Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Lactation Mastitis: Occurrence and Medical Management among 946 Breastfeeding Women in the United States

Lactation Mastitis: Occurrence and Medical Management among 946 Breastfeeding Women in the United... Abstract In 1994–1998, the authors followed 946 breastfeeding women from Michigan and Nebraska for the first 3 months postpartum or until they stopped breastfeeding to describe mastitis incidence, mastitis treatment, and any associations between mastitis occurrence and hypothesized host characteristics and behaviors. Participants were interviewed by telephone at 3, 6, 9, and 12 weeks postpartum or until they ceased breastfeeding. A total of 9.5% reported provider-diagnosed lactation mastitis at least once during the 12-week period, with 64% diagnosed via telephone. After adjustment in a logistic regression model, history of mastitis with a previous child (odds ratio (OR) = 4.0, 95% confidence interval (CI): 2.64, 6.11), cracks and nipple sores in the same week as mastitis (OR = 3.4, 95% CI: 2.04, 5.51), using an antifungal nipple cream (presumably for nipple thrush) in the same 3-week interval as mastitis (OR = 3.4, 95% CI: 1.37, 8.54), and (for women with no prior mastitis history) using a manual breast pump (OR = 3.3, 95% CI: 1.92, 5.62) strongly predicted mastitis. Feeding fewer than 10 times per day was protective regardless of whether or not feeding frequency in the same week or the week before mastitis was included in the model (for the same week: 7–9 times: OR = 0.6, 95% CI: 0.41, 1.01; ≤6 times: OR = 0.4, 95% CI: 0.19, 0.82). Duration of feeding was not associated with mastitis risk. breast feeding, human body, lactation, mastitis, risk factors CI, confidence interval, OR, odds ratio, RR, rate ratio Lactation mastitis is a cellulitis of the interlobular connective tissue within the mammary gland that usually occurs in the first 6 weeks postpartum (1, 2). The clinical spectrum ranges from focal inflammation with minimal systemic symptoms to abscess and septicemia. Abscesses are reported to occur in 11 percent of all affected women (2). Pain and discomfort associated with mastitis or concerns about treatment with antibiotics that pass through the breast milk may lead some women to cease breastfeeding. The signs and symptoms include fever of 38.5°C or greater; flu-like aches and chills; and a red, tender, hot, swollen, wedge-shaped area of the breast (3). In prospective studies of breastfeeding women, the incidence of lactation mastitis per infant nursed has been reported to be as low as 2.5 percent among United States women (1), 24 percent among Finnish women (4), and 27.1 percent among Australian women (5). In a population-based study conducted in Scotland in the pre-antibiotic era, the incidence was 8.9 percent (6). The agents most frequently cultured are Staphylococcus aureus or coagulase-negative staphylococci (3, 7). Several investigators (2, 3, 8) suggest a fissure in the nipple as a route of infection. Other proposed routes include through the lactiferous ducts into a lobule or by hematogenous spread (2, 3). As there have been few formal studies of lactation mastitis, most hypothesized risk factors are based on clinical impression. Risk factors fall into two general categories: poor breastfeeding technique (1–3, 8, 9) and lowered immune status secondary to stress and sleep deprivation (10). Poor breastfeeding technique may lead to poor drainage of a duct, insufficient emptying of the breast, milk stasis, and cracks or fissures of the nipple. In the one study that formally examined these factors, however, they occurred relatively infrequently: Milk stasis due to weaning or missed feeding preceded nine of the 65 infections, and in eight, the nipple of the involved breast was fissured (1). The number of nipple fissures in the unaffected breast and in women without mastitis was not reported nor was the incidence of milk stasis in women without mastitis. Increased stress and sleep deprivation, which are often complaints of new mothers, have also been reported as risk factors (3). In a retrospective study among participants at a breastfeeding conference, women with a history of mastitis associated their mastitis with fatigue, stress, a plugged duct, change in number of feedings, engorgement/stasis, an infection in the family, breast trauma, and poor diet (11). These findings are suggestive only; the study was retrospective, the occurrence of these factors among women without mastitis was not reported, and the population had a strong commitment to breastfeeding. We followed 946 breastfeeding women for the first 3 months postpartum or until they stopped breastfeeding to describe incidence of mastitis, mastitis treatment, and any associations between mastitis occurrence and hypothesized host characteristics and behaviors. MATERIALS AND METHODS Study protocol Study participants were recruited from women who gave birth at a family birthing center in Michigan and from pregnant women working at a single large company in Nebraska, which had recently implemented a breastfeeding support program for employees. Together, the two sites gave geographic, employment, and socioeconomic status diversity. Enrollment began in Michigan during July 1994, and in Nebraska in November 1994; the last interview was in March 1998. At both sites, potential participants were given information about the study during their pregnancy and a postcard to indicate refusal. The only requirement for enrollment was an intention to breastfeed. All women who did not refuse and who delivered a live infant were sent a letter describing the study and a calendar/diary designed to be used as a memory aid during study interviews. The calendar showed nursing positions and provided spaces to record information pertinent to the study. Women could participate more than once; 43 women did so. There were five sets of twins and one set of triplets. Most potential participants were contacted by telephone at 3 weeks postpartum, and their verbal consent was obtained at the time of the interview. Participants were interviewed at 3, 6, 9 and 12 weeks postpartum or until they ceased breastfeeding. After each interview, they were sent a small incentive, such as video coupons. Fifty-six women missed the first interview and were enrolled at week 6. Interviews were conducted by using computer-assisted technology that enforced skip patterns, checked answers for appropriate ranges, and allowed for textual comments. Interviewers received 4 hours of training, were monitored periodically for adherence to the study protocol, and were retrained annually. Each participant had the same female interviewer for all four interviews. The study protocol was approved by the institutional review boards of all participating institutions. Measuring instruments All interviews included questions on breastfeeding habits, breast-related symptoms, and breast care, including use of breast pumps, bottle-feeding, mastitis, sleeping habits, depression, smoking and drinking habits, and household duties. Questions about the infant's birth; history of mastitis with a previous child, including date of diagnosis; other medical history; and sociodemographic information were asked at enrollment. As much as possible, subjects were asked to recall breastfeeding information on a week-by-week basis. Case definition Mastitis was defined by self-report of mastitis diagnosed by a health care provider. The use of a more specific definition, which included antibiotic prescription and symptoms, did not change the results, and thus, we opted to use the more sensitive definition. Because of an error in the computer-assisted interviewing program, 44 women who ceased breastfeeding sometime during the 3 weeks prior to a study interview were not asked about the occurrence of mastitis. Once the error was discovered, a validation study was conducted; this study is described in detail elsewhere (12) (Gillespie et al., University of Michigan, manuscript submitted for publication). Briefly, we recontacted all women whom the error impacted, all known mastitis cases, and a sample of the remaining participants. Recall of mastitis was found to be quite reliable and was unrelated to age, race, income, parity, history of mastitis, and study site. We detected one additional mastitis case among those in whom the error occurred. This case was included in the analysis to the extent possible. Statistical analysis We describe crude associations between exposures and rate of mastitis with rate ratios and cumulative incidence (risk) ratios and their approximate 95 percent confidence intervals (13). Rate ratios were calculated for time-dependent covariates, for example, breastfeeding habits, and cumulative incidence ratios for covariates that did not change, for example, marital status. Relative risks and confidence intervals were calculated using dEPID software (14). We tested differences between groups for categorical variables with chi-square tests and for continuous variables with Student's t tests, with a two-sided significance level of α = 0.05. We modeled the time to first mastitis as a function of baseline and time-varying covariates. Discrete-time survival analysis was used, with a time-dependent logistic regression model (15) that included site and breastfeeding variables. We estimated the association between hypothesized risk factors in the week of mastitis with mastitis risk by using rate ratios. In cases in which the hypothesized risk factor could have resulted from the mastitis or early signs of mastitis (e.g., cracks and nipple sores), we also calculated (by using the same model) the association of mastitis with the hypothesized variables in the week prior to mastitis. Because women could enter the study more than once, we analyzed the data both considering all births by a single mother as independent and taking into account clustering utilizing generalized