Treatments for breast abscesses in breastfeeding women

Hayley Irusen; Anke C Rohwer; D Wilhelm Steyn; Taryn Young

doi:10.1002/14651858.CD010490.pub2

. 2015 Aug 17;2015(8):CD010490. doi: 10.1002/14651858.CD010490.pub2

Treatments for breast abscesses in breastfeeding women

Hayley Irusen ^1,^✉, Anke C Rohwer ², D Wilhelm Steyn ³, Taryn Young ^2,⁴

Editor: Cochrane Pregnancy and Childbirth Group

PMCID: PMC9226721 PMID: 26279276

Abstract

Background

The benefits of breastfeeding are well known, and the World Health Organization recommends exclusive breastfeeding for the first six months of life and continuing breastfeeding to age two. However, many women stop breastfeeding due to lactational breast abscesses. A breast abscess is a localised accumulation of infected fluid in breast tissue. Abscesses are commonly treated with antibiotics, incision and drainage (I&D) or ultrasound‐guided needle aspiration, but there is no consensus on the optimal treatment.

Objectives

To assess the effects of different treatments for the management of breast abscesses in breastfeeding women.

Search methods

We searched the Cochrane Pregnancy and Childbirth Group’s Trial Register (27 February 2015). In addition we searched African Journals Online (27 February 2015), Google Scholar (27 February 2015), ProQuest Dissertations and Theses Databases (27 February 2015) and the WHO International Clinical Trials Registry Platform (ICTRP) search portal (27 February 2015). We also checked reference lists of retrieved studies and contacted experts in the field as well as relevant pharmaceutical companies.

Selection criteria

Randomised controlled trials (RCTs) investigating any intervention for treating lactational breast abscesses compared with any other intervention. Studies published in abstract form, quasi‐RCTs and cluster‐RCTs were not eligible for inclusion.

Data collection and analysis

Two review authors independently assessed studies for inclusion, assessed risk of bias and extracted data. Data were checked for accuracy.

Main results

We included six studies. Overall, trials had an unclear risk of bias for most domains due to poor reporting. Two studies did not stratify data for lactational and non‐lactational breast abscesses, and these studies do not contribute to the results. This review is based on data from four studies involving 325 women.

Needle aspiration (with and without ultrasound guidance) versus incision and drainage (I&D)

Mean time (days) to complete resolution of breast abscess (three studies) ‐ there was substantial heterogeneity among these data (Tau² = 47.63, I² = 97%) and a clear difference between subgroups (with or without ultrasound guidance; Chi² = 56.88, I² = 98.2%, P = < 0.00001). We did not pool these data in a meta‐analysis. Two studies excluded women who had treatment failure when they calculated the mean time to complete resolution. One study found that the time to complete resolution of breast abscess favoured needle aspiration over I&D (mean difference (MD) ‐6.07; 95% confidence interval (CI) ‐7.81 to ‐4.33; n = 36), but excluded 9/22 (41%) women in the needle aspiration group due to treatment failure. Another study reported faster resolution in the needle aspiration group (MD ‐17.80; 95% CI ‐21.27 to ‐14.33; n = 64) but excluded 6/35 (17%) women in the needle aspiration group due to treatment failure. A third study also reported that needle aspiration was associated with a shorter time to complete resolution of breast abscess (MD ‐16.00; 95%CI ‐18.73 to ‐13.27; n = 60); however, the authors did not indicate the number of women who were lost to follow‐up for either group, and it is unclear how many women contributed to this result. Considering the limitations of the available data, we do not consider the results to be informative.

Continuation of breastfeeding, after treatment (success): results favoured the needle aspiration group, but we did not pool data from the two studies because of substantial unexplained heterogeneity (I² = 97%). One study reported that women in the needle aspiration group were more likely to continue breastfeeding (risk ratio (RR) 2.89; 95% CI 1.64 to 5.08; n = 60), whereas the other study found no clear difference (RR 1.09; 95% CI 0.97 to 1.22 n = 70).

Treatment failure was more common among women treated with needle aspiration compared to those who underwent I&D (RR 16.12; 95% CI 2.21 to 117.73; two studies, n = 115, low quality evidence). In one study, treatment with needle aspiration failed in 9/22 women who subsequently underwent I&D to treat their breast abscess. In another study, treatment with needle aspiration failed in 6/35 women, who subsequently underwent I&D. All abscesses in the I&D group were successfully treated.

The included studies provided limited data for the review's secondary outcomes. No data were reported for adverse events. One study (60 women) reported that women in the needle aspiration group were more satisfied with their treatment than women who received I&D to treat their breast abscesses.

Incision and drainage (I&D) with or without antibiotics

One study (150 women) compared the value of adding a broad‐spectrum cephalosporin (single dose or a course of treatment) to women who underwent I&D for breast abscesses.

The mean time to resolution of breast abscess was reported as being similar in all groups (although women with infection were excluded). Mean time to resolution for women who received a course of antibiotics was reported as 7.3 days, 6.9 days for women who received a single dose of antibiotics and 7.4 days for women who did not receive antibiotics. Standard deviations, P values and CIs were not reported and prevented further analysis. No data were reported for any continuation of breastfeeding after treatment (success). For treatment failure, there was no clear difference between the groups of women who received antibiotics (either a single dose or a course of antibiotics) and those who did not (RR 1.00; 95% CI 0.36 to 2.76).

Included studies rarely reported this review's secondary outcomes (including adverse events). For post‐operative complications/morbidity, there was no difference in the risk of wound infections between the antibiotics and no antibiotics groups (RR 0.58; 95% CI 0.29 to 1.17), irrespective of whether women received a single dose or a course of antibiotics.

Authors' conclusions

There is insufficient evidence to determine whether needle aspiration is a more effective option to I&D for lactational breast abscesses, or whether an antibiotic should be routinely added to women undergoing I&D for lactational breast abscesses. We graded the evidence for the primary outcome of treatment failure as low quality, with downgrading based on including small studies with few events and unclear risk of bias.

Plain language summary

Treatments for breast abscesses in breastfeeding women

Some women develop a breast abscess while breastfeeding, called a lactational breast abscess. An abscess is a collection of infected fluid within the breast tissue. The aim of treatment is to cure the abscess quickly and effectively, ensuring maximum benefit to the mother with minimal interruption of breastfeeding.

Presently, lactational breast abscesses are treated by incision and drainage or needle aspiration, with or without diagnostic ultrasound. Antibiotics may or may not be prescribed. For incision and drainage the abscess is cut open with a scalpel (blade) to release the infected fluid. A drain may be inserted into the wound to help the infected fluid drain or may be left open so that the infected fluid drains naturally. A less invasive way to treat the breast abscess is by needle aspiration. A needle is inserted into the cavity of the breast abscess and a syringe is used to draw out the infected fluid, often using ultrasound guidance. As there are advantages in using this method e.g. no scars, reduced hospitalisation etc. the trend is to use this method more often.

