Br. J. Surg. 1992, Vol. 79, January, 58-59
N. J. Bundred, M. S. Dover, S. Coley and J. M. Morrison Department of Surgery, Selly Oak Hospital, Birmingham, UK Correspondence to: M r N . J. Bundred, University Hospital of South Manchester, Nell Lane, Manchester M20 8LR.
UK
Breast abscesses and cigarette smoking An I I-year retrospective review of women with breast abscesses presenting to a district general hospital was performed. A total of 122 women were identiJied with a breast abscess: 8.5 (70 per cent) with non-lactational abscesses and 37 (30 per cent) with an abscess in the puerperium. Six of37 lactational and 24 of 8.5 non-lactational abscesses recurred. Sixteen mammillary fistulae developed, all following incision and drainage of non-lactational breast abscesses. Women with non-lactational breast abscesses were more likely to smoke cigarettes (P < 0.00.5). Breast abscesses containing anaerobic bacteria were signlJicantly more likely to occur in current cigarette smokers ( P < 0.05). Women with non-lactational breast abscesses who were heavy cigarette smokers were more likely to sufer recurrent abscesses ( P < 0.01). There was histological evidence of duct ectasia on biopsy in 25 women with non-lactational abscesses and 21 of 25 were current smokers. Mammillary ,fistulae developed more frequently in current smokers ( P < 0.03). Cigarette smoking is important in the natural history of non-lactational breast abscesses and may predispose to anaerobic breast injections and the development of mammillaryfistulae.
The incidence of lactational breast abscesses is declining but non-lactational breast abscesses are seen more frequently’,’. Non-lactational breast abscesses often recur and may be associated with mammillary fistulae or mammary duct ectasia’,3.4. This study examined the factors which might predispose women with non-lactational breast abscesses to develop further sepsis.
Patients and methods The case records of all women presenting with a diagnosis of a breast abscess over an 11-year period were reviewed ( 1 January 1980-31 December 1990). Breast abscess was defined as an acute inflammatory lump which yielded pus on incision. Details were recorded of the clinical features, smoking habits, surgical treatment and eventual outcome in all women. From 1980 the Nursing Kardex and Selly Oak Hospital Medical Admissions Chart required the patient’s smoking habits to be recorded. A patient was recorded as smoking currently and the number of cigarettes smoked per day noted. Light smokers were defined as women smoking 1 10 cigarettes/day and heavy smokers as women smoking 11 or more cigarettes per day. Women were recorded as ‘never smoked’ if at admission they did not smoke unless the case notes recorded that they had previously done so. When a breast biopsy had been taken histological details were recorded. Bacteriological details were noted for all patients. Histological specimens were reviewed to confirm the presence of dilated ducts and for acute/chronic inflammation surrounding the ducts5. ~
Nineteen women were non-smokers, eight women previously smoked and ten were currently cigarette smokers. Staphylococcus auveus was isolated from 27 patients, anaerobic bacteria from four, coliform bacteria from one woman and no growth was obtained from five women. Six women developed recurrence of their abscess, four who had never smoked, one who was a previous smoker and one who was currently smoking 20 cigarettes per day. Biopsy of the cavity wall from ten women (including all six with recurrent abscesses) failed to show any evidence of mammary duct ectasia. Non-lactational bveust ubscesses Eighty-five women (70 per cent j (mean age 35.8 (range 11 -69) years ) had non-lactational breast abscesses incised and drained. Twenty-eight (33 per cent) had never smoked, eight had previously smoked and 49 (58 per cent) were currently doing so regularly (Table 1 ). Smoking habits were not age-related (Table 2). Women with non-lactational breast abscesses were more likely to be current smokers (x’ = 9.6, P < 0.005) and more likely to have ever smoked (relative risk 2.15, 95 per cent confidence interval 2.04-2.26 ) than women with lactational breast abscesses. The bacteria isolated from the initial abscess are listed in Tuble 3. Fifteen of 18 women who developed anaerobic breast Table 1 Smokiny lzcrhirs q/\c.onwn with non-lacrarionul hrcvst ah.rc.esses
Results A total of 122 women presented with a breast abscess over the 11-year period. Several patients had experienced more than one episode of abscess formation but for the purpose of the study only the first presentation was considered. Lactutional breart abscesses
n ~~
-~
28 8
~ _ _ _ _ _ -
0007 1323’92/010058 02
Recurrent abscess ~~~~~
Never smoked smokers Current smokers Light ( < IO/day) Heavy ( > IO/day)
~~
58
~~~
(
Mammillary fistulae ~
3 1
2 1
13
4
3
36
16
10
~
_______
1992 Butterworth Helnerndnn Ltd
Breast abscesses and cigarette smoking:
Table 2 Frequency of smoking in women with non-lacrational breast abscrsses
Age (years )
Total
Current smokers
Previous smokers
15-24 25-34 35-44 45-54 55-64 > 65 Total