estimating equations using the SAS program, Proc Genmod (16). The results were very similar in both analyses, so we present them taking clustering into account. RESULTS Characteristics of the study population A total of 1,057 women were identified for study. One hundred could not be contacted after multiple attempts because of improper contact information, no answer, no response to messages left on telephone answering machines, etc. Of those initially contacted, 11 refused, for a response rate of 99 percent. Of the 946 contacted, consenting women, 112 (11.8 percent) had already ceased breastfeeding at the time of the initial interview and provided 3-week interview data only. Of the 834 women still breastfeeding at 3 weeks, 801 (96.0 percent) completed at least one additional interview, and 658 (78.9 percent) completed all four interviews. Overall, 840 of 946 (88.8 percent) were followed for 12 weeks or until they ceased breastfeeding. The majority (n = 711 (75.2 percent)) of study participants were enrolled from Michigan. Michigan participants tended to be slightly older, to have more education, and to have a slightly higher income than Nebraska participants (table 1). The majority of participants at both sites were White and living with their spouse or partner. Participants from Nebraska were more likely to be first-time mothers and, if they had children, to have fewer children living with them. TABLE 1. Demographic characteristics of 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998* Characteristics  Nebraska   Michigan   No.  %  No.  %  Overall  235  25  711  75  Age (years)           18–19  2  1  10  1   20–24  17  7  49  7   25–29  90  40  218  31   30–34  91  40  257  37   35–39  25  11  139  20   ≥40  2  1  32  5  Race           White  211  90  637  90   Black  14  6  45  6   Other  9  4  26  4  Marital status           Married  221  94  683  96   Living with partner  2  1  9  1   Separated  11  5  17  2  Hours worked outside the home when pregnant           None  1  0  252  36   <20  2  1  91  13   21–39  39  17  145  21   ≥40  192  82  214  30  Education           Less than high school  67  29  108  15   High school diploma  54  23  162  23   College degree  91  39  330  47   Professional or graduate training  23  10  110  16  Total family income (dollars)           <$25,000  33  14  78  12   $25,000–$49,999  88  38  235  35   $50,000–$99,999  103  45  319  47   ≥$100,000  6  3  48  7  Parity (no. of livebirths)           1  103  44  201  28   2  84  36  239  34   3  35  15  164  23   4  8  3  64  9   ≥5  4  2  42  6  Characteristics  Nebraska   Michigan   No.  %  No.  %  Overall  235  25  711  75  Age (years)           18–19  2  1  10  1   20–24  17  7  49  7   25–29  90  40  218  31   30–34  91  40  257  37   35–39  25  11  139  20   ≥40  2  1  32  5  Race           White  211  90  637  90   Black  14  6  45  6   Other  9  4  26  4  Marital status           Married  221  94  683  96   Living with partner  2  1  9  1   Separated  11  5  17  2  Hours worked outside the home when pregnant           None  1  0  252  36   <20  2  1  91  13   21–39  39  17  145  21   ≥40  192  82  214  30  Education           Less than high school  67  29  108  15   High school diploma  54  23  162  23   College degree  91  39  330  47   Professional or graduate training  23  10  110  16  Total family income (dollars)           <$25,000  33  14  78  12   $25,000–$49,999  88  38  235  35   $50,000–$99,999  103  45  319  47   ≥$100,000  6  3  48  7  Parity (no. of livebirths)           1  103  44  201  28   2  84  36  239  34   3  35  15  164  23   4  8  3  64  9   ≥5  4  2  42  6  * The two sites were significantly different with respect to age, working during pregnancy, education of the respondent, parity (p < 0.01), and income (p = 0.05). Numbers may not total 946 because of missing values. View Large Mastitis incidence During the 12-week study period, 77 women (8.1 percent) had one case of self-reported, health care provider-diagnosed mastitis; 12 (1.3 percent) had two cases, and one (0.1 percent) had three cases, for an overall incidence of first mastitis of 9.5 percent. The cumulative incidence of a first case of mastitis was 10.3 percent in Michigan and 7.2 percent in Nebraska. There was no statistically significant difference in incidence by site. Mastitis incidence was highest in the first few weeks postpartum and fell gradually thereafter (figure 1). Cracks and nipple sores occurred at a much higher frequency than mastitis, with the highest incidence in the first (36 percent) and second (14 percent) weeks postpartum. Women who reported a history of mastitis with a previous child were more likely to have mastitis with the current baby in every follow-up period (figure 2). FIGURE 1. View largeDownload slide Incidence of mastitis and nipple cracks or sores by week postpartum, Michigan and Nebraska, 1994–1998. FIGURE 1. View largeDownload slide Incidence of mastitis and nipple cracks or sores by week postpartum, Michigan and Nebraska, 1994–1998. FIGURE 2. View largeDownload slide Mastitis incidence, with the upper limits of the 95% confidence intervals shown by “whisker”, by 3-week postpartum interval and mastitis history, Michigan and Nebraska, 1994–1998. FIGURE 2. View largeDownload slide Mastitis incidence, with the upper limits of the 95% confidence intervals shown by “whisker”, by 3-week postpartum interval and mastitis history, Michigan and Nebraska, 1994–1998. Mastitis diagnosis, clinical presentation, disease impact, and treatment Women usually were diagnosed by their obstetrician/gynecologist (38 percent), but also by family practitioners (21 percent), nurses (23 percent), and others (18 percent). The majority (64 percent) were diagnosed after a telephone contact with their health care provider and did not receive a physical examination. Women diagnosed over the telephone and those diagnosed in person were similar with respect to symptoms and sociodemographic variables (data not shown). The most common mastitis symptoms were breast tenderness (98 percent), fever (82 percent), malaise (87 percent), chills (78 percent), redness (78 percent), and a hot spot (a localized area of warmth and tenderness on the affected breast) (62 percent). Women averaged 4.9 days when they had one or more symptoms (range, 1–21 days). These symptoms were severe enough to cause the vast majority of women to restrict their activities at least somewhat (77 percent) and to spend at least 1 day in bed (72 percent). Almost half of the women changed their breastfeeding habits, either feeding more often (36 percent), less often (11 percent), feeding more often (49 percent) or less often (8 percent) on the affected breast, changing nursing holds more frequently (33 percent), trying a new nursing hold (12 percent), or making some other change (19 percent). Most women with mastitis (88 percent) were prescribed one or more medications, either antibiotics (86 percent) and/or analgesics (17 percent); two woman reported receiving an antifungal medication, and another did not know the name of the drug prescribed. The most commonly prescribed antibiotics were cephalexin (46 percent), amoxicillin (7 percent), ampicillin (7 percent), and augmentin (7 percent). No cultures were performed. In addition, health care providers suggested applying hot compresses (83 percent), fully emptying the affected breast (65 percent), feeding more frequently (74 percent), and changing feeding positions often (48 percent). We did not ask whether women were advised to stop breastfeeding. Associations of risk factors with cumulative incidence of mastitis and cracks and nipple sores The overall cumulative incidence (risk) of mastitis among those who had never breastfed previously was 7.3 percent compared with 10.8 percent among those who had breastfed previously. For women with a history of mastitis, the risk was 23.9 percent compared with 8.3 percent among first-time mothers (rate ratio (RR) = 2.9, 95 percent confidence interval (CI): 1.79, 4.69) (table 2). Women with a history of mastitis were also more likely to report cracks and nipple sores (RR = 1.3, 95 percent CI: 1.04, 1.50). Neither the cumulative incidence of mastitis nor that of cracks and nipple sores was associated with age, income, returning to work, or participating in a breastfeeding support group (table 2). Women who reported having outside help were less likely to report cracks or nipple sores (odds ratio (OR) = 0.8, 95 percent CI: 0.71, 0.97), but having outside help was not associated with mastitis. TABLE 2. Relative risk (cumulative incidence ratio) of lactation mastitis and cracks and nipple sores by mastitis history and sociodemographic characteristics for 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998 Predictor  Mastitis   Cracks and nipple sores   RR*  95% CI*  RR  95% CI  Mastitis history           First-time mother (n = 303)  Reference    Reference     Parity >1, never breastfed before (n = 51)  0.2  0.03, 1.72  0.5  0.30, 0.83   Breastfed before, no mastitis (n = 458)  0.9  0.51, 1.40  0.9  0.77, 1.05   Mastitis in the past (n = 134)  2.9  1.79, 4.69  1.3  1.04, 1.50  Site           Nebraska (n = 235)  0.7  0.42, 1.17  0.8  0.64, 0.93   Michigan (n = 711)  Reference    Reference    Age (years) (missing for 14 women)          18–24 (n = 78)  Reference    Reference     25–29 (n = 308)  1.6  0.63, 3.91  0.7  0.55, 0.91   30–34 (n = 348)  1.4  0.58, 3.56  0.9  0.70, 1.11   ≥35 (n = 198)  1.7  0.68, 4.42  1.0  0.76, 1.25  Income (dollars) (missing for 36 women)           <$25,000 (n = 111)  Reference    Reference     $25,000–$49,999 (n = 323)  1.0  0.50, 1.98  1.1  0.88, 1.47   $50,000–$99,999 (n = 422)  0.9  0.47, 1.80  1.1  0.87, 1.43   ≥$100,000 (n = 54)  1.4  0.58, 3.57  1.0  0.67, 1.47  Marital status (missing for 3 women)           Not married (n = 39)  No mastitis in those not married    0.7  0.46, 1.14   Married (n = 904)  Reference    Reference    Ethnicity (missing for 4 