We wanted to find evidence on the effectiveness of different treatments. We looked at the time taken for the abscess to heal using the different types of treatments, the number of women who continued to breastfeed after treatment and how many women had healed in the each group after treatment. The definition of healing varied across the studies.

We found six studies, of which four studies with a total of 325 woman contributed data. These studies compared needle aspiration versus incision and drainage. Needle aspiration appeared to decrease the healing time compared to incision and drainage, but large proportions of women were excluded from the analysis and it was therefore difficult to make conclusions. For the outcome continuation of breastfeeding, both of the studies showed that women treated with needle aspiration were more likely to continue breastfeeding compared to incision and drainage. In two studies, breast abscesses did not heal in some women who had needle aspiration and had to be treated with incision and drainage (low quality evidence). All breast abscesses that were treated with incision and drainage healed. We were not able to make any conclusions regarding unwanted effects or complications. Studies did not report sufficiently on the number of follow‐up visits, duration of continuation of breastfeeding, post‐operative complications, duration of hospital stay and adverse events. However, it appeared that women were more satisfied when treated with needle aspiration.

One study compared different regimens of antibiotics versus no antibiotics in breastfeeding women who were treated with incision and drainage for breast abscesses. We did not find any difference between groups for the outcome resolution of breast abscesses and infections after the procedure.

All of the studies were poorly conducted and/or reported and did not address all of the outcomes that we were interested in. Studies with better design and reporting are needed to properly assess these outcomes.

Summary of findings

Summary of findings for the main comparison. Needle aspiration compared with incision and drainage for breast abscesses in breastfeeding women.

Needle aspiration compared with incision and drainage for breast abscesses in breastfeeding women
Patient or population: Breastfeeding women with breast abscesses  Intervention: Needle aspiration  Comparison: incision and drainage
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect  (95% CI)	No of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Assumed risk	Corresponding risk
	incision and drainage	Needle aspiration
Time to resolution of breast abscess (days)	This outcome was a dressed by three studies with severe heterogeneity (I² = 97%), therefore the result was not pooled.
Continuation of breastfeeding	The result for this outcome was not pooled as it was provided by two studies of small sample size with severe heterogeneity (I² = 97%)
Treatment failure	Study population		RR 16.12  (2.21 to 117.73)	115  (2 RCTs)	⊕⊕⊝⊝  LOW ^1,2
	0 per 1000	0 per 1000  (0 to 0)
	Moderate
	0 per 1000	0 per 1000  (0 to 0)
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  CI: Confidence interval;
GRADE Working Group grades of evidence  High quality: Further research is very unlikely to change our confidence in the estimate of effect.  Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.  Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.  Very low quality: We are very uncertain about the estimate.  CI: confidence interval; RR: risk ratio

Open in a new tab

¹Evidence provided by studies of unclear risk of bias

²Included studies were of small sample size with few events (<30)

Background

Description of the condition

The benefits of breastfeeding are well known and the World Health Organization (WHO) thus recommends exclusive breastfeeding for the first six months and continuing for up to two years and beyond. (WHO 2003; WHO 2012). However, there are many reasons why women stop breastfeeding; one of the most common being the complications of lactation (Dener 2003). Of note for this review, Amir 2004 found in a study of women who commenced breastfeeding that 0.4% (5/1183) developed a breast abscess.

Mastitis

Mastitis is an inflammatory condition of the breast that is usually associated with lactation and that can progress from the non‐infective stage, to infective mastitis and then to a breast abscess. The incidence of mastitis in lactating women is between 3% to 20% due to variations in the definition and follow‐up in the post partum period (Amir 2014). Mastitis is clinically characterised by a tender, hot, swollen, wedge‐shaped area of the breast associated with high temperatures (> 38.5°C ) and flu‐like symptoms. It may or may not be accompanied by an infection (Amir 2014). Some of the predisposing factors are limited feeding, poor positioning of the baby, illness of mother or baby, maternal malnutrition and cracked nipples. In infective mastitis, Staphylococcus aureus and Staphylococcus epidermidis are the commonest causative organisms (Amir 2014). Mastitis usually occurs during the first six weeks but can occur at any time during lactation (Amir 2014). The primary cause of mastitis is milk stasis (Hughes 1989). Conservative management includes efficient removal of milk, with the addition of antibiotics for possible bacterial infections (Baker 2010; Marchant 2002). Other measures include supportive care; rest and fluids, application of heat packs and analgesics. Antibiotics are recommended if symptoms have not improved (Lawrence 2005), although a Cochrane systematic review found insufficient evidence, due to a lack of studies, to confirm or refute when to use antibiotics in the treatment of mastitis (Jahanfar 2012).

Lactational breast abscess

A breast abscess is defined as a localised accumulation of infected fluid in breast tissue. Breast abscesses are usually puerperal (lactational) but can be non‐puerperal (Baker 2010). Three per cent of women with mastitis develop a lactational breast abscess ( Amir 2014).

The most common causative organism is Staphylococcus aureus (WHO 2003), although other organisms have been identified (Bertrand 1991; Dixon 1988; Karstrup 1993). A recent study has suggested that methicillin‐resistant Staphylococcus aureus (MRSA) is also beginning to play an important role (Branch‐Elliman 2012). Risk factors for developing lactational breast abscesses include: women over the age of 30, first pregnancies, gestational age ≥ 41 weeks and mastitis (Kvist 2005). A breast abscess usually presents as a hard, tender and sometimes fluctuant mass with overlying erythema (redness of the skin) (Barbosa‐Cesnick 2003). Diagnosis is usually made using ultrasound when a hypoechoic lesion with an irregular border is present (Dirbas 2011).

Three Cochrane reviews (Crepinsek 2012; Lumbiganon 2012; Mangesi 2010) have illustrated the need for education about breastfeeding during pregnancy and to determine effective treatments for the prevention of mastitis and engorgement, conditions which contribute towards the formation of lactational breast abscesses. For women at risk of developing a lactational breast abscesses, it is therefore necessary to examine existing studies on treatments for lactational breast abscesses, to understand its impact on maternal health, time to recovery and its effect on breastfeeding.

Description of the intervention

Approaches to treating breast abscesses include incision and drainage (I&D), usually carried out under general anaesthesia and needle aspiration, which may be a single aspiration with a drain left in situ or serial aspirations. Needle aspiration is usually done with a local anaesthetic. Antibiotics are recommended following either a needle aspiration or I&D (Abou‐Dakn 2010). Delayed, inappropriate or even inadequate treatment may result in more extensive lesions and permanent tissue damage, which could affect future lactation in about 10% of women. Breast abscesses that require extensive resection can cause disfigurement (World Health Organization 2000).