19 19 25 13 3 6 85
10 13 16 7 2 1 49
1 2 3 2 0 0 8
Never smoked 8 4 6 4 1
5 28
Table 3 The relationship between bacterial isolates and duct ecrasia in non-lactational abscesses
Staphylococcus aureus Anaerobic bacteria No growth Coliform bacteria Total
n
Recurrent abscess
Abscess biopsied ( n = 32); duct ectasia confirmed histologically
31 18 32 4 85
6 9* 7 2 24
2 9t 13t 1 25
*The presence of anaerobic bacteria in the initial abscess is associated with recurrence, x 2 = 4.1, P < 0.05; t a biopsy showing duct ectasia was associated with a current or previous abscess containing bacteria other than Staphylococcus aureus, x 2 = 9.8, P < 0.005
abscesses were current cigarette smokers compared with 34 of the 67 women who had other organisms isolated or showed no bacterial growth ( x 2 test with Yates' correction = 4.9, P < 0.05). Twenty-four women developed further breast abscesses following incision and drainage and 16 went on to develop mammillary fistulae. Heavy smokers were significantly more likely to develop further breast abscesses (Armitage test for trend, xz = 7.10, P < 0.01). Women with anaerobic breast abscesses were more likely to develop recurrent breast abscesses ( x 2 test with Yates' correction = 4.1, P < 0.05; Table 3 ) . Mammillary fistulae developed from non-lactational breast abscesses significantly more often in current smokers ( P < 0.03, Table 1 ). Thirty-two women underwent biopsy either at presentation or on recurrence of their abscess and review of the histology revealed 25 women had mammary duct ectasia. The majority of the biopsies were taken when the patients returned with a recurrent abscess or fistula but three women had a biopsy at first presentation which was diagnostic of duct ectasia. Among the 25 women with a histological diagnosis of duct ectasia, 21 were current cigarette smokers, one woman had previously smoked and three had never done so. A bacterial isolate other than Staphylococcus aurcus from the initial abscess was significantly more likely to be associated with a later histological confirmation of duct ectasia ( xz = 9.8, P < 0.005 ; Table 3 ).
Discussion Despite the reported fall in the incidence of lactational breast abscess elsewhere'.2 they accounted for one-third of the abscesses in this series. They were usually associated with Staphylococcus aureus and did not tend to recur. When lactational abscesses did recur this could not be related to smoking habit. Non-lactational breast abscesses tend to recur and three factors have been associated with recurrence ;subareolar site'.3, anaerobic bacteria in the abscess 1.3 and the presence of underlying duct ectasia'.3.
Br. J. Surg., Vol. 79, No. 1, January 1992
N. J.
Bundred st al.
There was no difference between peripheral and subareolar abscesses and although breast abscesses containing anaerobic bacteria were more likely to recur, it was not the most significant factor. The association of histological duct ectasia with recurrence may be because recurrent abscesses tend to be subject to repeated biopsy and this leads to the possibility of an ascertainment bias. It is likely that duct ectasia is responsible for the majority of these abscesses even when histological confirmation is not ~ b t a i n e d ~ . ~ . The aetiology of subareolar sepsis remains unknown. Cigarette smoking is reported to reduce the incidence of benign breast by lowering plasma and urinary oestrogen levels7. The Office of Population Censuses and Surveys reported that in 1988, 17-39 per cent of women were current smokers, the variance being related to social class and age'. A recent survey in the Selly Oak Hospital District has shown that approximately 30 per cent of the female population currently smoke and 10 per cent have previously done so1'. Regardless of age 67 per cent of women with non-lactational breast abscesses and 84 per cent of women with a histological diagnosis of duct ectasia currently smoke or have previously done so. Smoking may be an aetiological agent in this condition. This is supported by the higher recurrence rate of non-lactational breast abscesses in women who are heavy smokers and the association of smoking with anaerobic breast abscess. The recent increase in cigarette smoking over the last decade in young women aged 20-40 years may account for the apparent increase in the incidence of non-lactational breast abscesses in this period','.9. The mechanism is not clear but smoking leads to the production of toxins in ductal secretions and this may damage the lactiferous ducts".". The aetiology of mammillary fistula is ~ n k n o w n ' Nipple ~. inversion and squamous metaplasia of the nipple may be contributory but it is likely that such fistulae are complications of duct ectasiaI4. Only seven of the 16 women who developed mammillary fistulae in this series had nipple inversion, in keeping with previous reports4.l4. Women who were current smokers were significantly more likely to develop mammillary fistulae. The high frequency of sterile cultures from the non-lactational breast abscesses and the retrospective nature of this report require prospective studies of this phenomenon.