women)           Non-White (n = 94)  0.8  0.37, 1.62  0.7  0.50, 0.92   White (n = 848)  Reference    Reference    Education (missing for 1 woman)           Less than high school (n = 175)  Reference    Reference     High school diploma (n = 216)  1.2  0.60, 2.22  1.2  0.93, 1.54  College degree (n = 421)  1.2  0.65, 2.08  1.4  1.09, 1.70   Professional or graduate degree (n = 133)  1.6  0.82, 3.12  1.4  1.08, 1.81  Children living at home (missing for 3 women)           1 (n = 293)  Reference    Reference     2 (n = 330)  1.1  0.69, 1.87  1.0  0.83, 1.16   3 (n = 198)  1.3  0.76, 2.24  0.9  0.72, 1.09   4 (n = 75)  0.9  0.40, 2.20  1.0  0.75, 1.29   ≥5 (n = 47)  1.0  0.36, 2.74  0.9  0.59, 1.24  Smoking during any week (missing for 1 woman)           Yes (n = 61)  No mastitis in smokers    1.1  0.84, 1.42   No (n = 884)  Reference    Reference    Drinking during any week (missing for 82 women)           Yes (n = 442)  1.1  0.74, 1.61  1.0  0.86, 1.12   No (n = 422)  Reference    Reference    Return to work by week 12           Yes (n = 60)  0.7  0.28, 1.90  1.0  0.74, 1.32   No (n = 886)  Reference    Reference    Breastfeeding support group (missing for 83 women)           Yes (n = 96)  1.4  0.78, 2.33  1.0  0.81, 1.24   No (n = 767)  Reference    Reference    Outside help in any wave (missing for 1 woman)           Yes (n = 633)  1.2  0.79, 1.87  0.8  0.71, 0.97   No (n = 312)  Reference    Reference    Predictor  Mastitis   Cracks and nipple sores   RR*  95% CI*  RR  95% CI  Mastitis history           First-time mother (n = 303)  Reference    Reference     Parity >1, never breastfed before (n = 51)  0.2  0.03, 1.72  0.5  0.30, 0.83   Breastfed before, no mastitis (n = 458)  0.9  0.51, 1.40  0.9  0.77, 1.05   Mastitis in the past (n = 134)  2.9  1.79, 4.69  1.3  1.04, 1.50  Site           Nebraska (n = 235)  0.7  0.42, 1.17  0.8  0.64, 0.93   Michigan (n = 711)  Reference    Reference    Age (years) (missing for 14 women)          18–24 (n = 78)  Reference    Reference     25–29 (n = 308)  1.6  0.63, 3.91  0.7  0.55, 0.91   30–34 (n = 348)  1.4  0.58, 3.56  0.9  0.70, 1.11   ≥35 (n = 198)  1.7  0.68, 4.42  1.0  0.76, 1.25  Income (dollars) (missing for 36 women)           <$25,000 (n = 111)  Reference    Reference     $25,000–$49,999 (n = 323)  1.0  0.50, 1.98  1.1  0.88, 1.47   $50,000–$99,999 (n = 422)  0.9  0.47, 1.80  1.1  0.87, 1.43   ≥$100,000 (n = 54)  1.4  0.58, 3.57  1.0  0.67, 1.47  Marital status (missing for 3 women)           Not married (n = 39)  No mastitis in those not married    0.7  0.46, 1.14   Married (n = 904)  Reference    Reference    Ethnicity (missing for 4 women)           Non-White (n = 94)  0.8  0.37, 1.62  0.7  0.50, 0.92   White (n = 848)  Reference    Reference    Education (missing for 1 woman)           Less than high school (n = 175)  Reference    Reference     High school diploma (n = 216)  1.2  0.60, 2.22  1.2  0.93, 1.54  College degree (n = 421)  1.2  0.65, 2.08  1.4  1.09, 1.70   Professional or graduate degree (n = 133)  1.6  0.82, 3.12  1.4  1.08, 1.81  Children living at home (missing for 3 women)           1 (n = 293)  Reference    Reference     2 (n = 330)  1.1  0.69, 1.87  1.0  0.83, 1.16   3 (n = 198)  1.3  0.76, 2.24  0.9  0.72, 1.09   4 (n = 75)  0.9  0.40, 2.20  1.0  0.75, 1.29   ≥5 (n = 47)  1.0  0.36, 2.74  0.9  0.59, 1.24  Smoking during any week (missing for 1 woman)           Yes (n = 61)  No mastitis in smokers    1.1  0.84, 1.42   No (n = 884)  Reference    Reference    Drinking during any week (missing for 82 women)           Yes (n = 442)  1.1  0.74, 1.61  1.0  0.86, 1.12   No (n = 422)  Reference    Reference    Return to work by week 12           Yes (n = 60)  0.7  0.28, 1.90  1.0  0.74, 1.32   No (n = 886)  Reference    Reference    Breastfeeding support group (missing for 83 women)           Yes (n = 96)  1.4  0.78, 2.33  1.0  0.81, 1.24   No (n = 767)  Reference    Reference    Outside help in any wave (missing for 1 woman)           Yes (n = 633)  1.2  0.79, 1.87  0.8  0.71, 0.97   No (n = 312)  Reference    Reference    * RR, relative risk; CI, confidence interval. View Large Associations of breastfeeding and other health behaviors with rate of mastitis Women with previous mastitis compared with those without this history had significantly more weeks with nipple cracks or sores, more weeks with 10 or more feedings per day, and more weeks with feedings of 10 minutes or less; used fewer positions; and were less likely to pump, bottle-feed, wash their breasts either before or after feeding, and use nursing bras. Because of the strong association of mastitis with mastitis history, we present associations separately by mastitis history. Cracks and nipple sores that occurred during the same week as mastitis were associated with an almost sixfold increase in mastitis rate for women with no history of mastitis and a threefold increase in mastitis rate for women with such a history (table 3). Women without cracks and nipple sores in the same week as mastitis had similar rates of mastitis in the right (8.8/1,000 person-weeks) as in the left (8.5/1,000 person-weeks) breast and lower rates of mastitis in both breasts at the same time (1.5/1,000 person-weeks). Women with cracks and nipple sores in one breast in the same week as mastitis had higher rates of mastitis in the breast with cracks or nipple sores (48.9/1,000 person-weeks) than in the breast without them (13.3/1,000 person-weeks). Women with cracks or nipple sores in both breasts in the same week as mastitis experienced the highest mastitis rate (88.7/1,000 person-weeks). TABLE 3. Unadjusted rate ratios for the association of breastfeeding and other health behaviors with lactation mastitis by history of mastitis, Michigan and Nebraska, 1994–1998 Characteristic  No previous mastitis   Mastitis history   Person-weeks exposed  Mastitis rate/1,000  RR*  95% CI*  Person-weeks exposed  Mastitis rate/1,000  RR  95% CI  Nipple cracks or sores†                   Same week as mastitis                1.38, 6.84    Yes  591  30.5  5.6  3.21, 9.77  146  54.8  3.1      No  7,353  5.4      1,345  17.8       Week before mastitis                0.46, 3.14    Yes  849  20.0  3.5  1.99, 6.18  204  24.5  1.2      No  7,000  5.7      1,280  20.3      Any ointment use                0.26, 4.50    Yes  545  11.0  1.6  0.67, 3.64  87  23.0  1.1      No  7,384  7.0      1,404  21.4      Antifungal ointment                0.54, 29.12    Yes  63  47.6  6.8  2.13, 21.78  12  83.3  4.0      No  7,866  7.0      1,479  21.0      Frequency of feeding (times/day)                   Same week as mastitis                   <6  1,937  2.1  0.2  0.06, 0.46  339  14.7  0.4  0.15, 1.11    7–9  3,189  7.2  0.6  0.32, 0.98  654  15.3  0.4  0.20, 0.94   ≥10  2,183  12.8      476  35.7       Week before mastitis                   <6  1,694  2.4  0.2  0.07, 0.55  292  13.7  0.4  0.14, 1.19    7–9  3,211  6.5  0.53  0.03, 0.93  642  14.0  0.4  0.19, 0.92    ≥10  2,440  12.3      530  34.0      Duration of feeding (minutes)†                   Same week as mastitis                    <10  1, 037  5.8      284  31.7        11–20  3,605  8.0  1.4  0.58, 3.35  825  17.0  0.5  0.23, 1.24    21–30  1,836  7.1  1.2  0.47, 3.22  281  24.9  0.8  0.29, 2.11    ≥31  850  9.4  1.6  0.56, 4.69  78  25.6  0.8  0.17, 3.74   Week before mastitis                    <10  1,051  7.6      285  31.6        11–20  3,586  7.0  0.9  0.41, 2.03  809  18.5  0.6  0.26, 1.34    21–30  1,831  7.1  0.9  0.39, 2.25  279  17.9  0.6  0.19, 1.69    ≥31  896  11.2  1.5  0.58, 3.71  90  22.2  0.7  0.15, 3.26  Pain when latching on                    Yes  3,084  12.6  3.2  1.86, 5.58  588  35.7  2.9  1.41, 6.08    No  4,845  3.9      903  12.2      Pain when nursing                    Yes  2,979  14.4  4.8  2.65, 8.57  597  36.9  3.3  1.56, 6.96    No  4,950  3.0      894  11.2      Pain when not nursing                    Yes  1,938  17.0  4.1  2.43, 6.86  372  43.0  3.0  1.50, 6.01    No  5,991  4.2      1,119  14.2      Pain all the time                    Yes  381  21.0  3.2  1.50, 6.67  126  71.4  4.2  1.96, 2.91    No  7,548  6.6      1,365  1.7      Any breast pain in the week before mastitis                    Yes  1,540  8.4  1.2  0.65, 2.25  277  18.1  0.8  0.32, 3.18    No  6,305  7.0      1,205  21.6      No. of positions used†                    1  2,022  5.4      288  6.9        2  3,678  6.0  1.1  0.52, 2.22  690  15.9  2.3  0.51, 10.36    ≥3  4,433  5.6  1.0  0.51, 2.11  516  36.8  5.3  1.24, 22.76  Pumping‡                   Same week as mastitis                    Any  3,346  10.8  2.2  1.32, 3.81  478  20.9  1.0  0.46, 2.03    Electric pump  1,948  9.2  1.9  1.03, 3.59  198  25.3  1.2  0.44, 3.07    Manual pump  956  15.7  3.3  1.70, 6.30  195  15.4  0.7  0.21, 2.37    By hand  722  8.3  1.7  0.70, 4.27  126  15.9  0.7  0.17, 3.11    No pumping  4,583  4.8  Reference    1,013  21.7       Week before mastitis                    Any  3,265  7.7  1.1  0.65, 1.85  469  21.3  1.0  0.46, 2.19    Electric pump  1,850  6.5  0.9  0.48, 1.81  187  21.4  1.0  0.35, 3.01    Manual pump  949  11.6  1.7  0.84, 3.30  187  16.0  0.8  0.23, 2.60    By hand  750  5.3  0.8  0.27, 2.16  134  22.4  1.1  0.32, 3.63    No pumping  4,587  7.0      1,015  20.7      Bottle fed                    Yes  3,273  6.7  0.9  0.51, 1.48  445  18.0  0.8  0.35, 1.74    No  4,653  7.7      1,046  22.9      Engorgement                    Yes  564  12.4  1.8  0.81, 3.94  89  22.5  1.0  0.25, 4.26    No  7,350  6.9      1,402  21.4      Washed nipples before feeding                    Yes  198  15.2  2.1  0.67, 6.81  15  0§  1.5  0.09, 24.72    No  7,731  7.1      1,476  21.7      Washed nipples after feeding                    Yes  152  13.2  1.8  0.45, 7.49  14  0§  1.6  0.10, 26.50    No  7,777  7.2      1,477  21.7      Used nursing pads                    Yes  4,767  8.6  1.7  0.87, 3.16  927  22.7  1.1  0.51, 2.19    No  2,319  5.2      513  21.4      Wore a nursing bra                    