Factors related to breast abscess after delivery: a population-based study


L. J. Kvist, BFK-Department of Obstetrics and Gynaecology, Helsingborg Hospital, SE-251 87 Helsingborg, Sweden.


Objective  To investigate whether there are underlying factors, other than breastfeeding behaviours, which may contribute to the development of breast abscess during the year following delivery.

Design  A population-based study.

Setting  In Sweden.

Population  The 1,454,068 singleton deliveries during 1987–2000.

Methods  Data retrieval from two national patient registers: the Medical Birth Registry and the National Discharge Register. Stratified Mantel–Haenszel analysis.

Main outcome measures  Odds ratios with 95% confidence intervals for possible risk factors for breast abscess in the year following delivery.

Results  During the year following delivery 1401 women had surgery because of a breast abscess. This gives a rate of 0.1%. Sixty-five percent of cases occurred between three and eight weeks postpartum. An annual increase in the odds ratio (OR) was found between 1993 and 1999. Lowest risk was found among mothers who were ≤24 years and a significantly increased risk among mothers ≥30 years. There was a 3.6-fold increased risk for breast abscess associated with primiparity and OR for post-maturity (>41 weeks) was 5.

Conclusion  Primiparous women appear to be at a greater risk for the development of breast abscess during lactation than multiparous women. Mothers over the age of 30 years and those who give birth post-maturely are also at a significantly increased risk. Primiparous women and women over 30 years could be targeted for extra information in preparation for parenthood classes about how to avoid over distension of the breasts during breastfeeding. Further research is needed to understand the role of post-maturity and to investigate why some women recover spontaneously from inflammatory processes of the breast.


The two main objectives of treatment for inflammatory processes of the breast during lactation are to facilitate recovery in order to allow continued breastfeeding and to prevent the occurrence of breast abscess. Studies of inflammatory processes of the breast during lactation show an estimated incidence of between 2% and 33%.1–7 Some researchers have reported that delayed treatment of mastitis by antibiotic therapy is the major risk factor for development of breast abscess.8,9 Other researchers suggest that this may not necessarily be the case2,6,10,11 because mastitis may be a self-limiting disease in many cases. During the 1980s a Norwegian study found that bacterial counts had no ability to discriminate between patients who had short-lived or protracted symptoms10 and that subsequent medication with phenoxymethylpenicillin did not prevent prolonged illness or abscess formation. Marchant9 writes that the distinction between infective and non-infective forms of mastitis has no practical use because the treatment is the same for both groups. However, he questions the use of antibiotics without attempts to correct the underlying cause of the disease.

The aim of the present study was to identify factors, other than current breastfeeding practices, that may contribute to the development of breast abscess. Identification of these factors may increase our understanding of the aetiology of the disease and suggest more preventative measures for specific groups of women.


Information on pregnancy and delivery was collected from the Medical Birth Registry (MBR), which is held by the National Board of Health and Welfare (NBHW), Stockholm, Sweden. The MBR holds records of all deliveries, including stillbirths, with a gestational duration of at least 28 weeks. When compared with the official statistics from the country's Central Bureau of Statistics, less than 1% of all births are missing from the MBR.12

In a study of the quality of the MBR, Cnattingius et al.13 examined a 0.5% random sample of deliveries from 1974 to 1986. Copies of the original medical records were requested from the hospitals and compared with data held at the MBR. Compared with the original medical records, the data regarding gestational duration held at the MBR were acceptable. A new evaluation of the MBR has recently been carried out.12

Data on year of delivery, maternal age, parity, smoking, years of education, gestational duration, birthweight, mode of delivery and body mass index (BMI) were collected from the MBR. All singleton pregnancies with delivery between 1987 and 2000 were included. The identification of cases, that is, mothers operated upon for breast abscess, was performed by using information stored in the National Discharge Register (NDR), also held by the NBHW, Stockholm. Identification of the women who had breast surgery within one year after delivery was made possible by the unique personal number allocated to each person who is permanently resident in Sweden.

The cases were identified in the NDR by the use of the NBHW code for operative incision and drainage (3800) with or without a diagnosis of inflammatory disease in the breast, infection in the breast related to delivery or breast engorgement. The codes for these are, respectively, 611.0, 675 and 676.2 according to the International Classification of diseases (ICD9) before 1997. The corresponding codes for the International Statistical Classification of Diseases and Related Health Problems (ICD10) after 1997 were HAA00, HAA01 and O91 or N61, respectively. If the case was not identified in the MBR, which would mean that the operation had not occurred in conjunction with a recent delivery (<1 year), the case was excluded from the analysis.