References 1. 3
i .
3. 4.
5. 6. 7. 8. 9. 10. 11.
12. 13. 14.
Scholefield JH, Duncan JL, Rogers K. Review of a hospital experience of breast abscess. Br J Sury 1987; 74: 469-70. Bates T, Down RHL, Tant DR, Fiddian RV. The current treatment of breast abscesses in hospital and in general practice. Practitioner 1973; 211: 541-7. Ekland DA, Zeigler MG. Abscess in the nonlactating breast. Arch Surg 1973; 107: 398-401. Lambert ME, Betts CD, Sellwood RA. Mammillary fistulae. Br J Surg 1986; 73: 367-8. Dixon J M , Anderson TJ, Lumsden AB et al. Mammary duct ectasia. Br J Surg 1983; 70: 601-3. Bundred NJ, Dixon JM, Lumsden AB et ul. Are the lesions of duct ectasia sterile. Br J Sury 1985 ; 72 : 844-5. Baron JA. Smoking and oestrogen related disease. Am J Epidemiol 1984; 119: 9-22. Berkowitz GS, Canny P F , Vivolsi VA et al. Cigarette smoking and benign breast disease. J Epidemiol Community V - d f h 1985 : 39: 308-13. Office of Population Censuses and Surveys England and Wales. London: HMSO, 1988. South Birmingham Health Authority. Annual Public Health Report 19YU. Parker: Canterbury, 1990. Wynder EL, Hill P. Nicotine and coninine in breast fluid. Cancer Letr 1979; 6 : 251-4. Petrakis NL, Dupuy ME, Lee RE. Mutagens in nipple aspirate of breast fluid. Bumbury Report 1985 ; 13: 67-82. Editorial. Mammillary fistulae. Lancet 1986; ii: 438-9. Bundred NJ, Dixon JM, Chetty U, Forrest APM. Mammillary fistula. Br J Sury 1987; 74: 844-5.
Paper accepted 28 September 1991
59
N. J. Bundred, M. S. Dover, S. Coley and J. M. Morrison Department of Surgery, Selly Oak Hospital, Birmingham, UK Correspondence to: M r N . J. Bundred, University Hospital of South Manchester, Nell Lane, Manchester M20 8LR.
UK
Breast abscesses and cigarette smoking An I I-year retrospective review of women with breast abscesses presenting to a district general hospital was performed. A total of 122 women were identiJied with a breast abscess: 8.5 (70 per cent) with non-lactational abscesses and 37 (30 per cent) with an abscess in the puerperium. Six of37 lactational and 24 of 8.5 non-lactational abscesses recurred. Sixteen mammillary fistulae developed, all following incision and drainage of non-lactational breast abscesses. Women with non-lactational breast abscesses were more likely to smoke cigarettes (P < 0.00.5). Breast abscesses containing anaerobic bacteria were signlJicantly more likely to occur in current cigarette smokers ( P < 0.05). Women with non-lactational breast abscesses who were heavy cigarette smokers were more likely to sufer recurrent abscesses ( P < 0.01). There was histological evidence of duct ectasia on biopsy in 25 women with non-lactational abscesses and 21 of 25 were current smokers. Mammillary ,fistulae developed more frequently in current smokers ( P < 0.03). Cigarette smoking is important in the natural history of non-lactational breast abscesses and may predispose to anaerobic breast injections and the development of mammillaryfistulae.
The incidence of lactational breast abscesses is declining but non-lactational breast abscesses are seen more frequently’,’. Non-lactational breast abscesses often recur and may be associated with mammillary fistulae or mammary duct ectasia’,3.4. This study examined the factors which might predispose women with non-lactational breast abscesses to develop further sepsis.
