SOC. Sci. Med. Vol. 33, NO. 6, pp. 701-705, 1991 0277-9536/91 S3.00 + 0.00 Princcd in Great Britain. All rights reserved Copyright 0 1991 Pcrgamon Press plc MASTITIS AMONG LACTATING WOMEN: OCCURRENCE AND RISK FACTORS RACHEL KNFMANN and BETSYFOXMAN* Department of Epidemiology, University of Michigan School of Public Health, 109 Observatory Street, Ann Arbor, MI 48109-2029, U.S.A. Abstract-Puerperal mastitis is a potentially serious illness among lactating women which traditionally has been thought to be associated with primiparity, stress, improper nursing technique, and incomplete emptying of the breast. However, none of these putative associations has been examined analytically in recent years. Further, the incidence of mastitis in the United States has not been estimated since 1975, although the prevalence of breastfeeding has increased dramatically since then. In this retrospective cohort studv of 966 lactating women, the cumulative incidence of mastitis in the first seven weeks postmrtum was 3.9%. This incidkce was associated with professional, technical, or managerial occupaiion in both parents (rate ratio = 12.29; 95% CI: 1.62,93.43) and with giving birth in the hospital delivery room, rather than the labor room (rate ratio = 4.05; 95% CI: 0.92, 17.83). Parity was not associated with risk of mastitis in this sample. Key words-mastitis. breastfeeding, stress, epidemiology, occurrence Over the past 15 years, the prevalence of breastfeed- ing in the United States has increased, from 22-28% of all newborns in 1972 [l] to 63% of all newborns in 1984 [2]. More recent national data are not avail- able; however, a 1988 study of primiparae giving birth in three hospitals in Washington, DC found that 84% of white women and 49% of black women breastfed their infants [3]. Despite the increase in breastfeeding, recent medi- cal literature contains few references to mastitis in lactating women. Sporadic puerperal mastitis is an infectious condition of the lactating breast character- ized by local swelling and pain, and flu-like symptoms such as fever, malaise, nausea, and vomiting [4]. Onset of mastitis can occur at any time, but is most likely to develop during the second and third weeks following delivery [5,6]. In most cases, mastitis can be successfully treated with oral antibiotics, and need not interrupt breastfeeding [6,7]. However, in S-1 1% of cases, and particularly when treatment is delayed, an abscess may develop, requiring surgical drainage [5,8]. Occasionally, an abscess may result in sep- ticemia. Thus, puerperal mastitis is a potentially serious complication of breastfeeding. The current incidence of sporadic puerperal masti- tis in the United States is unknown. A 1975 study of women receiving care at a large medical center in California found an incidence rate of 2.5% among breastfeeding mothers [8]. Ogle and Davis [9] have suggested that many cases of mastitis are not reported to physicians and are therefore uncounted, although Devereux [5] claims this is unlikely on the ground that the illness is too severe to be tolerated without treatment. This paper describes an historical cohort study which was designed to investigate the incidence and *To whom reprint requests should be addressed. timing of sporadic puerperal mastitis, as well as generate hypotheses regarding associated risk factors. The cohort consisted of women who gave birth at Women’s Hospital at the University of Michigan during 1984 and 1985; data were collected from delivery and postnatal follow-up medical records. Data collection METHODS Study population. All women who gave birth at Women’s Hospital during 1984 or 1985, who were seen prenatally at the Obstetrics and Gynecology Clinic, and who breastfed their newborn infants were eligible for the study. Review of delivery records identified 3243 women who gave birth at Women’s Hospital between 1 January 1984 and 31 December 1985. According to the delivery records, 60.5% of these women breastfed at time of delivery. About half (1535) were not eligible for the study, because they were not delivered by providers associated with the Obstetrics and Gynecology Clinic. Ultimately, after further excluding women who had ceased breastfeed- ing by time of discharge and women who had not received prenatal care at the Clinic (late referrals to Clinic providers), 966 women met all eligibility cri- teria and were included in the study. Data were abstracted from pre- and post-delivery medical records of eligible women. Variables. Data items collected included: subject’s age, race, marital status and reproductive history; subject’s and partner’s occupation and significant medical history; up to five complications of any type during each of the prenatal, perinatal, and postpar- tum periods; infant’s weight, sex, and health status; results of the routine postpartum checkup; and any clinical diagnosis of mastitis following delivery. (It is hospital procedure to ask new mothers to return for a checkup at about six weeks postpartum. Since not 701