Odds ratios (OR) were calculated with 95% confidence intervals according to Miettinen.14 To control for confounding factors such as parity, maternal age and year of delivery, a stratified analysis was performed using the Mantel–Haenszel technique.15 In each and every analysis all factors were stratified for each other, except for educational level. The reason for this was that statistics for educational level were available only up to 1995, causing a large proportion of the sample statistics for education to be unknown. Statistics for BMI were available between the years 1992 and 2000.

The committee for medical research ethics in Lund, Sweden, gave permission for the study.


During the study period, the number of singleton deliveries in Sweden was 1,454,068. A total of 1401 women had surgery because of a breast abscess related to a recent lactation. This gives a rate of 0.1%. The OR for breast abscess seemed to be significantly reduced before 1991 and significantly increased after 1992 (Table 1).Figure 1 shows the distribution of cases of abscess by week postpartum. Sixty-five percent of cases occurred between three and eight weeks postpartum. A clear effect of maternal age could be seen, with least risk among mothers who were ≤24 years and a significantly increased risk among mothers ≥30 years (Table 2). Primiparity entailed a 3.6 increased risk for breast abscess compared with multiparity (Table 2).

Table 1.  OR for breast abscess related to year of delivery and stratified for maternal age, parity, smoking, gestational age, birthweight, mode of delivery and BMI. Each year is compared with all other years.
YearNo. of singleton deliveriesNo. of breast incision and drainage operationsOR95% CI
Figure 1.

Number of cases of breast abscess by week postpartum.

Table 2.  OR for breast abscess related to maternal age and parity stratified for year of delivery, smoking, gestational age, birthweight, mode of delivery and BMI. Maternal age is also stratified for parity and parity is stratified for maternal age.
 OR95% CI
Maternal age
Para 13.63.2–4.0
Para 20.50.4–0.5
Para 30.40.4–0.5
Para 4+0.30.2–0.4

For gestational age, there was significantly increased risk among mothers whose babies were post-mature, >41 weeks (Table 3). Mothers of infants weighing between 1500 and 2499 g were significantly less likely to suffer from breast abscess (Table 3). A significantly decreased OR for breast abscess was seen in mothers with a BMI of ≥30 (OR 0.6, 95% CI 0.4–1.0).

Table 3.  OR for breast abscess related to gestational age and birthweight, stratified for year of delivery, maternal age, parity, smoking, mode of delivery and BMI. Gestational age is also stratified for birthweight and birthweight for gestational age.
 OR95% CI
Gestational age (weeks)
Birthweight (g)

To further elucidate the relationship between primiparity, post-maturity and maternal age ≥30 years, the ORs for these groups of mothers were calculated (Table 4). The overall OR was 6.6. For primiparity and maternal age ≥30 years, OR was 5.7, indicating that this combination is important. Primiparity exclusively had an OR of 3.5.

Table 4.  OR for breast abscess related to the three identified risk groups: primiparity, maternal age ≥30 years and post-maturity, stratified for each other.
 OR95% CI
Primiparity + maternal age ≥ 30 years + post-maturity6.64.9–8.8
Primiparity + maternal age ≥ 30 years5.74.8–6.7
Primiparity + post-maturity4.33.5–5.3
Primiparity exclusively3.53.1–3.9
Post-maturity + maternal age ≥ 30 years2.11.6–2.7
Maternal age ≥ 30 years exclusively1.61.5–1.8
Post-maturity exclusively1.21.0–1.5

The mothers' educational level had no effect on the OR for development of breast abscess. No increased risk was seen among smokers. The mode of delivery did not influence the incidence; caesarean section versus vaginal delivery (OR 1.1, 95% CI 0.9–1.3) and instrumental vaginal delivery versus non-instrumental vaginal delivery (OR 1.1, 95% CI 0.9–1.3).


In this material, first time mothers appear to be those at greatest risk for the development of breast abscess. Even mothers over 30 years of age and those whose babies are born post-maturely are at an increased risk. Evans and Head16 reported an increased risk for lactation mastitis in first-time mothers although others found that parity was not related.3

