Chen 2016.

Methods	3‐arm RCT; conducted in Rabies Prophylaxis and Immunity Clinic of Beijing, China Follow‐up: unclear, however, study duration was 6 months
Participants	586 participants with dog bites to their limbs (6 participants receiving standard care were lost to follow‐up and excluded from the analysis). Participants had an average of 4.6 wounds. Inclusion criteria: aged > 18 years with lacerated limb wounds demanding surgical treatment (lacerations with lengths of ≥ 5 cm, multiple penetrating lacerations of soft tissues, or lacerations accompanied by damage to ≥ 1 of the following: muscles, tendons, ligaments, nerves, articular capsules, fractures, important blood vessels, acra [sic] amputation) Exclusion criteria: puncture wounds (< 2 mm); medium or small lacerations (< 5 cm); infected wounds at presentation; having visited a physician's office > 8 h after the injury; wounds with skin loss requiring plastic surgery; or patients with immune deficiency, using immunosuppressive agents, or with autoimmune disorders or diabetes; refusal to give consent
Interventions	Group A (n = 335 participants): standard care – sterilised dressings Group B (n = 123 participants): polyvinyl alcohol shrink formaldehyde bubble dressing + NPWT set at 125 mm Hg Group C (n = 128 participants): polyvinyl alcohol shrink formaldehyde bubble dressing + NPWT set at 75 mm Hg Co‐intervention: wounds were cleaned and disinfected. This was followed by debridement and important tissue repair after which wounds were covered. All participants were given rabies prophylactic active immunity and/or passive immunity, tetanus antitoxin was also given where necessary. Drainage was also performed, however, removed 24‐48 h before surgery. Antibiotics were administered only after wound infection occurred.
Outcomes	Primary review outcomes: wound infection*; time to clinical healing (defined as the interval from being bitten to the wound achieving clinical healing) Secondary review outcomes: none reported *defined as the satisfaction of 1 of the following 3 major criteria: fever (body temperature ≥ 38°C), abscess, and lymphangitis, or 4/5 minor criteria: wound‐associated erythema that extended > 3 cm from the edge of the wound, tenderness at the wound site, swelling at the site, purulent drainage, and a white blood cell count in the peripheral blood of 12,000/mL
Notes	The study authors declare no conflicts of interest. Funding source was not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “All participants were subjected to block randomisation…using a block random digits table” Comment: a block random digits table used to ensure the randomisation
Allocation concealment (selection bias)	Unclear risk	Quote: see above Comment: methods of allocation concealment unclear
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: no quote Comment: no mention of blinding but would not be possible to blind health professionals to the different treatments so without further detail considered high risk
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: no quote Comment: no mention of blinding in the paper
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: no quote Comment: 6 participants in the standard care group without clarifying the reasons vs 0 were lost to follow‐up in the NPWT group
Selective reporting (reporting bias)	Unclear risk	Quote: no quote Comment: participants had an average of 4.6 wounds, which would usually result in unit of analysis issues, however, wound infection data suggest the unit of analysis was the participants.
Other bias	Low risk	Quote: no quote Comment: there was a difference in number of participants with wounds in multiple locations, however, this is unlikely to have introduced bias since the unit of randomisation was the participant; no other areas of potential risk were noted.