Karmakar 2014.

Methods	Blinded (outcome assessor), RCT Sequence generation by computer‐generated allocation number Follow‐up: 6 months
Participants	Participants: 180 adult women in University Hospital in Hong Kong, China Operation: modified radical mastectomy (including axillary lymph node clearance) 3 groups, size: 60, 57, 60 Age (± SD), group 1, 2, 3: 51 (± 9), 54 (± 9), 53 (± 8) All female participants
Interventions	Group 1 (GA group): standardized GA as described below Group 2 (GA + single shot PVB + placebo infusion): pre‐op thoracic paravertebral catheter placed opposite third thoracic spine, ipsilateral to side of surgery, ropivacaine (2 mg/kg) + epinephrine (5 µg/mL) in total volume of 20 mL with normal saline injected slowly then epidural catheter inserted into thoracic paravertebral space. Intraoperatively, continuous infusion of 0.9% saline started at 0.10 mL/kg/h via catheter and maintained constant until 72 h post‐op. Group 3 (GA+ PVB): pre‐op thoracic paravertebral catheter placed opposite third thoracic spine, ipsilateral to side of surgery, ropivacaine (2 mg/kg) + epinephrine (5 µg/mL) in total volume of 20 mL with normal saline injected slowly then epidural catheter inserted into thoracic paravertebral space. Intraoperatively, continuous infusion of ropivacaine 0.25% started at 0.10 mL/kg/h via catheter, maintained constant until 72 h post‐op. All participants had standardized GA, which included IV fentanyl, propofol and rocuronium. Intraoperative morphine (0.1 mg/kg) IV to every participant, then morphine (1 mg IV) as needed, ondansetron 4 mg IV 30 min before end of surgery. In the PACU, all participants had nurse‐administered IV morphine for rescue analgesia as needed. On post‐op ward, analgesia was with diclofenac (75 mg) oral 2 x 72 h, IM morphine (0.1 mg/kg, as needed every 3 h) or Dologesic (paracetamol 325 mg and dextropropoxyphene 32.5 mg, 2 tablets as needed every 6 h) as rescue. Adjuvants: none Immediate pain control: not significantly improved
Outcomes	Dichotomous: incidence of chronic pain at all sites (operated site, axilla, arm) and over operated site at 3 and 6 months Continuous: chronic pain scores at rest and on movement at all sites (operated site, axilla, arm) and over operated site at 3 and 6 months Other reported outcomes: HRQOL (Chinese‐HK version of SF‐36) at 3 and 6 months, Chronic pain symptom and sign score at 3 and 6 months, physical health summary score, mental health summary score (of SF‐36) at 3 and 6 months
Notes	Funding sources: this research work was fully funded by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (RGC reference no. CUHK4406/05, project code 2140452). Conflicts of interest: the study authors declare no conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "patients were randomized to 1 of 3 study groups... with a computer‐generated allocation number"
Allocation concealment (selection bias)	Low risk	Quote: "sequentially numbered, coded, sealed opaque envelopes...The sealed envelopes were prepared by a third party (research assistant) who took no further part in the study"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "patients in group1, who had received standardized GA with no paravertebral intervention, could not be blinded for obvious reasons..For the other 2 study groups that had a thoracic paravertebral catheter placed, we adopted a double‐blind methodology... The principal investigator performed all the thoracic paravertebral catheter placements, collected procedural data, injected the ropivacaine bolus for the TPVB [thoracic paravertebral block], conducted the GA, and took no further part in data collection.. The paravertebral infusion (ropivacaine 0.25% or 0.9% saline) was prepared.. by a postanaesthetic care unit (PACU) nurse not involved in the study ... A single surgeon, who was also blinded to the group allocation, performed or supervised all the surgical procedures"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "a research nurse blinded to the group allocation recorded data preoperatively, in the PACU, and at regular intervals in the postoperative ward...The telephone interview at 3 and 6 months after surgery was also conducted by the same research nurse (blind to group allocation)"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "the primary analyses were performed on a modified intention‐to‐ treat basis (i.e., patients were analysed according to their randomized allocated groups but were excluded from the analysis if they did not adhere to the protocol after randomization)". 1 participant lost to follow‐up in group 2 and reason given (returned overseas after surgery). 2 excluded from the analysis in group 2 because of protocol violation/diagnosed contralateral breast cancer. Very small numbers of attrition, with reasons reported for each exclusion and modified ITT protocol used.
Selective reporting (reporting bias)	Low risk	All primary outcomes in protocol were fully reported on
Null bias	High risk	Quote: "there was no significant difference in acute pain scores at rest (Fig. 2) or on movement (Fig. 3) between the study groups (both P = 0.22) during the 72 hours after surgery".