Yes  6,396  7.8  1.6  0.70, 3.80  1,272  22.8  1.4  0.42, 4.52    Some  276  7.2  1.5  0.31, 7.52  39  0.0  0.7  0.03, 12.55    No  1,257  4.8      180  16.7      Took daytime naps                   Same week as mastitis                    Yes  4,358  9.4  1.8  1.01, 3.13  771  23.3  1.2  0.75, 1.90    No  3,214  5.3      713  19.6       Week before mastitis                    Yes  4,681  8.5  1.4  0.82, 2.54  833  22.8  1.2  0.59, 2.52    No  2,862  5.9      644  18.6      Characteristic  No previous mastitis   Mastitis history   Person-weeks exposed  Mastitis rate/1,000  RR*  95% CI*  Person-weeks exposed  Mastitis rate/1,000  RR  95% CI  Nipple cracks or sores†                   Same week as mastitis                1.38, 6.84    Yes  591  30.5  5.6  3.21, 9.77  146  54.8  3.1      No  7,353  5.4      1,345  17.8       Week before mastitis                0.46, 3.14    Yes  849  20.0  3.5  1.99, 6.18  204  24.5  1.2      No  7,000  5.7      1,280  20.3      Any ointment use                0.26, 4.50    Yes  545  11.0  1.6  0.67, 3.64  87  23.0  1.1      No  7,384  7.0      1,404  21.4      Antifungal ointment                0.54, 29.12    Yes  63  47.6  6.8  2.13, 21.78  12  83.3  4.0      No  7,866  7.0      1,479  21.0      Frequency of feeding (times/day)                   Same week as mastitis                   <6  1,937  2.1  0.2  0.06, 0.46  339  14.7  0.4  0.15, 1.11    7–9  3,189  7.2  0.6  0.32, 0.98  654  15.3  0.4  0.20, 0.94   ≥10  2,183  12.8      476  35.7       Week before mastitis                   <6  1,694  2.4  0.2  0.07, 0.55  292  13.7  0.4  0.14, 1.19    7–9  3,211  6.5  0.53  0.03, 0.93  642  14.0  0.4  0.19, 0.92    ≥10  2,440  12.3      530  34.0      Duration of feeding (minutes)†                   Same week as mastitis                    <10  1, 037  5.8      284  31.7        11–20  3,605  8.0  1.4  0.58, 3.35  825  17.0  0.5  0.23, 1.24    21–30  1,836  7.1  1.2  0.47, 3.22  281  24.9  0.8  0.29, 2.11    ≥31  850  9.4  1.6  0.56, 4.69  78  25.6  0.8  0.17, 3.74   Week before mastitis                    <10  1,051  7.6      285  31.6        11–20  3,586  7.0  0.9  0.41, 2.03  809  18.5  0.6  0.26, 1.34    21–30  1,831  7.1  0.9  0.39, 2.25  279  17.9  0.6  0.19, 1.69    ≥31  896  11.2  1.5  0.58, 3.71  90  22.2  0.7  0.15, 3.26  Pain when latching on                    Yes  3,084  12.6  3.2  1.86, 5.58  588  35.7  2.9  1.41, 6.08    No  4,845  3.9      903  12.2      Pain when nursing                    Yes  2,979  14.4  4.8  2.65, 8.57  597  36.9  3.3  1.56, 6.96    No  4,950  3.0      894  11.2      Pain when not nursing                    Yes  1,938  17.0  4.1  2.43, 6.86  372  43.0  3.0  1.50, 6.01    No  5,991  4.2      1,119  14.2      Pain all the time                    Yes  381  21.0  3.2  1.50, 6.67  126  71.4  4.2  1.96, 2.91    No  7,548  6.6      1,365  1.7      Any breast pain in the week before mastitis                    Yes  1,540  8.4  1.2  0.65, 2.25  277  18.1  0.8  0.32, 3.18    No  6,305  7.0      1,205  21.6      No. of positions used†                    1  2,022  5.4      288  6.9        2  3,678  6.0  1.1  0.52, 2.22  690  15.9  2.3  0.51, 10.36    ≥3  4,433  5.6  1.0  0.51, 2.11  516  36.8  5.3  1.24, 22.76  Pumping‡                   Same week as mastitis                    Any  3,346  10.8  2.2  1.32, 3.81  478  20.9  1.0  0.46, 2.03    Electric pump  1,948  9.2  1.9  1.03, 3.59  198  25.3  1.2  0.44, 3.07    Manual pump  956  15.7  3.3  1.70, 6.30  195  15.4  0.7  0.21, 2.37    By hand  722  8.3  1.7  0.70, 4.27  126  15.9  0.7  0.17, 3.11    No pumping  4,583  4.8  Reference    1,013  21.7       Week before mastitis                    Any  3,265  7.7  1.1  0.65, 1.85  469  21.3  1.0  0.46, 2.19    Electric pump  1,850  6.5  0.9  0.48, 1.81  187  21.4  1.0  0.35, 3.01    Manual pump  949  11.6  1.7  0.84, 3.30  187  16.0  0.8  0.23, 2.60    By hand  750  5.3  0.8  0.27, 2.16  134  22.4  1.1  0.32, 3.63    No pumping  4,587  7.0      1,015  20.7      Bottle fed                    Yes  3,273  6.7  0.9  0.51, 1.48  445  18.0  0.8  0.35, 1.74    No  4,653  7.7      1,046  22.9      Engorgement                    Yes  564  12.4  1.8  0.81, 3.94  89  22.5  1.0  0.25, 4.26    No  7,350  6.9      1,402  21.4      Washed nipples before feeding                    Yes  198  15.2  2.1  0.67, 6.81  15  0§  1.5  0.09, 24.72    No  7,731  7.1      1,476  21.7      Washed nipples after feeding                    Yes  152  13.2  1.8  0.45, 7.49  14  0§  1.6  0.10, 26.50    No  7,777  7.2      1,477  21.7      Used nursing pads                    Yes  4,767  8.6  1.7  0.87, 3.16  927  22.7  1.1  0.51, 2.19    No  2,319  5.2      513  21.4      Wore a nursing bra                    Yes  6,396  7.8  1.6  0.70, 3.80  1,272  22.8  1.4  0.42, 4.52    Some  276  7.2  1.5  0.31, 7.52  39  0.0  0.7  0.03, 12.55    No  1,257  4.8      180  16.7      Took daytime naps                   Same week as mastitis                    Yes  4,358  9.4  1.8  1.01, 3.13  771  23.3  1.2  0.75, 1.90    No  3,214  5.3      713  19.6       Week before mastitis                    Yes  4,681  8.5  1.4  0.82, 2.54  833  22.8  1.2  0.59, 2.52    No  2,862  5.9      644  18.6      * RR, relative risk; CI, confidence interval. † Rate ratios for no previous mastitis and mastitis history were significantly different for nipple cracks and sores, feeding for 11–20 minutes, and number of positions used. ‡ Categories are not mutually exclusive. All comparisons are to no pumping. § For tables with zero cells, odds ratios were calculated by adding 1/2 to all cells. View Large Cracks and nipple sores may be treated with ointments; however, there was little association with any ointment use during the same week as mastitis. There was a twofold increase (RR = 2.0, 95 percent CI: 1.00, 4.17) for any ointment use in the week before mastitis among women with no history of mastitis. Using an antifungal ointment during the same week as mastitis, but not during the previous week, was statistically significantly associated with mastitis among women with no mastitis history (RR = 6.8, 95 percent CI: 2.13, 21.78). We do not know whether the antifungal creams were self- or physician-prescribed. Average frequency of feeding per day, but not duration of feeding, was associated with mastitis rate; this was true whether or not the feeding habits were in the same week or the week before mastitis. Women without a history of mastitis who fed six or fewer times a day had a rate of mastitis five times lower than those who fed 10 or more times a day (same week as mastitis: RR = 0.2, 95 percent CI: 0.06, 0.46; week before mastitis: RR = 0.2, 95 percent CI: 0.07, 0.55); for women with a history of mastitis, the rate was 2.5 times lower (same week as mastitis: RR = 0.4, 95 percent CI: 0.15, 1.11; week before mastitis: RR = 0.4, 95 percent CI: 0.14, 1.19) (figure 3). FIGURE 3. View largeDownload slide Three-month incidence of mastitis, with the upper limits of the 95 percent confidence intervals shown by “whisker”, by average number of feedings per day and mastitis history, Michigan and Nebraska, 1994–1998. FIGURE 3. View largeDownload slide Three-month incidence of mastitis, with the upper limits of the 95 percent confidence intervals shown by “whisker”, by average number of feedings per day and mastitis history, Michigan and Nebraska, 1994–1998. Breast pain when latching on, nursing, not nursing, or all the time also were associated with an increased rate of mastitis; however, if lagged by 1 week, they were not associated, suggesting that the breast pain may have been a symptom of mastitis rather than a risk factor of acquisition. Similarly, women who reported using more nursing positions had a higher rate of mastitis among those with a history of mastitis, but an increased number of positions likely reflects a response to nipple cracks or sores, breast pain, or mastitis. This variable could not be lagged because it was not asked on a week-by-week basis. Among women with no history of mastitis, use of a breast pump, particularly a manual pump, resulted in an increased rate of mastitis. A history of using a breast pump for the previous week was also associated with mastitis, but the rate ratio was reduced and was no longer statistically significant; for other breast pumps, use of a pump during the previous week also noticeably reduced the association. For women with a history of mastitis, using a pump had no association with mastitis rate, with the point estimates for manual pumping suggesting a protective effect for both the same and the previous week. Other breastfeeding behaviors, including bottle-feeding, engorgement, washing breasts before or after feeding, using nursing pads, and wearing a nursing bra had little association with mastitis rate; none of the observed associations were statistically significant. Women who reported taking daytime naps had higher rates of mastitis, but, again, this result could be an effect rather than a cause of mastitis. When information from the previous week was used, there was still a modest increase in the rate for women with no history of mastitis, but it was no longer statistically significant. Discrete survival regression analysis To explore further the relations between variables identified in the bivariate and stratified analyses, we fit a time-dependent logistic regression model (table 4), including mastitis history, weeks since birth, and the following variables measured in the same week as mastitis: nipple sores or cracks, antifungal nipple cream, frequency of feeding, and use of a manual breast pump. Duration of feeding and daytime naps were not associated with risk of mastitis after adjustment for other variables and, thus, were not included in the model. After adjustment for covariates and clustering between responses for women who participated more than once, mastitis history, cracks and nipple sores in the same week as mastitis, and use of antifungal nipple cream in the same 3-week