Patients and methods The case records of all women presenting with a diagnosis of a breast abscess over an 11-year period were reviewed ( 1 January 1980-31 December 1990). Breast abscess was defined as an acute inflammatory lump which yielded pus on incision. Details were recorded of the clinical features, smoking habits, surgical treatment and eventual outcome in all women. From 1980 the Nursing Kardex and Selly Oak Hospital Medical Admissions Chart required the patient’s smoking habits to be recorded. A patient was recorded as smoking currently and the number of cigarettes smoked per day noted. Light smokers were defined as women smoking 1 10 cigarettes/day and heavy smokers as women smoking 11 or more cigarettes per day. Women were recorded as ‘never smoked’ if at admission they did not smoke unless the case notes recorded that they had previously done so. When a breast biopsy had been taken histological details were recorded. Bacteriological details were noted for all patients. Histological specimens were reviewed to confirm the presence of dilated ducts and for acute/chronic inflammation surrounding the ducts5. ~
Nineteen women were non-smokers, eight women previously smoked and ten were currently cigarette smokers. Staphylococcus auveus was isolated from 27 patients, anaerobic bacteria from four, coliform bacteria from one woman and no growth was obtained from five women. Six women developed recurrence of their abscess, four who had never smoked, one who was a previous smoker and one who was currently smoking 20 cigarettes per day. Biopsy of the cavity wall from ten women (including all six with recurrent abscesses) failed to show any evidence of mammary duct ectasia. Non-lactational bveust ubscesses Eighty-five women (70 per cent j (mean age 35.8 (range 11 -69) years ) had non-lactational breast abscesses incised and drained. Twenty-eight (33 per cent) had never smoked, eight had previously smoked and 49 (58 per cent) were currently doing so regularly (Table 1 ). Smoking habits were not age-related (Table 2). Women with non-lactational breast abscesses were more likely to be current smokers (x’ = 9.6, P < 0.005) and more likely to have ever smoked (relative risk 2.15, 95 per cent confidence interval 2.04-2.26 ) than women with lactational breast abscesses. The bacteria isolated from the initial abscess are listed in Tuble 3. Fifteen of 18 women who developed anaerobic breast Table 1 Smokiny lzcrhirs q/\c.onwn with non-lacrarionul hrcvst ah.rc.esses
Results A total of 122 women presented with a breast abscess over the 11-year period. Several patients had experienced more than one episode of abscess formation but for the purpose of the study only the first presentation was considered. Lactutional breart abscesses
n ~~
-~
28 8
~ _ _ _ _ _ -
0007 1323’92/010058 02
Recurrent abscess ~~~~~
Never smoked smokers Current smokers Light ( < IO/day) Heavy ( > IO/day)
~~
58
~~~
(
Mammillary fistulae ~
3 1
2 1
13
4
3
36
16
10
~
_______
1992 Butterworth Helnerndnn Ltd
Breast abscesses and cigarette smoking:
Table 2 Frequency of smoking in women with non-lacrational breast abscrsses
Age (years )
Total
Current smokers
Previous smokers
15-24 25-34 35-44 45-54 55-64 > 65 Total
19 19 25 13 3 6 85
10 13 16 7 2 1 49
1 2 3 2 0 0 8
Never smoked 8 4 6 4 1
5 28
Table 3 The relationship between bacterial isolates and duct ecrasia in non-lactational abscesses
Staphylococcus aureus Anaerobic bacteria No growth Coliform bacteria Total
n
Recurrent abscess
Abscess biopsied ( n = 32); duct ectasia confirmed histologically
31 18 32 4 85
6 9* 7 2 24
2 9t 13t 1 25
*The presence of anaerobic bacteria in the initial abscess is associated with recurrence, x 2 = 4.1, P < 0.05; t a biopsy showing duct ectasia was associated with a current or previous abscess containing bacteria other than Staphylococcus aureus, x 2 = 9.8, P < 0.005
abscesses were current cigarette smokers compared with 34 of the 67 women who had other organisms isolated or showed no bacterial growth ( x 2 test with Yates' correction = 4.9, P < 0.05). Twenty-four women developed further breast abscesses following incision and drainage and 16 went on to develop mammillary fistulae. Heavy smokers were significantly more likely to develop further breast abscesses (Armitage test for trend, xz = 7.10, P < 0.01). Women with anaerobic breast abscesses were more likely to develop recurrent breast abscesses ( x 2 test with Yates' correction = 4.1, P < 0.05; Table 3 ) . Mammillary fistulae developed from non-lactational breast abscesses significantly more often in current smokers ( P < 0.03, Table 1 ). Thirty-two women underwent biopsy either at presentation or on recurrence of their abscess and review of the histology revealed 25 women had mammary duct ectasia. The majority of the biopsies were taken when the patients returned with a recurrent abscess or fistula but three women had a biopsy at first presentation which was diagnostic of duct ectasia. Among the 25 women with a histological diagnosis of duct ectasia, 21 were current cigarette smokers, one woman had previously smoked and three had never done so. A bacterial isolate other than Staphylococcus aurcus from the initial abscess was significantly more likely to be associated with a later histological confirmation of duct ectasia ( xz = 9.8, P < 0.005 ; Table 3 ).