Treatments

1. Antibiotics

Treatment of lactational breast abscesses with antibiotics, without removal of pus is considered to be ineffective (World Health Organization 2000). Following diagnostic or interventional ultrasound or I&D of breast abscesses, breast milk and fluid samples should be sent for culture to detect the presence of bacteria or resistant pathogens (Amir 2014). The most commonly found organism in a lactational breast abscess is Staphylococcus aureus with Steptococcus orEscherichia coli being less common. Antibiotics of choice such as dicloxacillin or flucloxacillin 500mg four times daily orally, or the recommended sensitive local antibiotic may be prescribed. First generation cephalosporins may also be an alternative. Women who may be allergic to penicillin may be prescribed cephalexin or clindamycin. In cases where Staphylococcus Aureus is resistant to penicillinase‐resistant penicillins (methicillin‐resistant Staphylococcus Aureus (MRSA)) is suspected, breast milk culture and assay of antibiotics sensitivities should be undertaken. Most strains of MRSA are sensitive to vancomycin or trimethoprim/sulphamethoxazole and less so to rifampin. One should presume that MRSA is resistant to treatment with macrolides and quinolones regardless of susceptibility test results (Amir 2014)

2. Surgical

Lactational breast abscesses have traditionally been treated with I&D, but more recently there is a growing tendency to use less invasive procedures. Where possible, all women with a suspected lactational breast abscess should have an ultrasound, which will be helpful in identifying all pockets of fluid. Management may depend on the state of the overlying skin. For skin that appears normal, drainage of the abscess is done by needle aspiration usually with ultrasound (see below). If the skin over the abscess is thin and shiny or the abscess appears as if it will burst, then I&D is recommended (Dirbas 2011).

Incision and drainage is done with local or general anaesthetic. An incision is made to allow for drainage of the infected fluid and if a drain is required a counter incision is then made. Daily washing out of the wound may be required until secretions decrease or are clear. By week four, the wound should be closed and without complications. I&D is recommended when the abscess is large or if there are multiple abscesses. A course of antibiotics is also advised (Abou‐Dakn 2010) (see below).

3. Needle aspiration

Breast abscesses are also treated with needle aspiration, using a local anaesthetic and under sterile conditions, with or without ultrasound guidance. The Society of Interventional Radiology (SIR) defines image‐guided percutaneous aspiration as "evacuation or diagnostic sampling of a fluid collection with the use of a catheter or a needle during a single imaging session, with removal of the catheter or needle immediately after the aspiration" while image‐guided percutaneous drainage is defined as "the placement of a catheter with the use of image guidance to provide continuous drainage of a fluid collection" (Wallace 2010, p432). It may be performed during a single session or as a staged procedure during multiple sessions (Wallace 2010).

The (WHO (World Health Organization 2000), supports the use of ultrasound guidance for diagnosis and treatment of lactational breast abscesses. Ulitzsch 2004 has shown that abscesses of less than three cm in diameter can be treated with single aspiration or serial aspirations until resolution. Failure was seen with abscesses greater than five cm in diameter. A probe or a drain is an alternative to using a needle to remove the infected fluid. If the aspirate is viscous, then a saline or antibiotic solution can be used to assist with the aspiration. Daily aspirations are recommended until the wound cannot be punctured anymore (< 4 mm). Serial aspirations are done between two to nine times. A course of antibiotics is usually recommended (Abou‐Dakn 2010).

Although needle aspiration is considered as being less invasive, not all lactational breast abscesses have been successfully treated by this method and have subsequently needed I&D (Ozseker 2008; Ulitzsch 2004). Some of the reasons cited for treatment failure include lack of clinical improvement, recurrence of abscess or formation of fistulas (Giess 2014).

Breastfeeding

Prior to drainage of the breast abscess, breastfeeding should continue from the unaffected breast. Breastfeeding from the affected breast should resume soon after drainage to prevent stasis of milk and relapse of the infection (World Health Organization 2000). Feeding from the affected breast is recommended, even if a drain is in place but care should be taken to ensure that the infant's mouth is not in contact with the infected fluid or breast tissue (Amir 2014). Failure to allow breastfeeding lends itself to the production of fluid that is viscous, which aggravates engorgement. Breastfeeding ensures drainage of the affected area and speedy resolution of the abscess (Walker 2011). Giess 2014 recommends that breastfeeding can and should continue from the lactating breast with the proviso that the prescribed antibiotics are safe for the infant. This encourages adequate drainage, which facilitates clearing of the infection and limits the bacterial culture medium.

How the intervention might work

The objective of any of the interventions employed in treating an abscess is to remove the infected fluid as speedily as possible, hastening resolution, thereby reducing the pain and discomfort and allowing the woman to continue breastfeeding her infant with little or no interruption. Maintaining the integrity of the breast is also important, i.e. the procedure should leave the woman complication‐free, with minimal or preferably no scarring, and the function of breastfeeding should be maintained.

Antibiotics and I&D have been viewed as standard therapy in managing lactational breast abscesses. More recently, however, there has been an emergence of studies favouring treatment of lactational abscesses with needle aspiration, which is considered a less invasive technique.

Christensen 2005 favoured the use of ultrasound‐guided drainage of breast abscesses as it caused less scarring, did not affect breastfeeding, did not require anaesthesia or hospitalisation, and was less expensive than surgery. Although I&D has the advantage of breaking down the loculi, if the procedure is carried out under a general anaesthetic, it will also involve hospitalisation and regular dressings. This is thought to cause considerable distress to both mother and baby during what is already a difficult time and the final cosmetic result may be unsatisfactory (Benson 1989; Dixon 1998). Scholefield 1987 expressed a similar view, suggesting that I&D is associated with a prolonged healing time, regular dressings, difficulties in breastfeeding, and the possibility of an unsatisfactory cosmetic outcome. Conversely Jones 1976 and Ajao 1994 found that I&D, curettage and primary closure of the abscess cavity had better scar formation and a reduction of cost of treatment.

Effective management of a lactational breast abscess is necessary to eliminate discomfort and reduce the risk of discontinuation of breastfeeding. Breastfeeding is regarded as fundamental to the growth and development of an infant, and it is therefore important that whatever the intervention is, it should not disrupt any momentum gained by the mother with regards to breastfeeding (Walker 2011).

Why it is important to do this review

There have been a number of Cochrane reviews addressing questions around prevention and treatment of breastfeeding complications (Crepinsek 2012; Lumbiganon 2012; Mangesi 2010).

Mangesi 2010 examined treatments for breast engorgement during lactation with one of its key outcomes being mastitis and the secondary outcome as breast abscess formation. One study showed that there was a difference in breast abscesses between the group that received acupuncture and those that did not, however, this study was underpowered and the results were not statistically significant.