interval as mastitis remained positively associated with mastitis, increasing risk more than threefold. Those who fed fewer than 10 times per day in the same week as mastitis were at significantly decreased risk of mastitis, with those feeding fewer than six times per day having a risk 2.5 times lower than women feeding 10 or more times a day. Use of a manual breast pump in women without mastitis history was also significantly associated with mastitis rate. The interaction with mastitis history (p = 0.02) indicates that those with mastitis history do not share the increased risk with manual breast pump use: Using a manual breast pump increased risk of mastitis 2.1 times (95 percent CI: 1.09, 3.86) among women with no mastitis history, but had no association with mastitis among women who had experienced mastitis with a previous baby. When the model was fit using covariates from the previous week, the results were similar, although cracks and nipple sores were only marginally statistically significant (p = 0.06). The interaction between mastitis history and use of a manual breast pump was also no longer statistically significant. In both models, the time variables showed mastitis rates to be significantly higher in weeks 1–6 relative to weeks 10–12. TABLE 4. Odds ratios and 95 percent confidence intervals for the associations of breastfeeding practices and other health behaviors with lactation mastitis: results from a time-dependent logistic regression model using characteristics from the same week as mastitis and from the week before mastitis* for 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998 Week of time-dependent covariates  OR†  95% CI†  p value  Covariates from the same week as mastitis         Mastitis history  4.0  2.64, 6.11  0.0000   Nipple cracks or sores  3.4  2.04, 5.51  0.0000   Antifungal nipple cream  3.4  1.37, 8.54  0.009   Feed (times/day)          <6  0.4  0.19, 0.82  0.01    7–9  0.6  0.41, 1.01  0.06    ≥10  Reference      Using a manual breast pump          No mastitis history  3.3  1.92, 5.62  0.0000    Mastitis history  0.9  0.21, 4.04     Manual breast pump* mastitis history (weeks)      0.02    1–3  3.3  1.41, 7.74  0.006    4–6  3.0  1.31, 6.89  0.009    7–9  2.4  1.02, 5.73  0.05    10–12  Reference      Covariates from the week before mastitis         Mastitis history  3.3  2.12, 5.16  0.0000   Nipple cracks or sores  1.7  0.96, 2.86  0.07   Antifungal nipple cream‡  3.4  1.06, 10.80  0.04   Feed (times/day)          <6  0.4  0.20, 0.89  0.02    7–9  0.6  0.40, 0.99  0.05    ≥10  Reference       Using a manual breast pump          No mastitis history  2.1  1.09, 3.86  0.03    Mastitis history  0.5  0.12, 1.65  0.23   Manual breast pump* mastitis history (weeks)          1–3  3.7  1.53, 8.96  0.004    4–6  3.0  1.22, 7.23  0.02    7–9  2.3  0.93, 5.86  0.07    10–12  Reference      Week of time-dependent covariates  OR†  95% CI†  p value  Covariates from the same week as mastitis         Mastitis history  4.0  2.64, 6.11  0.0000   Nipple cracks or sores  3.4  2.04, 5.51  0.0000   Antifungal nipple cream  3.4  1.37, 8.54  0.009   Feed (times/day)          <6  0.4  0.19, 0.82  0.01    7–9  0.6  0.41, 1.01  0.06    ≥10  Reference      Using a manual breast pump          No mastitis history  3.3  1.92, 5.62  0.0000    Mastitis history  0.9  0.21, 4.04     Manual breast pump* mastitis history (weeks)      0.02    1–3  3.3  1.41, 7.74  0.006    4–6  3.0  1.31, 6.89  0.009    7–9  2.4  1.02, 5.73  0.05    10–12  Reference      Covariates from the week before mastitis         Mastitis history  3.3  2.12, 5.16  0.0000   Nipple cracks or sores  1.7  0.96, 2.86  0.07   Antifungal nipple cream‡  3.4  1.06, 10.80  0.04   Feed (times/day)          <6  0.4  0.20, 0.89  0.02    7–9  0.6  0.40, 0.99  0.05    ≥10  Reference       Using a manual breast pump          No mastitis history  2.1  1.09, 3.86  0.03    Mastitis history  0.5  0.12, 1.65  0.23   Manual breast pump* mastitis history (weeks)          1–3  3.7  1.53, 8.96  0.004    4–6  3.0  1.22, 7.23  0.02    7–9  2.3  0.93, 5.86  0.07    10–12  Reference      * Adjusted for clustering between the 43 women who participated twice, using generalized estimating equations. † OR, odds ratio; CI, confidence interval. ‡ This variable was not lagged, as measured in 3-week intervals. View Large DISCUSSION In this prospective cohort study of 946 breastfeeding women followed for the first 3 months postpartum, 9.5 percent of the population self-reported health care provider-diagnosed lactation mastitis at least once. The strongest risk factors were history of mastitis with a previous child, cracks and nipple sores in the same week, using an antifungal nipple cream (presumably for nipple thrush) in the same week, feeding the baby more frequently and, for women with no prior mastitis history, using a manual breast pump in the same week. The incidence was essentially the same by sociodemographic variables, study site, sleeping habits, and breastfeeding practices other than frequency of feeding. This result is in contrast to a study in southwest Finland that found an increased incidence among women aged 21–35 years relative to younger and older women (4). The incidence of mastitis reported here is approximately half that reported from prospective cohorts in Australia (27 percent, 95 percent CI: 22, 32 percent, n = 306 (5)), New South Wales (20 percent, 95 percent CI: 18, 22 percent, n = 1,075 (17)), or southwest Finland (24 percent, 95 percent CI: 21, 27 percent, n = 664) (4)), but was more than three times the incidence reported in a United States cohort (2.5 percent, 95 percent CI: 1.8–3.0 percent, n = 2,534) (1) . This may be explained by differences in case definition, population, temporal variation, selection biases, or other unmeasured factors. Nevertheless, there is general agreement that mastitis is a common bacterial infection among breastfeeding women. One of the enduring myths of mastitis risk is that it results from inexperience with breastfeeding. However, in our study, women with a history of mastitis—by definition, previous breastfeeders—had a greater risk. Adjustment for the observed breastfeeding practices did not remove the significant effect of mastitis history. Standard breastfeeding literature recommends frequent, short feedings to build milk supply (18). While each additional 10 minutes per feeding increased risk of developing nipple fissures, cracks, or sores, longer feedings were not associated with increased incidence of mastitis. More frequent, rather than longer, feedings were associated with nipple fissures, cracks, or sores and mastitis incidence. Nipple trauma may result from latching the baby onto and removing the baby from the nipple. It is also possible that the act of latching on or removing the baby from the nipple results in increased exposure to potential mastitis pathogens on the mother's hands. Interestingly, washing the nipple either before or after feeding was not associated with mastitis incidence. We did not evaluate hand-washing practices. Nipple cracks and sores have been identified previously as mastitis risk factors (1, 4); cracks and sores enhance the entry of bacteria into breast tissue. Nipple cracks and sores were reported by more than one third of all participants in the first week postpartum. The observed threefold increase in mastitis incidence with mastitis history is similar to the threefold increase found among 664 women in southwest Finland who were also followed for 12 weeks. Women with a history of mastitis may have a breastfeeding style that puts them at greater risk of mastitis; indeed, they had significantly different breastfeeding practices than do women without mastitis history. Alternatively, the increase in risk with mastitis history may reflect persistent colonization with a potential pathogen. Breast anatomy may also predispose to mastitis or skin type prone to nipple cracks or sores. Siblings may nurse with similar frequency either due to genetic predisposition or to learned behaviors on the part of the mother. Mastitis caused considerable pain, worry, days in bed, and physician visits. Further, mastitis can lead to early weaning (5, 19). It was frequently diagnosed and treated over the telephone. As antibiotic resistance is an increasing concern, better algorithms to distinguish between milk stasis and mastitis without the benefit of physical examination are required. Moreover, the field would benefit from controlled studies of mastitis treatment; such studies are virtually nonexistent in the general medical literature. Our sample size was large, with very stable estimates. We had frequent, regular contact (every 3 weeks) with study participants to assess breastfeeding practices and health behaviors. We had a high initial response (99 percent) and follow-up rate (89 percent). Study participants were recruited from two sites. While Michigan women had much higher rates of breastfeeding and tended to breastfeed longer, there was no difference in mastitis incidence rate or the incidence of nipple cracks, fissures, or sores by site. Although the women in our study are not representative of all breastfeeding women because they are primarily White and middle class, the associations should be broadly generalizable, especially since we did not find any difference between sites in mastitis incidence or any associations with sociodemographic characteristics. Mastitis is a multiagent syndrome. To the best of our knowledge, the role of specific agents has yet to be described. The most common etiologic agents, S. aureus, Staphylococcus epidermidis, and Streptococcus species, inhabit either the skin or nasopharynx but are not uniformly present among all persons. For example, an estimated 20 percent of the general population are chronic carriers of S. aureus (20). Given that risk of mastitis is independent of sociodemographic variables and site, that mastitis history is a strong predictor, and