The fact that the majority of cases of abscess occurred during three to eight weeks postpartum could imply that the problem started soon after delivery. The bacterial content of breast milk has been examined earlier but although S. aureus has been implicated, researchers have been unable to reach consensus on which concentration of bacteria in breast milk requires antibiotic therapy. Matheson et al. concluded that phenoxymethylpenicillin failed to stop abscess formation in one out of five patients8 and that the majority of mothers suffering from various levels of mastitis recovered without antibiotics. Because the rate of abscess formation in our study was 0.1% the question arises as to whether antibiotic treatment is appropriate for all mothers (2–33%) with inflammatory symptoms of the breast. There is no consensus in the literature on this subject. A Swedish study17 from 2000 suggested that 61% of the mothers with mastitis should have been treated by antibiotic therapy because of potentially pathogenic bacteria in their milk. However, a more recent Swedish study18 on care of women with breast inflammation showed a use of antibiotics for only 9% of participants. Fetherston19 reported from Australia that antibiotics were prescribed for 85% of mothers with inflammatory symptoms. These differences are so large and the need to avoid over-prescription of antibiotics so pressing that more research is needed before recommendations on optimal treatment can be safely given. Because there is a natural portal in the nipple for invading pathogens, the incidence rate for breast abscess of 0.1% could be interpreted as surprisingly low. The question to be asked now might be ‘what is different about the few mothers who succumb to the invasion of pathogens’?

It has been demonstrated earlier that stress to the mother and fetus during labour and delivery are risk factors for delayed lactogenesis.20 We hypothesised that mother and infant dyads with a sub-optimal start to breastfeeding may have increased lactation problems later on. This could have explained why mothers of post-mature babies had an increased risk for abscess formation. However, somewhat surprisingly to us, the mode of delivery (which also would have implied a sub-optimal breastfeeding start) had no effect on the OR for breast abscess and thus the hypothesis may have to be rejected. It is possible that the release of oxytocin as a result of suckling may be somewhat slower in those mothers whose labour may have been induced with artificial oxytocin, pre-disposing them to engorged breasts, subsequent inflammation and abscess formation. It was suggested in 1994 that oxytocin administration during labour might be implicated in breastfeeding problems.21

What roles then do increased maternal age and primiparity play? The delaying of pregnancy may play a part in lactation difficulties. One may only speculate that nature's intention is that a first lactation should occur much earlier than 30 years plus. Breastfeeding is a time-consuming activity and one that the infant dictates. The advent of a first child always gives cause for the restructuring of ones lifestyle. It is possible that younger women are less set in their ways and therefore more receptive to the babies demands to be fed at varying intervals and for varying lengths of time. More information about the avoidance of inflammatory breast disease during lactation by the use of good breastfeeding practices could be given at preparation for parenthood classes.

Although Sweden is a country with a good track record for the initiation of various medical registers, the maintenance of these registers requires continuous vigilance. The human error factor means that we cannot be certain that the true incidence of breast abscess as a result of lactation in this population is 0.1%. For curiosity we checked on the incidence of breast abscess at our own unit, which has 3000 deliveries annually. For 2003 the incidence was 0.1% (unpublished information). This may suggest that our material gives a generally true picture of the situation in Sweden. The apparent increase in the number of cases between 1993 and 1999 may simply be a result of better registration during those years and may reflect an increase in the number of women continuing to breastfeed for longer periods.

Our study showed a protective effect of giving birth to an immature baby. There is great awareness on Neonatal Intensive Care Units (NICU) in Sweden of the advantages of breast milk for immature infants and therefore all encouragement and support is given to mothers to allow continued breastfeeding (unpublished). The fact that these mothers and infants spend much longer time in contact with care professionals may increase the mothers' knowledge and understanding of breastfeeding practices that avert the occurrence of inflammatory processes that may lead to abscess formation. Hand hygiene is of paramount importance on the NICU and lessons learned there might also be a protective factor against invading bacteria. It seems clear that the infants weighing 1500–2499 g are the same population as these prematurely born babies. The fact that mode of delivery did not entail increased risk of breast abscess may in part be explained by the fact that infants born by caesarean section are often admitted to the NICU.

Three areas for further research have become apparent during this study. We need to know more about the possible effects of administration of oxytocin during labour on breastfeeding. As electronic documentation systems develop, it will be possible to check for correlation between oxytocin and breastfeeding problems. Our understanding of the reasons why primiparous women and those over 30 years of age are more vulnerable to breast abscess is important and qualitative studies using, for example, a narrative approach might reveal the breastfeeding practices and overall life situation of these mothers. We need also to further investigate the differences in bacteriological and immunoglobulin content of breast milk in mothers who recover spontaneously from inflammatory processes of the breast and mothers who develop breast abscess.


Primiparous women, mothers over 30 years of age and those giving birth post-maturely may be more likely to develop breast abscess during lactation than other groups. Primiparous and older women could benefit from extra information in preparation for parenthood classes about the need for lifestyle adjustment in order to follow the dictates of nature and avoid over distension of the lactiferous sinuses. Increased awareness of the importance of hand hygiene could be beneficial to new mothers. It may also be helpful, when problems have already occurred, to give extra guidance about self-care to avoid recurrences of inflammatory symptoms of the breast.

Accepted 13 February 2005