Lumbiganon 2012 looked at antenatal breastfeeding education in increasing breastfeeding duration. As a secondary outcome, they also listed breastfeeding complications such as mastitis and breast abscesses. The authors reported that compared to formal breastfeeding education plus lactation consultation versus routine breastfeeding, education showed no significant difference in mastitis but a significant reduction in nipple pain. Crepinsek 2012 examined the effect of different interventions for the prevention of mastitis following childbirth. They showed that none of the interventions were effective in preventing mastitis. As appropriate studies were not available at the time the review by Jahanfar 2012 was done, the author was unable to support or deny the role antibiotics played in treating mastitis.

A recently published non‐Cochrane systematic review on the treatment of breast abscesses Lam 2014 included randomised controlled trials, non‐randomised trials as well as case series. Participants had lactational or non‐lactational breast abscesses and one study included men. Although the authors used SORT (Strength of Recommendation Taxonomy) to grade the quality of evidence and the recommendations made, it is not clear how judgements about risk of bias were made. The authors recommend the use of needle aspiration with or without the use of ultrasound as first line treatment of breast abscesses. No meta‐analysis was conducted to measure treatment effects.

Lam 2014 does not recommend breastfeeding from the affected breast due to Staphylococcal organisms, which places the infant at risk of pneumonia, lung abscesses and death. This recommendation contradicts other current literature (Amir 2014; Giess 2014).

Effective interventions for the prevention of engorgement and mastitis are still to be determined. In the absence of these interventions there is an increased likelihood of developing a breast abscess. Currently, there appears to be no consensus on which the best treatment for lactational breast abscess is and to this end there is a need to rigorously synthesise existing research to obtain clarity.

Objectives

To assess the effects of different treatments for the management of breast abscesses in breastfeeding women.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs). Trials using cluster‐randomised or cross‐over designs were not eligible for inclusion. As per protocol, quasi‐randomised trials were also excluded as we identified RCTs. Studies only reported as abstracts were not included in the review. Future updates of this review may consider including quasi‐RCTs and cluster‐RCTs (due to paucity of data).

Types of participants

Breastfeeding women (exclusive breastfeeding or mixed‐feeding) presenting with breast abscesses in one or both breasts. Women with co‐morbidities were included (e.g. HIV, diabetes).

Types of interventions

Any intervention, surgical, non‐surgical, pharmacological, non‐pharmacological, invasive, non‐invasive, or a combination of treatments, to treat lactational breast abscesses, compared with any other intervention, surgical, non‐surgical, pharmacological, non‐pharmacological, invasive, non‐invasive, or a combination of treatments, aimed at treating lactational breast abscesses.

Types of outcome measures

Primary outcomes

Time to complete resolution of breast abscess (resolution of abscess was defined as no recurrence of abscess or need for any intervention). Time was defined by the authors as time of presentation for care or from time of hospitalisation.
Any continuation of breastfeeding after treatment (success).
Treatment failure.

Secondary outcomes

Number of follow‐up visits.
Duration of continuation of breastfeeding after treatment.
Maternal satisfaction with treatment.
Post‐operative complications/morbidity.
Duration of hospital stay.
Adverse events.

Search methods for identification of studies

The following methods section of this review is based on a standard template used by the Cochrane Pregnancy and Childbirth Group.

Electronic searches

We searched the Cochrane Pregnancy and Childbirth Group’s Trials Register by contacting the Trials Search Co‐ordinator (27 February 2015).

The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co‐ordinator and contains trials identified from:

monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
weekly searches of MEDLINE (Ovid);
weekly searches of Embase (Ovid);
monthly searches of CINAHL (EBSCO);
handsearches of 30 journals and the proceedings of major conferences;
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Details of the search strategies for CENTRAL, MEDLINE, Embase and CINAHL, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group.

Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co‐ordinator searches the register for each review using the topic list rather than keywords.

In addition, we carried out supplementary searches of African Journals Online (27 February 2015), dissertation databases, trial registries for ongoing studies and Google Scholar (27 February 2015). For dissertations we searchedProQuest Dissertations and Theses Databases (27 February 2015). For ongoing trials, we searched the WHO International Clinical Trials Registry Platform Search Portal (ICTRP) (27 February 2015). See Appendix 1 for search terms used in these databases.

Searching other resources

We checked the reference lists of included studies for relevant citations and contacted experts in the field in order to find any unpublished studies. We also contacted the following pharmaceutical companies: Aspen, Glaxo Smithkline, Novartis, Pfizer and Roche for relevant studies.

We did not apply any language or date restrictions.

Data collection and analysis

Selection of studies

Two review authors (Hayley Irusen (HI) and Anke Rohwer (AR)) independently assessed for inclusion all the potential studies identified as a result of the search strategy. We screened titles and abstracts of search results to exclude irrelevant studies. We then retrieved full text articles of seemingly relevant studies and examined them to see whether they met the inclusion criteria. We resolved any disagreement through discussion and by consultation with the third review author (Taryn Young (TY)).

Data extraction and management

We designed a form to extract data (Appendix 2). For eligible studies, two review authors (HI and AR) extracted the data using the agreed form. We resolved discrepancies through discussion and consultation of a third author (TY). We entered data into Review Manager software (RevMan 2014) and checked them for accuracy.    When information regarding any of the above was unclear, we contacted authors of the original reports to provide further details. Of the six included studies (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012), only one author responded in part (Naeem 2012). Where studies reported ranges, we used BMJ online 2014 as a resource to provide the statistical method to convert ranges to mean and standard deviations (Suthar 2012).

Assessment of risk of bias in included studies

Two review authors (HI and AR) independently made judgements about risk of bias. Discrepancies were resolved through discussion and by consultation with the third review author (TY) if they were not resolved.

(1) Random sequence generation (checking for possible selection bias)

We described for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it produced comparable groups.

We assessed the method as:

low risk of bias (any truly random process, e.g. random number table; computer random number generator);
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
unclear risk of bias.

(2) Allocation concealment (checking for possible selection bias)

We described for each included study the method used to conceal allocation to interventions prior to assignment and assessed whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.

We assessed the methods as:

low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
unclear risk of bias.

(3.1) Blinding of participants and personnel (checking for possible performance bias)

We described for each included study the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We considered studies at low risk of bias if they were blinded, or if we judged that the lack of blinding would be unlikely to affect results. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed the methods as:

low, high or unclear risk of bias for participants;
low, high or unclear risk of bias for personnel.

(3.2) Blinding of outcome assessment (checking for possible detection bias)

We described for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed methods used to blind outcome assessment as:

low, high or unclear risk of bias.

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

We described for each included study, and for each outcome or class of outcomes, the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported, or could be supplied by the trial authors, we re‐included missing data in the analyses which we undertook.

