Naeem 2012.

Methods	Study design: RCT Duration of study: August 2008 and August 2010
Participants	Sample size: 64 (only 52 (86.67%) were lactational breast abscesses. From the text it appears that 11 had non‐puerperal breast abscesses, but that the percentages do not correspond to the actual numbers. We have contacted authors for clarification. Inclusion criteria: Any female with a single abscess smaller than 5 cm in a reproductive‐aged group who was not pregnant at the time and not being treated for any other breast pathology. Exclusion criteria: Women with sinus/fistula of breast abscess, prolonged history and necrosis of the skin. Setting: KVVS hospital, Karachi. Country: Pakistan.
Interventions	1. Incision and drainage (n = 32) The intervention was carried out in a hospital setting. Women were admitted between 1 to 3 days. A general anaesthetic was given. Before surgery the infected fluid was sent for culture and sensitivity and cytology. 625 mg capsule of co‐amoxiclav 3 times a day were and 400 mg metronidazole (non‐lactational breast abscesses) until culture reports were received. IV analgesia was initially given. IM analgesia was initially given and then followed by oral analgesia. Daily dressings were for 1 to 3  weeks  Women were followed up for 8 weeks after the procedure. Women were unable to feed with the affected breast so milk was discarded after expressing with breast pump according to the authors. 2. Ultrasound‐guided needle aspiration. (n = 32) The procedure was done in the OPD therefore requiring no hospital admission. No anaesthetic was given. The aspiration was done using ultrasound guidance. Prior to the procedure, samples were sent for culture and sensitivity and cytology. 625 mg capsule of co‐amoxiclav 3 times a day and (400 mg metronidazole) was given to non‐puerperal abscesses. The cavity was washed with normal saline and a follow‐up ultrasound was done on the third day. If the abscess was seen, the canula was left in place. IM analgesia was initially given. Follow‐up was done for 8 weeks after the procedure. Women continued to feed.
Outcomes	Authors did not differentiate between primary and secondary outcomes. 1.Time taken to resolve symptoms (point tenderness; erythema; hyperthermia) 2.Recurrence of breast abscess 3.Healing time
Notes	Funding: not reported. Conflict of interest: authors did not declare any competing interests. Ethics approval: not stated. Author contacted: Partial response received from corresponding author below: "Our review is assessing puerperal breast abscesses only and the management thereof".
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	It is not stated how they were randomly divided.
Allocation concealment (selection bias)	Unclear risk	Not stated.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Although there was no blinding the resolution of the abscess would not be affected.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessors is not possible due to the nature of the interventions and therefore difficult to judge whether this has an impact.
Incomplete outcome data (attrition bias) All outcomes	High risk	Table 1: Group A, n = 32 however outcomes for 30 only addressed. For group B there is also one missing, the values only add up to 31. The response from the author was that in the analysis there were:  Group A ‐ 30 women and group B 31. Upon requesting information from the corresponding author he mentioned that 3 participants were missing at week 8 follow‐up ‐ no reasons were given as to what happened to them. This information was not in the study.
Selective reporting (reporting bias)	Low risk	Healing time was addressed and abscess recurrence. However with regards to time to resolving of symptoms, only breast pain was addressed, erythema and hyperthermia not addressed.
Other bias	Unclear risk	The authors report that skin indurations around abscesses were present in Group A in 93.75% whereas in Group B only 71.8%. They say the P value is significant, but it is 0.20. There is also no table of baseline characteristics of both groups, so we cannot really judge whether there were any baseline differences between groups.