that nipple sores occur frequently in the absence of infection, agent characteristics must be an important determinant of mastitis risk. Unfortunately, this study was not designed to examine the etiologic agents. Published articles on the bacteria-causing lactation mastitis are few, and most are at least 10 years old. The frequency of these bacteria in mothers' milk and on their breasts and in the nasopharynx among women with and those without mastitis deserves attention. Mastitis is common but poorly understood. Our study highlights how little is known about the epidemiology and pathogenesis of this common condition. The popularly reported risk factors explain very little. Future studies should describe the role of nasal and skin flora on mastitis risk. Editor's note: An invited commentary on this article appears on page 115. Reprint requests to Dr. Betsy Foxman, Department of Epidemiology, University of Michigan School of Public Health, 109 Observatory Street, Ann Arbor, MI 48109-2029 (e-mail: [email protected]). Supported by National Institute of Child Health and Human Development grant HD30866 to Dr. Betsy Foxman. The authors thank the Providence Family Birthing Center and the Mutual of Omaha for allowing us to identify study participants. Telephone interviews were conducted by the Survey Unit at the Institute of Social Research. REFERENCES 1. Marshall BR, Heppa JK, Zirbel CC. Sporadic puerperal mastitis: an infection that need not interrupt lactation. JAMA  1975; 233: 1377–9. Google Scholar 2. Devereux WP. Acute puerperal mastitis: evaluation of its management. Am J Obstet Gynecol  1970; 108: 78–81. Google Scholar 3. Lawrence RA. Breastfeeding: a guide for the medical profession. St Louis, MO: C. V. Mosby, 1989. Google Scholar 4. Jonsson S, Pulkkinen MO. Mastitis today: incidence, prevention and treatment. Ann Chir Gynaecol Suppl  1994; 208: 84–7. Google Scholar 5. Fetherston C. Characteristics of lactation mastitis in a Western Australian cohort. Breastfeed Rev  1997; 52: 5–11. Google Scholar 6. Fulton AA. Incidence of puerperal and lactational mastitis in an industrial town of some 43,000 inhabitants. Br Med J  1945; May 19: 693–6. Google Scholar 7. Matheson I, Aursnes I, Horgen M, et al. Bacteriological findings and clinical symptoms in relation to clinical outcome in puerperal mastitis. Acta Obstet Gynecol Scand  1988; 67: 723–6. Google Scholar 8. Niebyl JR, Spence MR, Parmley TH. Sporadic nonepidemic puerperal mastitis. J Reprod Med  1978; 20: 97–100. Google Scholar 9. Thomsen AC, Espersen T, Maigaard S. Course and treatment of milk stasis, noninfectious inflammation of the breast, and infectious mastitis in nursing women. Am J Obstet Gynecol  1984; 149: 492–5. Google Scholar 10. Dilts CL. Nursing management of mastitis due to breastfeeding. J Obstet Gynecol Neonatal Nurs  1985; 14: 286–8. Google Scholar 11. Riordan JM, Nichols FH. A descriptive study of lactation mastitis in long-term breastfeeding women. J Human Lact  1990; 6: 53–8. Google Scholar 12. d'Arcy H, Gillespie B, Foxman B. Respiratory symptoms in mothers of young children: Association with outside childcare. Pediatrics  2000; 106: 1013–16. Google Scholar 13. Rothman KJ, Greenland S. Modern epidemiology. Second ed. Philadelphia, PA: Lippincott-Raven, 1998. Google Scholar 14. Sullivan KM, Foster DA. dEPID A Program for stratified and standardized analysis. Version 2.1. October 30 (shareware), 1987. Google Scholar 15. Allison PD. Logistic regression using the SAS System. Cary, NC: SAS Institute, Inc, 1999. Google Scholar 16. SAS Institute, Inc. SAS/STAT user's guide. Version 8. Cary, NC: SAS Institute Inc, 1999. Google Scholar 17. Kinlay JR. O'Connell DL. Kinlay S. Incidence of mastitis in breastfeeding women during the six months after delivery: a prospective cohort study. Med J Aust  1998; 169: 310–12. Google Scholar 18. Klaus MH. The frequency of suckling. A neglected but essential ingredient of breast-feeding. Obstet Gynecol Clin North Am  1987; 14: 623–33. Google Scholar 19. Schwartz K, d'Arcy HJ, Gillespie B, et al. Factors associated with weaning in the first three months post-partum. (In press). Google Scholar 20. Kluytmans J, van Belkum A, Verbrugh H. Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin Microbiol Rev  1997; 10: 505–20. Google Scholar http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Epidemiology Oxford University Press

Lactation Mastitis: Occurrence and Medical Management among 946 Breastfeeding Women in the United States

Download PDF
12 pages

Loading next page...
 
/lp/oxford-university-press/lactation-mastitis-occurrence-and-medical-management-among-946-T54TUM20my

References (22)

Publisher
Oxford University Press
ISSN
0002-9262
eISSN
1476-6256
DOI
10.1093/aje/155.2.103
Publisher site
See Article on Publisher Site

Abstract

Abstract In 1994–1998, the authors followed 946 breastfeeding women from Michigan and Nebraska for the first 3 months postpartum or until they stopped breastfeeding to describe mastitis incidence, mastitis treatment, and any associations between mastitis occurrence and hypothesized host characteristics and behaviors. Participants were interviewed by telephone at 3, 6, 9, and 12 weeks postpartum or until they ceased breastfeeding. A total of 9.5% reported provider-diagnosed lactation mastitis at least once during the 12-week period, with 64% diagnosed via telephone. After adjustment in a logistic regression model, history of mastitis with a previous child (odds ratio (OR) = 4.0, 95% confidence interval (CI): 2.64, 6.11), cracks and nipple sores in the same week as mastitis (OR = 3.4, 95% CI: 2.04, 5.51), using an antifungal nipple cream (presumably for nipple thrush) in the same 3-week interval as mastitis (OR = 3.4, 95% CI: 1.37, 8.54), and (for women with no prior mastitis history) using a manual breast pump (OR = 3.3, 95% CI: 1.92, 5.62) strongly predicted mastitis. Feeding fewer than 10 times per day was protective regardless of whether or not feeding frequency in the same week or the week before mastitis was included in the model (for the same week: 7–9 times: OR = 0.6, 95% CI: 0.41, 1.01; ≤6 times: OR = 0.4, 95% CI: 0.19, 0.82). Duration of feeding was not associated with mastitis risk. breast feeding, human body, lactation, mastitis, risk factors CI, confidence interval, OR, odds ratio, RR, rate ratio Lactation mastitis is a cellulitis of the interlobular connective tissue within the mammary gland that usually occurs in the first 6 weeks postpartum (1, 2). The clinical spectrum ranges from focal inflammation with minimal systemic symptoms to abscess and septicemia. Abscesses are reported to occur in 11 percent of all affected women (2). Pain and discomfort associated with mastitis or concerns about treatment with antibiotics that pass through the breast milk may lead some women to cease breastfeeding. The signs and symptoms include fever of 38.5°C or greater; flu-like aches and chills; and a red, tender, hot, swollen, wedge-shaped area of the breast (3). In prospective studies of breastfeeding women, the incidence of lactation mastitis per infant nursed has been reported to be as low as 2.5 percent among United States women (1), 24 percent among Finnish women (4), and 27.1 percent among Australian women (5). In a population-based study conducted in Scotland in the pre-antibiotic era, the incidence was 8.9 percent (6). The agents most frequently cultured are Staphylococcus aureus or coagulase-negative staphylococci (3, 7). Several investigators (2, 3, 8) suggest a fissure in the nipple as a route of infection. Other proposed routes include through the lactiferous ducts into a lobule or by hematogenous spread (2, 3). As there have been few formal studies of lactation mastitis, most hypothesized risk factors are based on clinical impression. Risk factors fall into two general categories: poor breastfeeding technique (1–3, 8, 9) and lowered immune status secondary to stress and sleep deprivation (10). Poor breastfeeding technique may lead to poor drainage of a duct, insufficient emptying of the breast, milk stasis, and cracks or fissures of the nipple. In the one study that formally examined these factors, however, they occurred relatively infrequently: Milk stasis due to weaning or missed feeding preceded nine of the 65 infections, and in eight, the nipple of the involved breast was fissured (1). The number of nipple fissures in the unaffected breast and in women without mastitis was not reported nor was the incidence of milk stasis in women without mastitis. Increased stress and sleep deprivation, which are often complaints of new mothers, have also been reported as risk factors (3). In a retrospective study among participants at a breastfeeding conference, women with a history of mastitis associated their mastitis with fatigue, stress, a plugged duct, change in number of feedings, engorgement/stasis, an infection in the family, breast trauma, and poor diet (11). These findings are suggestive only; the study was retrospective, the occurrence of these factors among women without mastitis was not reported, and the population had a strong commitment to breastfeeding. We followed 946 breastfeeding women for the first 3 months postpartum or until they stopped breastfeeding to describe incidence of mastitis, mastitis treatment, and any associations between mastitis occurrence and hypothesized host characteristics and behaviors. MATERIALS AND METHODS Study protocol Study participants were recruited from women who gave birth at a family birthing center in Michigan and from pregnant women working at a single large company in Nebraska, which had recently implemented a breastfeeding support program for employees. Together, the two sites gave geographic, employment, and socioeconomic status diversity. Enrollment began in Michigan during July 1994, and in Nebraska in November 1994; the last interview was in March 1998. At both sites, potential participants were given information