We assessed methods as:

low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of intervention received from that assigned at randomisation);
unclear risk of bias.

(5) Selective reporting (checking for reporting bias)

We described for each included study how we investigated the possibility of selective outcome reporting bias and what we found.

We assessed the methods as:

low risk of bias (where it was clear that all of the study’s pre‐specified outcomes and all expected outcomes of interest to the review had been reported);
high risk of bias (where not all the study’s pre‐specified outcomes had been reported; one or more reported primary outcomes were not pre‐specified; outcomes of interest were reported incompletely and so could not be used; study failed to include results of a key outcome that would have been expected to have been reported);
unclear risk of bias.

(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)

We described for each included study any important concerns we had about other possible sources of bias.

We assessed whether each study was free of other problems that could have put it at risk of bias:

low risk of other bias;
high risk of other bias;
unclear whether there is risk of other bias.

(7) Overall risk of bias

We made explicit judgements about whether studies were at high risk of bias, according to the criteria given in the Cochrane Handbook for Systematic Interventions Reviews (Higgins 2011). With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it likely to impact on the findings.

The quality of the evidence was assessed using the GRADE approach (Schunemann 2009) in order to assess the quality of the body of evidence relating to the following outcomes for the main comparison of needle aspiration compared to I&D.

Time to complete resolution of breast abscess
Any continuation of breastfeeding after treatment (success)
Treatment failure

We used GRADE profiler (GRADEpro 2014) to import data from Review Manager 5.3 (RevMan 2014) in order to create a ’Summary of findings’ table. A summary of the intervention effect and a measure of quality for each of the above outcomes was produced using the GRADE approach. The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of evidence for each outcome. The evidence was downgraded from 'high quality' by one level for serious (or by two levels for very serious) limitations, depending on assessments for risk of bias, indirectness of evidence, serious inconsistency, imprecision of effect estimates or potential publication bias.

Measures of treatment effect

Dichotomous data

For dichotomous data, we presented results as summary risk ratios with 95% confidence intervals. Outcomes with dichotomous data included any continuation of breastfeeding after treatment, resolution of abscess, and post‐operative complications/morbidity.

Continuous data

For continuous data, we used the mean difference, since outcomes were measured in the same way between trials. Outcomes with continuous data included time to complete resolution of abscess, number of follow‐up visits, duration of continuation of breastfeeding after treatment, maternal satisfaction with treatment and duration of hospital stay. For length of time to resolution of abscess, two studies (Chandika 2012; Saleem 2008) reported the range of values only. In these circumstances, we estimated mean and SD (BMJ online 2014).

Unit of analysis issues

Cluster‐randomised trials were not eligible for inclusion. However, due to paucity of data, cluster‐randomised trials will be eligible for inclusion in future updates.

Cluster‐randomised trials

In future updates, we will include cluster‐randomised trials in the analyses along with individually randomised trials. We will adjust their sample sizes or standard errors using the methods described in the Handbook using an estimate of the intracluster correlation co‐efficient (ICC) derived from the trial (if possible), from a similar trial or from a study of a similar population. If we use ICCs from other sources, we will report this and conduct sensitivity analyses to investigate the effect of variation in the ICC. If we identify both cluster‐randomised trials and individually‐randomised trials, we plan to synthesise the relevant information. We will consider it reasonable to combine the results from both if there is little heterogeneity between the study designs and the interaction between the effect of intervention and the choice of randomisation unit is considered to be unlikely.

We will also acknowledge heterogeneity in the randomisation unit and perform a sensitivity analysis to investigate the effects of the randomisation unit.

Cross‐over trials

Trials using a cross‐over design were not eligible for inclusion in this review as this is not an appropriate study design for the interventions in this review.

Other unit of analysis issues

Studies with two or more than two treatment groups were included and were dealt with as recommended by the Cochrane Handbook for Systematic Interventions Reviews (Higgins 2011).When a multi‐arm study contributed more than one comparison to a particular meta‐analysis, we either combined treatment groups or divided the control group, so that the inclusion of data from the same woman more than once in the same analysis was avoided.

Dealing with missing data

No imputation of missing data was done. Where the required summary statistics were not reported, these were calculated from the available data according to the Cochrane Handbook Chapter 7.7 (Higgins 2011), specifically where means and confidence intervals and sample sizes per group were reported, standard deviations were calculated in the recommended manner.

For included studies, we noted levels of attrition.    For all outcomes, we carried out analyses, as far as possible, on an intention‐to‐treat basis, i.e. we attempted to include all participants randomised to each group in the analyses, and analysed all participants in the group to which they were allocated, regardless of whether or not they received the allocated intervention. The denominator for each outcome in each trial was the number of participants randomised.

Assessment of heterogeneity

We assessed statistical heterogeneity in each meta‐analysis using the T², I² and Chi² statistics. We regarded heterogeneity as substantial if the T² was greater than zero and either the I² was greater than 30% or there was a low P (less than 0.10) in the Chi² test for heterogeneity. For significant heterogeneity, we used the random‐effects model or reported results narratively.

Assessment of reporting biases

We did not investigate reporting biases due to the limited number of included studies. In future updates of this review, if there are 10 or more studies in the meta‐analysis we will investigate reporting biases (such as publication bias) using funnel plots. We will assess funnel plot asymmetry visually. If asymmetry is suggested by a visual assessment, we will perform exploratory analyses to investigate it.

Data synthesis

We carried out statistical analysis using the Review Manager software (RevMan 2014). We planned to use fixed‐effect meta‐analysis for combining data where it was reasonable to assume that studies were estimating the same underlying treatment effect: i.e. where trials were examining the same intervention, and the trials’ populations and methods were judged sufficiently similar. Where there was clinical heterogeneity sufficient to expect that the underlying treatment effects differed between trials, or where we detected substantial statistical heterogeneity (if the T² was greater than zero and either the I² was greater than 30% or there was a low P value (less than 0.10) in the Chi² test for heterogeneity), we used random‐effects meta‐analysis to produce an overall summary, if an average treatment effect across trials was considered clinically meaningful. The random‐effects summary was treated as the average range of possible treatment effects and we discussed the clinical implications of treatment effects differing between trials. If the average treatment effect was not clinically meaningful, we did not combine trials.

For random‐effects analyses, the results are presented as the average treatment effect with 95% confidence intervals, and the estimates of T² and I².

Subgroup analysis and investigation of heterogeneity

Where we identified substantial heterogeneity, we investigated it using subgroup analyses and sensitivity analyses. We considered whether an overall summary was meaningful, and if it was, random‐effects analysis was used to produce it.    We planned to carry out the following subgroup analyses. We assessed different definitions for the primary outcome.