Discussion Despite the reported fall in the incidence of lactational breast abscess elsewhere'.2 they accounted for one-third of the abscesses in this series. They were usually associated with Staphylococcus aureus and did not tend to recur. When lactational abscesses did recur this could not be related to smoking habit. Non-lactational breast abscesses tend to recur and three factors have been associated with recurrence ;subareolar site'.3, anaerobic bacteria in the abscess 1.3 and the presence of underlying duct ectasia'.3.
Br. J. Surg., Vol. 79, No. 1, January 1992
N. J.
Bundred st al.
There was no difference between peripheral and subareolar abscesses and although breast abscesses containing anaerobic bacteria were more likely to recur, it was not the most significant factor. The association of histological duct ectasia with recurrence may be because recurrent abscesses tend to be subject to repeated biopsy and this leads to the possibility of an ascertainment bias. It is likely that duct ectasia is responsible for the majority of these abscesses even when histological confirmation is not ~ b t a i n e d ~ . ~ . The aetiology of subareolar sepsis remains unknown. Cigarette smoking is reported to reduce the incidence of benign breast by lowering plasma and urinary oestrogen levels7. The Office of Population Censuses and Surveys reported that in 1988, 17-39 per cent of women were current smokers, the variance being related to social class and age'. A recent survey in the Selly Oak Hospital District has shown that approximately 30 per cent of the female population currently smoke and 10 per cent have previously done so1'. Regardless of age 67 per cent of women with non-lactational breast abscesses and 84 per cent of women with a histological diagnosis of duct ectasia currently smoke or have previously done so. Smoking may be an aetiological agent in this condition. This is supported by the higher recurrence rate of non-lactational breast abscesses in women who are heavy smokers and the association of smoking with anaerobic breast abscess. The recent increase in cigarette smoking over the last decade in young women aged 20-40 years may account for the apparent increase in the incidence of non-lactational breast abscesses in this period','.9. The mechanism is not clear but smoking leads to the production of toxins in ductal secretions and this may damage the lactiferous ducts".". The aetiology of mammillary fistula is ~ n k n o w n ' Nipple ~. inversion and squamous metaplasia of the nipple may be contributory but it is likely that such fistulae are complications of duct ectasiaI4. Only seven of the 16 women who developed mammillary fistulae in this series had nipple inversion, in keeping with previous reports4.l4. Women who were current smokers were significantly more likely to develop mammillary fistulae. The high frequency of sterile cultures from the non-lactational breast abscesses and the retrospective nature of this report require prospective studies of this phenomenon.
References 1. 3
i .
3. 4.
5. 6. 7. 8. 9. 10. 11.
12. 13. 14.
Scholefield JH, Duncan JL, Rogers K. Review of a hospital experience of breast abscess. Br J Sury 1987; 74: 469-70. Bates T, Down RHL, Tant DR, Fiddian RV. The current treatment of breast abscesses in hospital and in general practice. Practitioner 1973; 211: 541-7. Ekland DA, Zeigler MG. Abscess in the nonlactating breast. Arch Surg 1973; 107: 398-401. Lambert ME, Betts CD, Sellwood RA. Mammillary fistulae. Br J Surg 1986; 73: 367-8. Dixon J M , Anderson TJ, Lumsden AB et al. Mammary duct ectasia. Br J Surg 1983; 70: 601-3. Bundred NJ, Dixon JM, Lumsden AB et ul. Are the lesions of duct ectasia sterile. Br J Sury 1985 ; 72 : 844-5. Baron JA. Smoking and oestrogen related disease. Am J Epidemiol 1984; 119: 9-22. Berkowitz GS, Canny P F , Vivolsi VA et al. Cigarette smoking and benign breast disease. J Epidemiol Community V - d f h 1985 : 39: 308-13. Office of Population Censuses and Surveys England and Wales. London: HMSO, 1988. South Birmingham Health Authority. Annual Public Health Report 19YU. Parker: Canterbury, 1990. Wynder EL, Hill P. Nicotine and coninine in breast fluid. Cancer Letr 1979; 6 : 251-4. Petrakis NL, Dupuy ME, Lee RE. Mutagens in nipple aspirate of breast fluid. Bumbury Report 1985 ; 13: 67-82. Editorial. Mammillary fistulae. Lancet 1986; ii: 438-9. Bundred NJ, Dixon JM, Chetty U, Forrest APM. Mammillary fistula. Br J Sury 1987; 74: 844-5.
Paper accepted 28 September 1991
59