about the study during their pregnancy and a postcard to indicate refusal. The only requirement for enrollment was an intention to breastfeed. All women who did not refuse and who delivered a live infant were sent a letter describing the study and a calendar/diary designed to be used as a memory aid during study interviews. The calendar showed nursing positions and provided spaces to record information pertinent to the study. Women could participate more than once; 43 women did so. There were five sets of twins and one set of triplets. Most potential participants were contacted by telephone at 3 weeks postpartum, and their verbal consent was obtained at the time of the interview. Participants were interviewed at 3, 6, 9 and 12 weeks postpartum or until they ceased breastfeeding. After each interview, they were sent a small incentive, such as video coupons. Fifty-six women missed the first interview and were enrolled at week 6. Interviews were conducted by using computer-assisted technology that enforced skip patterns, checked answers for appropriate ranges, and allowed for textual comments. Interviewers received 4 hours of training, were monitored periodically for adherence to the study protocol, and were retrained annually. Each participant had the same female interviewer for all four interviews. The study protocol was approved by the institutional review boards of all participating institutions. Measuring instruments All interviews included questions on breastfeeding habits, breast-related symptoms, and breast care, including use of breast pumps, bottle-feeding, mastitis, sleeping habits, depression, smoking and drinking habits, and household duties. Questions about the infant's birth; history of mastitis with a previous child, including date of diagnosis; other medical history; and sociodemographic information were asked at enrollment. As much as possible, subjects were asked to recall breastfeeding information on a week-by-week basis. Case definition Mastitis was defined by self-report of mastitis diagnosed by a health care provider. The use of a more specific definition, which included antibiotic prescription and symptoms, did not change the results, and thus, we opted to use the more sensitive definition. Because of an error in the computer-assisted interviewing program, 44 women who ceased breastfeeding sometime during the 3 weeks prior to a study interview were not asked about the occurrence of mastitis. Once the error was discovered, a validation study was conducted; this study is described in detail elsewhere (12) (Gillespie et al., University of Michigan, manuscript submitted for publication). Briefly, we recontacted all women whom the error impacted, all known mastitis cases, and a sample of the remaining participants. Recall of mastitis was found to be quite reliable and was unrelated to age, race, income, parity, history of mastitis, and study site. We detected one additional mastitis case among those in whom the error occurred. This case was included in the analysis to the extent possible. Statistical analysis We describe crude associations between exposures and rate of mastitis with rate ratios and cumulative incidence (risk) ratios and their approximate 95 percent confidence intervals (13). Rate ratios were calculated for time-dependent covariates, for example, breastfeeding habits, and cumulative incidence ratios for covariates that did not change, for example, marital status. Relative risks and confidence intervals were calculated using dEPID software (14). We tested differences between groups for categorical variables with chi-square tests and for continuous variables with Student's t tests, with a two-sided significance level of α = 0.05. We modeled the time to first mastitis as a function of baseline and time-varying covariates. Discrete-time survival analysis was used, with a time-dependent logistic regression model (15) that included site and breastfeeding variables. We estimated the association between hypothesized risk factors in the week of mastitis with mastitis risk by using rate ratios. In cases in which the hypothesized risk factor could have resulted from the mastitis or early signs of mastitis (e.g., cracks and nipple sores), we also calculated (by using the same model) the association of mastitis with the hypothesized variables in the week prior to mastitis. Because women could enter the study more than once, we analyzed the data both considering all births by a single mother as independent and taking into account clustering utilizing generalized estimating equations using the SAS program, Proc Genmod (16). The results were very similar in both analyses, so we present them taking clustering into account. RESULTS Characteristics of the study population A total of 1,057 women were identified for study. One hundred could not be contacted after multiple attempts because of improper contact information, no answer, no response to messages left on telephone answering machines, etc. Of those initially contacted, 11 refused, for a response rate of 99 percent. Of the 946 contacted, consenting women, 112 (11.8 percent) had already ceased breastfeeding at the time of the initial interview and provided 3-week interview data only. Of the 834 women still breastfeeding at 3 weeks, 801 (96.0 percent) completed at least one additional interview, and 658 (78.9 percent) completed all four interviews. Overall, 840 of 946 (88.8 percent) were followed for 12 weeks or until they ceased breastfeeding. The majority (n = 711 (75.2 percent)) of study participants were enrolled from Michigan. Michigan participants tended to be slightly older, to have more education, and to have a slightly higher income than Nebraska participants (table 1). The majority of participants at both sites were White and living with their spouse or partner. Participants from Nebraska were more likely to be first-time mothers and, if they had children, to have fewer children living with them. TABLE 1. Demographic characteristics of 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998* Characteristics  Nebraska   Michigan   No.  %  No.  %  Overall  235  25  711  75  Age (years)           18–19  2  1  10  1   20–24  17  7  49  7   25–29  90  40  218  31   30–34  91  40  257  37   35–39  25  11  139  20   ≥40  2  1  32  5  Race           White  211  90  637  90   Black  14  6  45  6   Other  9  4  26  4  Marital status           Married  221  94  683  96   Living with partner  2  1  9  1   Separated  11  5  17  2  Hours worked outside the home when pregnant           None  1  0  252  36   <20  2  1  91  13   21–39  39  17  145  21   ≥40  192  82  214  30  Education           Less than high school  67  29  108  15   High school diploma  54  23  162  23   College degree  91  39  330  47   Professional or graduate training  23  10  110  16  Total family income (dollars)           <$25,000  33  14  78  12   $25,000–$49,999  88  38  235  35   $50,000–$99,999  103  45  319  47   ≥$100,000  6  3  48  7  Parity (no. of livebirths)           1  103  44  201  28   2  84  36  239  34   3  35  15  164  23   4  8  3  64  9   ≥5  4  2  42  6  Characteristics  Nebraska   Michigan   No.  %  No.  %  Overall  235  25  711  75  Age (years)           18–19  2  1  10  1   20–24  17  7  49  7   25–29  90  40  218  31   30–34  91  40  257  37   35–39  25  11  139  20   ≥40  2  1  32  5  Race           White  211  90  637  90   Black  14  6  45  6   Other  9  4  26  4  Marital status           Married  221  94  683  96   Living with partner  2  1  9  1   Separated  11  5  17  2  Hours worked outside the home when pregnant           None  1  0  252  36   <20  2  1  91  13   21–39  39  17  145  21   ≥40  192  82  214  30  Education           Less than high school  67  29  108  15   High school diploma  54  23  162  23   College degree  91  39  330  47   Professional or graduate training  23  10  110  16  Total family income (dollars)           <$25,000  33  14  78  12   $25,000–$49,999  88  38  235  35   $50,000–$99,999  103  45  319  47   ≥$100,000  6  3  48  7  Parity (no. of livebirths)           1  103  44  201  28   2  84  36  239  34   3  35  15  164  23   4  8  3  64  9   ≥5  4  2  42  6  * The two sites were significantly different with respect to age, working during pregnancy, education of the respondent, parity (p < 0.01), and income (p = 0.05). Numbers may not total 946 because of missing values. View Large Mastitis incidence During the 12-week study period, 77 women (8.1 percent) had one case of self-reported, health care provider-diagnosed mastitis; 12 (1.3 percent) had two cases, and one (0.1 percent) had three cases, for an overall incidence of first mastitis of 9.5 percent. The cumulative incidence of a first case of mastitis was 10.3 percent in Michigan and 7.2 percent in Nebraska. There was no statistically significant difference in incidence by site. Mastitis incidence was highest in the first few weeks postpartum and fell gradually thereafter (figure 1). Cracks and nipple sores occurred at a much higher frequency than mastitis, with the highest incidence in the first (36 percent) and second (14 percent) weeks postpartum. Women who reported a history of mastitis with a previous child were more likely to have mastitis with the current baby in every follow-up period (figure 2). FIGURE 1. View largeDownload slide Incidence of mastitis and nipple cracks or sores by week postpartum, Michigan and Nebraska, 1994–1998. FIGURE 1. View largeDownload slide Incidence of mastitis and nipple cracks or sores by week postpartum, Michigan and Nebraska, 1994–1998. FIGURE 2. View largeDownload slide Mastitis incidence, with the upper limits of the 95% confidence intervals shown by “whisker”, by 3-week postpartum interval and mastitis history, Michigan and Nebraska, 1994–1998. FIGURE 2. View largeDownload slide Mastitis incidence, with the upper limits of the 95% confidence intervals shown by “whisker”, by 3-week postpartum interval and mastitis