 1: transcutaneous electrical acupoint stimulation at dominant wrist P6 acupoints 30 min before induction of anaesthesia, left on for 24 hours after surgery (n = 65). Group 2: sham transcutaneous electrical acupoint stimulation at dominant wrist P6 acupoints 30 min before induction of anaesthesia but no electrical stimulation activated, left on for 24 h after surgery (n = 65). Ondansetron 4 mg IV and dexamethasone 10 mg given as routine antiemetic treatment for each participant during surgery in both groups.
Outcomes	Nausea (0 ‐ 24 h), vomiting (0 ‐ 24 h), rescue antiemetic (metoclopramide 10 mg IM)
Notes	Power calculation done. Authors reported that no adverse effects (cutaneous irritation, bleeding, nerve injury) occurred associated with treatment groups. Authors have no conflicts of interest. Study supported by grants from Major State Basic Research Development Program of China (973 Program No. 2007CB12502) and National Natural Science Foundation of China (No. 81171235/H0914).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random number table.
Allocation concealment (selection bias)	Low risk	Sequentially‐numbered, opaque sealed envelopes.
Blinding of patients (performance bias) All outcomes	Low risk	"Display screens of the units were concealed from view for patients and other investigators... All patients were told that a tingling or numbing sensation might or might not be felt, regardless of the the group assignment."
Blinding of healthcare providers (performance bias) All outcomes	Low risk	"Display screens of the units were concealed from view for patients and other investigators."
Blinding of outcome assessor (detection bias) All outcomes	Low risk	"Trained nursing staff, who were blinded to the group assignments, assessed PONV..."
Incomplete outcome data (attrition bias) All outcomes	Low risk	11 (8%) participants withdrew, probably due to those who could not be extubated within 2 h after surgery or had impaired consciousness in the neurological intensive care unit.
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported.
Other bias	Low risk	"No differences in patient demographics, risk factors for PONV, duration of anaesthesia, intraoperative opioids and postoperative analgesic consumption between the two groups."