Primiparas versus multigravidas.
Catheter aspiration (abscess ≥ 3 cm) versus needle aspiration (abscess < 3 cm).
Women under 30 years of age versus those over 30 years of age.
Urban settings versus rural settings.
Co‐morbidities versus no co‐morbidities.
Exclusive breastfeeding versus mixed breast‐bottle feeding.
High‐income settings versus low‐income settings.

Due to the limited amount of data in the included studies, we were not able to perform any of the pre‐specified subgroup analysis. In future updates of this review, we will assess subgroup differences by interaction tests available within RevMan (RevMan 2014). We reported the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value.

Sensitivity analysis

We planned to perform sensitivity analysis on primary outcomes, to examine what effect excluding those studies at high risk of bias (for allocation concealment, and incomplete outcome data might have on the overall result of the meta‐analysis. However, since all of the included studies were of poor quality, we did not perform sensitivity analysis. We will carry out our planned sensitivity analysis in future updates of this review, if appropriate. In future updates, if we include cluster‐randomised trials in with the indivually randomised trials, we will also carry out sensitivity analysis to investigate the effect of the randomisation unit.

Results

Description of studies

Results of the search

The search of the Cochrane Pregnancy and Childbirth Group’s Register retrieved four trial reports.The search for trial reports on the Proquest dissertation and theses databases yielded 2005 studies, Google Scholar retrieved 2501 studies, African Journals online database retrieved 22 studies and the WHO ICTRP search retrieved 38 studies. Screening of reference lists yielded 30 extra studies, while contact with experts yielded no studies. Of the pharmaceutical companies we contacted, Pfizer responded with seven reports, which were unsuitable for inclusion as they did not fulfil the inclusion criteria for this review of RCTS and Novartis was unable to assist. After screening abstracts for eligibility, 15 full text articles of seemingly relevant studies were obtained. Of these, we included six published studies (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012) and excluded nine studies (Blick 1980; Edino 2001; Florey 1946; Ozseker 2008; Peters 1991; Sheih 2009; Strauss 2003; Tewari 2006; Wang 2013). For included studies, we contacted all of the corresponding authors (seeCharacteristics of included studies).

For a summary of the search results, see (Figure 1).

Included studies

Six studies met our inclusion criteria. Four of the studies included 325 women (Table 2) and contributed data to the analyses (Eryilmaz 2005; Saleem 2008; Singla 2002; Suthar 2012). As the remaining two studies included lactational and non‐lactational breast abscesses and the results for the outcomes were not recorded separately for each abscess type, the studies were not included in the quantitative analysis (Chandika 2012; Naeem 2012) but are described qualitatively.

1. Summary of characteristics of studies.

Study ID

Total number of participants

Intervention (n)

Comparison (N)

Country

*Primary outcomes reported

**Secondary outcomes reported

Eryilmaz 2005

Needle aspiration without ultrasound (n = 22)

I&D (n = 23)

Turkey

Saleem 2008

Percutaneous ultrasound‐guided drainage (n = 30)

I&D (n = 30)

Pakistan

Singla 2002

150

All women underwent (I&D) Antibiotics (Group 1 n = 50; Group 2 n = 50)

Control (n = 50)

India

Suthar 2012

Percutaneous ultrasound‐guided needle aspiration (n = 35)

Open surgical drainage (n = 35)

India

Open in a new tab

*Primary and **Secondary outcomes

Time to complete resolution.
Any continuation of breastfeeding after treatment.
Treatment failure .
Number of follow‐up visits.
Duration of continuation of breastfeeding after treatment.
Maternal satisfaction with treatment.
Post‐operative complications/morbidity.
Duration of hospital stay.
Adverse events.

I&D: incision and drainage

Study location

One study was based in Turkey (Eryilmaz 2005), a second in Pakistan (Saleem 2008) and two in India (Singla 2002; Suthar 2012). (Chandika 2012) was based in Uganda and (Naeem 2012) in Pakistan. All of the included studies were conducted within a hospital setting.

Types of intervention

The interventions included surgical (Chandika 2012; Eryilmaz 2005; Saleem 2008; Suthar 2012) as well as pharmacological interventions (Singla 2002). Only one study investigated two interventions against a control (Singla 2002) while each of the remaining five compared one intervention with another (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Suthar 2012).

Surgical interventions

Surgical interventions included incision and drainage (I&D) (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Suthar 2012), ultrasound‐guided needle aspiration/drainage (Saleem 2008; Suthar 2012;Chandika 2012; Naeem 2012), and needle aspiration without ultrasound (Eryilmaz 2005). Needle aspiration was compared with I&D in five studies (Eryilmaz 2005; Saleem 2008; Suthar 2012; Chandika 2012; Naeem 2012). All of the women in the studies randomised to I&D underwent general anaesthesia (Chandika 2012; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012). Eryilmaz 2005 reported that women in the I&D group received a local anaesthetic. Women in the ultrasound‐guided needle aspiration group (Chandika 2012 Saleem 2008; Suthar 2012), received a local anaesthetic prior to the intervention. Participants in Naeem 2012 did not receive any local anaesthetic. It was unclear whether the participants in Eryilmaz 2005 study received any anaesthetic.

Intervention with antibiotics

Singla 2002 investigated different treatment regimens of a broad spectrum antibiotic (cefazolin) compared with no antibiotics. All three groups of women in Singla 2002 underwent I&D where the objective of the study was to evaluate the role of antibiotics in the management of lactational breast abscesses. One group received intravenous cefazolin during the procedure, followed by oral cefazolin for six days; the second intervention group received a single dose of intravenous cefazolin before the procedure and the control group did not receive any antibiotics.

Participants

The total number of women in the four studies that contributed data was 325. Two studies included women with non‐lactational breast abscesses (Chandika 2012; Naeem 2012). In Chandika 2012 66% (43/65) of women and in Naeem 2012 83% (53/64) presented with lactational breast abscesses. The outcomes in both studies were not stratified according to abscess types and therefore the data were not included in the meta‐analysis.

Age and parity of women

Most women were between the ages of 20 and 30 years. One study did not report on the age of women (Singla 2002). Saleem 2008; Chandika 2012 and Naeem 2012 included primiparous and multiparous women. Singla 2002 and Suthar 2012 did not report on the parity of women in their studies.

Abscess size

Abscess sizes differed between studies. Chandika 2012 and Naeem 2012 excluded all women with breast abscesses that were greater than 5 cm. The median abscess size in Saleem 2008 was 5.5 cm (range 2 cm to 12 cm). The mean abscess size in Suthar 2012 was 4.9 cm ± 2.5 cm (range 1 cm to 15 cm). The mean abscess size in Eryilmaz 2005 was 6.5 ± 2.7 cm (I&D) and 6.1 ± 2.8 cm (needle aspiration group). Singla 2002 did not report on the abscess size of participating women.

