Loudon 2014.

Methods	Study design: multi‐centre RCT Numbers allocated, 28; 15 to exercise intervention, 13 to control Study start: February 2011; stop date: May 2011 Length of intervention: 8 weeks Length of follow‐up: at 12 weeks Country: Tasmania, Australia
Participants	Baseline data available for 12 in intervention and 11 in control: Age, years (mean SD): Intervention: 55.1 (2.5) Control: 60.5 (3.6) Stage, n (%): Intervention: stage 0, 0 (0); stage I, 3 (25); stage II, 6 (50); stage III, 3 (25) Control: stage 0, 1 (9); stage I, 4 (3); stage II, 5 (45); stage III, 1 (9) Inclusion criteria: • Stage I unilateral secondary lymphoedema of the arm, as defined by the International Society of Lymphology and confirmed by a professional lymphoedema therapist • Completed treatment for breast cancer (surgery, radiotherapy, and chemotherapy) at least 6 months previously • Over 18 years of age • Sufficient English literacy to provide informed consent Exclusion criteria: • Recurrent cancer • Infection • Receiving complex lymphoedema therapy • Pregnancy • Wore a pacemaker, which would affect bioimpedance spectroscopy (BIS) readings • Severe psychological illness
Interventions	15 participants assigned to exercise intervention: 8‐Week Yoga intervention consisted of 1 supervised sessions per week (90 minutes) and 6 home‐based sessions per week (45 minutes). Yoga session consisted of documented breathing practices, physical postures, meditation, and relaxation techniques according to the Satyananda Yoga tradition. Participants were given a DVD with a 45‐minute yoga session and were instructed to perform it daily. The DVD followed the same sequence of practices as the class, with fewer postures and shorter relaxation. Participants received a log book in which they recorded their daily practice along with any relevant comments. Adherence: Attendance at group yoga sessions was high (97%), as was self‐reported compliance with the home practice DVD (86%). 13 participants assigned to control: Participants randomised to the control group maintained their usual self‐care as advised by their lymphoedema therapist. Self‐care included wearing of compression sleeves, self‐massage, skin protection, and continued usual lymphatic treatment. Control group was offered yoga classes at completion of the final measurement.
Outcomes	Primary outcome: Arm volume of lymphoedema measured by circumference; extracellular fluid measured by bioimpedance spectroscopy Secondary outcomes: Tissue induration measured by tonometry Severity of sensations, pain, and fatigue, and degree to which sensations, pain, and fatigue limited activity on the day of measurement on a 10‐cm visual analogue scale (VAS). A score of 0 cm indicated “no discomfort”, and a score of 10 cm indicated “the worst imaginable”. Quality of life based on the Lymphoedema Quality of Life Tool (LYMQOL). Total QoL was self‐recorded with scores from 0 to 10, 10 being the best and 0 the worst rating on the day of testing. Subscales, each consisting of several questions, for function, symptoms, appearance, and emotions were also self‐recorded. Each question was scaled from 1 to 4, with 4 being the worst. The score for each subscale was based on the mean of ratings for subscale‐related questions. A higher score indicates a lower QoL rating for that subscale. Numbers of participants assessed: Intervention: baseline, 15; at 8 weeks, 12; at 12 weeks, 9 Control: baseline, 13; at 8 weeks, 11; at 12 weeks, 10 Adverse events: No adverse events were attributable to the yoga or to the control intervention.
Notes	Trial registration link: none available Trial authors contacted: no Intention‐to‐treat analysis: no https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12611000202965 Funding: Swan Research Institute (SRI) and Faculty of Health Sciences Seed Funding, UTAS. Equipment was provided by Flinders University and University of Tasmania.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“randomisation based on a computer‐generated random number system”
Allocation concealment (selection bias)	Low risk	“An individual not associated with the trial will perform the randomisation”. “Group notification will be in a sealed envelope given to women after completion of the baseline measurement”.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Owing to the nature of the intervention, it was not possible to conceal allocation to the intervention from participants.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	“Measurements, based on validated instruments and protocols, were taken by trained researchers blinded to the group allocation and previous results”.
Incomplete outcome data (attrition bias) All outcomes	High risk	Only those who completed post‐intervention and 1‐month‐after‐cessation assessments were included in analysis.
Selective reporting (reporting bias)	Low risk	No selective reporting of outcomes is apparent.
Other bias	Low risk	Trial appears to be free of other problems that could put it at high risk of bias.