Duration of symptoms

Duration of symptoms varied across studies.

Methods used to diagnose breast abscess

Preliminary diagnosis of an abscess was made based on clinical features of pain, swelling, and redness of the breast associated with localised tenderness in Suthar 2012. These women had the diagnosis and size of breast abscess confirmed by ultrasound evidence. Eryilmaz 2005 and Naeem 2012 made the diagnosis via a clinical examination and ultrasound was not used. Saleem 2008 made the diagnosis based on presence of a palpable mass or focal tenderness in the clinical setting of mastitis. Singla 2002 did not report how the diagnosis of a lactational breast abscess was made.

Outcomes

None of the studies separated outcomes into primary and secondary outcomes.

Five studies reported on time to resolution of breast abscess (Eryilmaz 2005; Saleem 2008; Suthar 2012; Chandika 2012; Naeem 2012), three studies reported on continuation of breastfeeding (Saleem 2008; Naeem 2012; Suthar 2012), and three reported on resolution of breast abscess (Chandika 2012; Eryilmaz 2005; Suthar 2012). Secondary outcomes were not uniformly reported on in all studies. Only two of the six secondary outcomes were addressed. Two studies reported on maternal satisfaction with treatment (Eryilmaz 2005; Saleem 2008). Four studies reported on post‐operative complications (Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012).

Definitions employed by authors

Chandika 2012 defined breast resolution as absence of symptoms of inflammation and absence of fluid on sonar.

Naeem 2012 considered resolution of symptoms.

Saleem 2008 defined resolution as no recurrent abscess or need for surgery.

Suthar 2012 defined resolution in the needle aspiration group as absence of symptoms after four aspirations with no evidence of liquefaction using ultrasound. The definitions of resolution for the I&D group were unclear.

Eryilmaz 2005 defined time to healing as time from I&D to closure for the I&D group, and time until complete resolution of the mass up to a maximum of five aspirations for the needle aspiration group.

For this review, we considered healing time to be the same as time to resolution of abscess.

Excluded studies

Nine studies were excluded from the review (Blick 1980; Edino 2001; Florey 1946; Ozseker 2008; Peters 1991; Sheih 2009; Strauss 2003; Tewari 2006; Wang 2013). Reasons for study exclusions were non randomised controlled studies, non probability sampling, absence of comparator, case series and inclusion of non‐lactational breast abscesses. Excluded studies are summarised in the table of Characteristics of excluded studies.

Risk of bias in included studies

The risk of bias for each included study is presented in the 'Risk of bias’ tables in the Characteristics of included studies. Figure 2 and Figure 3 illustrate the summary of risk of bias in all the studies. Across studies there was unclear risk of bias for most domains due to poor reporting.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation was assessed as adequate in one study Chandika 2012, where a table of random numbers was generated via a computer (Microsoft excel version 5.0).

Random sequence generation was unclear in the remaining five studies as the authors did not adequately report on methods used to generate a random sequence (Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012).

Allocation concealment was unclear in all six studies as methods for performing allocation concealment were not described (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012).

Blinding

The nature of the intervention would have rendered it difficult to blind personnel, participants and outcome assessors involved in these studies. If data analysts were used and were independent of the research team, this was not made clear in the study and we therefore judged all studies as having an unclear risk for performance and detection bias. (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012).

Incomplete outcome data

Four studies were judged as having an unclear risk of attrition bias (Chandika 2012; Eryilmaz 2005; Singla 2002; Suthar 2012). In Chandika 2012 four participants were lost to follow‐up in the needle aspiration group, while in the I&D group one was lost to follow‐up. It is unclear what the outcomes for these abscesses were as this was not reported. In Eryilmaz 2005, nine participants in the needle aspiration group were excluded from the analysis. Their healing times were not included in the results as it was not known how long these abscesses took to heal. The study reported the healing rate as 41% (13/22), which is incorrect. In Singla 2002, all results were given as percentages and therefore made it difficult to comment on the attrition rate. In Suthar 2012, the authors indicated that the lactational breast abscesses for 29 women had resolved and six women were excluded from the study. It is unclear what happened to the six women.

Naeem 2012 was judged as having a high risk of attrition due to the fact that there were missing participants in both groups. In addition, correspondence by the author revealed that an additional three participants were missing. This information was not reported in the study. In Saleem 2008, risk of incomplete outcome data was judged high due to exclusions of participants. The authors stated that four abscesses perforated before treatment and three women underwent surgery because the abscesses were not suited for ultrasound. It is unclear to which groups these women belonged and whether they were included in the analysis. Nine women who had mastitis did not have breast abscesses and it is not known whether they were exposed to any intervention and included in the analysis.

Selective reporting

Four of the studies reported adequately on specified outcomes (Chandika 2012; Eryilmaz 2005; Naeem 2012; Singla 2002). Saleem 2008 prespecified resolution and complication rates but reported on other outcomes as well and we were therefore uncertain if these fell under complication and resolution rate. We therefore judged the study as having an unclear risk of bias. Suthar 2012 was judged as having unclear risk of bias as none of the outcomes were prespecified in the methods section.

Other potential sources of bias

We judged all six studies as having an unclear risk of other potential sources of bias (Chandika 2012; Eryilmaz 2005; Naeem 2012; Saleem 2008; Singla 2002; Suthar 2012). In Chandika 2012, women who were resistant to cloxacillin were removed from the study after randomisation and the authors have not described how many were resistant to cloxacillin and what happened to these women with regards to the abscess. In Naeem 2012, the authors reported that skin indurations around abscesses were present in 93.75% of abscesses in the I&D group, whereas 71.8% of women in the needle aspiration group had indurations. The authors report that this baseline difference was significant, yet the reported P value was 0.20, which is not statistically significant. We contacted the authors regarding this inconsistency but no response was received. In Saleem 2008, a table comparing baseline characteristics between groups was not available, which made it difficult to judge whether the groups were similar at the start of the study. In addition, groups were not treated similarly whereby in the ultrasound‐drainage group women had an ultrasound to confirm diagnosis and resolution but it is unclear what was done for the I&D group to confirm diagnosis and resolution. Also, the authors reported that women who had abscesses greater than 5 cm had a catheter inserted, but they did not report how many women had a catheter inserted. In Singla 2002, the authors reported most of the results as percentages and absolute numbers were not provided. The report also did not contain a table of participant characteristics to judge whether these groups were similar or not.

In Suthar 2012, a table of baseline characteristics was absent and it was therefore difficult to judge whether women in both groups were similar at the start of the study.

Effects of interventions

See: Table 1

1. Needle aspiration versus incision and drainage I&D)

Three studies (n = 160) were included under this comparison (Eryilmaz 2005; Saleem 2008; Suthar 2012). In Eryilmaz 2005, needle aspiration was done without ultrasound, while Saleem 2008 and Suthar 2012 both used ultrasound guidance for needle aspiration. In light of heterogeneity for all three outcomes, we did not pool data and the overall effected was not reported. Results have been summarised in the Table 1.

Primary outcomes

1.1 Time to complete resolution of breast abscess

Three studies reported on the mean time to complete resolution of breast abscess (Eryilmaz 2005; Saleem 2008; Suthar 2012) ‐ see Analysis 1.1. Eryilmaz 2005 and Suthar 2012 excluded women who had treatment failure when they calculated the mean time to complete resolution.

1.1 — Comparison 1 Needle aspiration versus incision and drainage (ultrasound guided versus no ultrasound guided), Outcome 1 Time to resolution of breast abscess (days).

Eryilmaz 2005 found that the time to complete resolution of breast abscess was significantly less in the needle aspiration group compared to the I&D group (mean difference (MD) ‐6.07; 95% confidence interval (CI) ‐7.81 to ‐4.33; n = 36), but excluded 9/22 (41%) women in the needle aspiration group due to treatment failure.

Suthar 2012 found a significant reduction in time to complete resolution in the needle aspiration group (MD ‐17.80; 95% CI ‐21.27 to ‐14.33; n = 64), but excluded 6/35 (17%) women in the needle aspiration group due to treatment failure.

Saleem 2008 also found a significant reduction in time to complete resolution of breast abscess in the needle aspiration group (MD ‐16.00; 95% CI ‐18.73 to ‐13.27; n = 60), but did not indicate the number of women who were lost to follow‐up for either group and it is therefore not known on how many women the calculation of average time to resolution of breast abscess was based on.

Taking into consideration the limitations of the available data of all three studies, we do not consider the results to be informative.

1.2 Any continuation of breastfeeding after treatment (success)

Two studies (n = 130) reported on this outcome (Saleem 2008; Suthar 2012). We did not pool the data, since there were high levels of unexplained heterogeneity in the random‐effects meta‐analysis (Tau² = 1.38; Chi² = 32.88: P < 0.00001; I² = 97%) (Analysis 1.2).

1.2 — Comparison 1 Needle aspiration versus incision and drainage (ultrasound guided versus no ultrasound guided), Outcome 2 Continuation of breastfeeding.

In Saleem 2008, women in the needle aspiration group were more likely to continue breastfeeding (risk ratio (RR) 2.89; 95% CI 1.64 to 5.08; n = 60). In Suthar 2012, continuation of breastfeeding showed a trend towards needle aspiration, however this was not statistically significant (RR 1.09; 95% CI 0.97 to 1.22 n = 70).

1.3 Treatment failure

Two studies (n = 115) reported on treatment failure (Eryilmaz 2005; Suthar 2012). In Eryilmaz 2005, treatment with needle aspiration failed in 9/22 women who proceeded to have I&D. All abscesses in the I&D group were successfully treated. In Suthar 2012, treatment with needle aspiration failed in 6/35 women, who then underwent I&D. Treatment failure rate was high among women who were treated with needle aspiration (RR 16.12; 95% CI 2.21 to 117.73; participants = 115; studies = two) Analysis 1.3. We graded this evidence as low quality (see Table 1).

1.3 — Comparison 1 Needle aspiration versus incision and drainage (ultrasound guided versus no ultrasound guided), Outcome 3 Treatment failure.

Secondary outcomes

Secondary outcomes were poorly reported in all studies and only limited data were available to include in the analysis.

1.4 Number of follow‐up visits

An assessment of the number of follow‐up visits would have provided information on the recovery following the intervention. However, no data were available. Post intervention follow‐up visits were not specified in Eryilmaz 2005, Singla 2002, and Suthar 2012. Saleem 2008 followed up women for up to two months in the needle aspiration group, at weeks four and eight after the procedure.

1.5 Duration of continuation of breastfeeding after treatment

None of the studies reported on the duration of continuation of breastfeeding.

1.6 Maternal satisfaction with treatment

Singla 2002 and Suthar 2012 did not discuss maternal satisfaction with the procedure. Eryilmaz 2005 reported that 16/23 (70%) of women in the I&D group were satisfied with the outcome but did not report on the needle aspiration group. Saleem 2008 indicated that there was 100% (30/30) satisfaction with treatment in the percutaneous ultrasound group whereas in the I&D group only 17/30 (55%) women were satisfied with the procedure (RR 1.74; 95% CI 1.28 to 2.38) (Analysis 1.4).

1.4 — Comparison 1 Needle aspiration versus incision and drainage (ultrasound guided versus no ultrasound guided), Outcome 4 Maternal satisfaction with treatment.

1.7 Post‐operative complications/morbidity

In Saleem 2008, one woman developed a milk fistula, 1/30 (3%) woman had a residual abscess in the needle aspiration group, and 5/30 (16%) women developed milk fistulas in the I&D group. In Suthar 2012, 20% (14/70) of women complained of intolerable pain in the needle aspiration group and one woman developed a milk fistula in the I&D group.

1.8 Duration of hospital stay

Duration of hospital stay was not reported in Suthar 2012. Eryilmaz 2005 treated both groups of women on a outpatient basis. Saleem 2008, described having admitted (4/30) 13% of women in the ultrasound group, but did not state for how long. The rest of this group were treated as outpatients. The women in the I&D group were admitted for a mean of four days (two to eight days).

1.9 Adverse events

An adverse event for this review was considered in the context of events arising from drugs that may have been prescribed for women during the interventions and complications associated with the prescription thereof. None of the studies reported on adverse events. Complications arising from the procedure itself were described under post‐operative complications and morbidity.

2. Incision and drainage (I&D): antibiotic use versus no antibiotic use

One study (Singla 2002) involving 150 women, compared two different antibiotic regimens to no antibiotic administration. All three groups of women underwent I&D. Two of the groups were given antibiotics and were compared with a similar group of women who were not given any antibiotics. Group A (n = 50) received cefazolin 1 g intravenously (IV) at the time of induction of anaesthesia and 500 mg eight hourly IV for 24 hours. This was followed by oral cefalexin 500 mg six hourly for six days. Group B (n = 50) received a single dose of cefazolin 1 g IV 30 minutes before surgery. Group C (n = 50) did not receive any antibiotics.

Primary outcomes

2.1 Time to resolution of breast abscess

The mean time to resolution of breast abscess was similar in all groups, although women with an infection were excluded. Mean time to resolution for women who received a course of antibiotics was 7.3 days, 6.9 days for women who received a single dose of antibiotics, and 7.4 days for women who did not receive antibiotics. Standard deviations (SDs), P values and confidence intervals (CIs) were not reported and prevented further analysis.