history, Michigan and Nebraska, 1994–1998. Mastitis diagnosis, clinical presentation, disease impact, and treatment Women usually were diagnosed by their obstetrician/gynecologist (38 percent), but also by family practitioners (21 percent), nurses (23 percent), and others (18 percent). The majority (64 percent) were diagnosed after a telephone contact with their health care provider and did not receive a physical examination. Women diagnosed over the telephone and those diagnosed in person were similar with respect to symptoms and sociodemographic variables (data not shown). The most common mastitis symptoms were breast tenderness (98 percent), fever (82 percent), malaise (87 percent), chills (78 percent), redness (78 percent), and a hot spot (a localized area of warmth and tenderness on the affected breast) (62 percent). Women averaged 4.9 days when they had one or more symptoms (range, 1–21 days). These symptoms were severe enough to cause the vast majority of women to restrict their activities at least somewhat (77 percent) and to spend at least 1 day in bed (72 percent). Almost half of the women changed their breastfeeding habits, either feeding more often (36 percent), less often (11 percent), feeding more often (49 percent) or less often (8 percent) on the affected breast, changing nursing holds more frequently (33 percent), trying a new nursing hold (12 percent), or making some other change (19 percent). Most women with mastitis (88 percent) were prescribed one or more medications, either antibiotics (86 percent) and/or analgesics (17 percent); two woman reported receiving an antifungal medication, and another did not know the name of the drug prescribed. The most commonly prescribed antibiotics were cephalexin (46 percent), amoxicillin (7 percent), ampicillin (7 percent), and augmentin (7 percent). No cultures were performed. In addition, health care providers suggested applying hot compresses (83 percent), fully emptying the affected breast (65 percent), feeding more frequently (74 percent), and changing feeding positions often (48 percent). We did not ask whether women were advised to stop breastfeeding. Associations of risk factors with cumulative incidence of mastitis and cracks and nipple sores The overall cumulative incidence (risk) of mastitis among those who had never breastfed previously was 7.3 percent compared with 10.8 percent among those who had breastfed previously. For women with a history of mastitis, the risk was 23.9 percent compared with 8.3 percent among first-time mothers (rate ratio (RR) = 2.9, 95 percent confidence interval (CI): 1.79, 4.69) (table 2). Women with a history of mastitis were also more likely to report cracks and nipple sores (RR = 1.3, 95 percent CI: 1.04, 1.50). Neither the cumulative incidence of mastitis nor that of cracks and nipple sores was associated with age, income, returning to work, or participating in a breastfeeding support group (table 2). Women who reported having outside help were less likely to report cracks or nipple sores (odds ratio (OR) = 0.8, 95 percent CI: 0.71, 0.97), but having outside help was not associated with mastitis. TABLE 2. Relative risk (cumulative incidence ratio) of lactation mastitis and cracks and nipple sores by mastitis history and sociodemographic characteristics for 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998 Predictor  Mastitis   Cracks and nipple sores   RR*  95% CI*  RR  95% CI  Mastitis history           First-time mother (n = 303)  Reference    Reference     Parity >1, never breastfed before (n = 51)  0.2  0.03, 1.72  0.5  0.30, 0.83   Breastfed before, no mastitis (n = 458)  0.9  0.51, 1.40  0.9  0.77, 1.05   Mastitis in the past (n = 134)  2.9  1.79, 4.69  1.3  1.04, 1.50  Site           Nebraska (n = 235)  0.7  0.42, 1.17  0.8  0.64, 0.93   Michigan (n = 711)  Reference    Reference    Age (years) (missing for 14 women)          18–24 (n = 78)  Reference    Reference     25–29 (n = 308)  1.6  0.63, 3.91  0.7  0.55, 0.91   30–34 (n = 348)  1.4  0.58, 3.56  0.9  0.70, 1.11   ≥35 (n = 198)  1.7  0.68, 4.42  1.0  0.76, 1.25  Income (dollars) (missing for 36 women)           <$25,000 (n = 111)  Reference    Reference     $25,000–$49,999 (n = 323)  1.0  0.50, 1.98  1.1  0.88, 1.47   $50,000–$99,999 (n = 422)  0.9  0.47, 1.80  1.1  0.87, 1.43   ≥$100,000 (n = 54)  1.4  0.58, 3.57  1.0  0.67, 1.47  Marital status (missing for 3 women)           Not married (n = 39)  No mastitis in those not married    0.7  0.46, 1.14   Married (n = 904)  Reference    Reference    Ethnicity (missing for 4 women)           Non-White (n = 94)  0.8  0.37, 1.62  0.7  0.50, 0.92   White (n = 848)  Reference    Reference    Education (missing for 1 woman)           Less than high school (n = 175)  Reference    Reference     High school diploma (n = 216)  1.2  0.60, 2.22  1.2  0.93, 1.54  College degree (n = 421)  1.2  0.65, 2.08  1.4  1.09, 1.70   Professional or graduate degree (n = 133)  1.6  0.82, 3.12  1.4  1.08, 1.81  Children living at home (missing for 3 women)           1 (n = 293)  Reference    Reference     2 (n = 330)  1.1  0.69, 1.87  1.0  0.83, 1.16   3 (n = 198)  1.3  0.76, 2.24  0.9  0.72, 1.09   4 (n = 75)  0.9  0.40, 2.20  1.0  0.75, 1.29   ≥5 (n = 47)  1.0  0.36, 2.74  0.9  0.59, 1.24  Smoking during any week (missing for 1 woman)           Yes (n = 61)  No mastitis in smokers    1.1  0.84, 1.42   No (n = 884)  Reference    Reference    Drinking during any week (missing for 82 women)           Yes (n = 442)  1.1  0.74, 1.61  1.0  0.86, 1.12   No (n = 422)  Reference    Reference    Return to work by week 12           Yes (n = 60)  0.7  0.28, 1.90  1.0  0.74, 1.32   No (n = 886)  Reference    Reference    Breastfeeding support group (missing for 83 women)           Yes (n = 96)  1.4  0.78, 2.33  1.0  0.81, 1.24   No (n = 767)  Reference    Reference    Outside help in any wave (missing for 1 woman)           Yes (n = 633)  1.2  0.79, 1.87  0.8  0.71, 0.97   No (n = 312)  Reference    Reference    Predictor  Mastitis   Cracks and nipple sores   RR*  95% CI*  RR  95% CI  Mastitis history           First-time mother (n = 303)  Reference    Reference     Parity >1, never breastfed before (n = 51)  0.2  0.03, 1.72  0.5  0.30, 0.83   Breastfed before, no mastitis (n = 458)  0.9  0.51, 1.40  0.9  0.77, 1.05   Mastitis in the past (n = 134)  2.9  1.79, 4.69  1.3  1.04, 1.50  Site           Nebraska (n = 235)  0.7  0.42, 1.17  0.8  0.64, 0.93   Michigan (n = 711)  Reference    Reference    Age (years) (missing for 14 women)          18–24 (n = 78)  Reference    Reference     25–29 (n = 308)  1.6  0.63, 3.91  0.7  0.55, 0.91   30–34 (n = 348)  1.4  0.58, 3.56  0.9  0.70, 1.11   ≥35 (n = 198)  1.7  0.68, 4.42  1.0  0.76, 1.25  Income (dollars) (missing for 36 women)           <$25,000 (n = 111)  Reference    Reference     $25,000–$49,999 (n = 323)  1.0  0.50, 1.98  1.1  0.88, 1.47   $50,000–$99,999 (n = 422)  0.9  0.47, 1.80  1.1  0.87, 1.43   ≥$100,000 (n = 54)  1.4  0.58, 3.57  1.0  0.67, 1.47  Marital status (missing for 3 women)           Not married (n = 39)  No mastitis in those not married    0.7  0.46, 1.14   Married (n = 904)  Reference    Reference    Ethnicity (missing for 4 women)           Non-White (n = 94)  0.8  0.37, 1.62  0.7  0.50, 0.92   White (n = 848)  Reference    Reference    Education (missing for 1 woman)           Less than high school (n = 175)  Reference    Reference     High school diploma (n = 216)  1.2  0.60, 2.22  1.2  0.93, 1.54  College degree (n = 421)  1.2  0.65, 2.08  1.4  1.09, 1.70   Professional or graduate degree (n = 133)  1.6  0.82, 3.12  1.4  1.08, 1.81  Children living at home (missing for 3 women)           1 (n = 293)  Reference    Reference     2 (n = 330)  1.1  0.69, 1.87  1.0  0.83, 1.16   3 (n = 198)  1.3  0.76, 2.24  0.9  0.72, 1.09   4 (n = 75)  0.9  0.40, 2.20  1.0  0.75, 1.29   ≥5 (n = 47)  1.0  0.36, 2.74  0.9  0.59, 1.24  Smoking during any week (missing for 1 woman)           Yes (n = 61)  No mastitis in smokers    1.1  0.84, 1.42   No (n = 884)  Reference    Reference    Drinking during any week (missing for 82 women)           Yes (n = 442)  1.1  0.74, 1.61  1.0  0.86, 1.12   No (n = 422)  Reference    Reference    Return to work by week 12           Yes (n = 60)  0.7  0.28, 1.90  1.0  0.74, 1.32   No (n = 886)  Reference    Reference    Breastfeeding support group (missing for 83 women)           Yes (n = 96)  1.4  0.78, 2.33  1.0  0.81, 1.24   No (n = 767)  Reference    Reference    Outside help in any wave (missing for 1 woman)           Yes (n = 633)  1.2  0.79, 1.87  0.8  0.71, 0.97   No (n = 312)  Reference    Reference    * RR, relative risk; CI, confidence interval. View Large Associations of breastfeeding and other health behaviors with rate of mastitis Women with previous mastitis compared with those without this history had significantly more weeks with nipple cracks or sores, more weeks with 10 or more feedings per day, and more weeks with feedings of 10 minutes or less; used fewer positions; and were less likely to pump, bottle-feed, wash their breasts either before or after feeding, and use nursing bras. Because of the strong association of mastitis with mastitis history, we present associations separately by mastitis history. Cracks and nipple sores that occurred during the same week as mastitis were associated with an almost sixfold increase in mastitis rate for women with no history of mastitis and a threefold increase in mastitis rate for women with such a history (table 3). Women without cracks and nipple sores in the same week as mastitis had similar rates of mastitis in the right (8.8/1,000 person-weeks) as in the left (8.5/1,000 person-weeks) breast and lower rates of mastitis in both breasts at the same time (1.5/1,000 person-weeks). Women with cracks and nipple sores in one breast in the same week as mastitis had higher rates of mastitis in the breast with cracks or nipple sores (48.9/1,000 person-weeks) than in the breast without them (13.3/1,000 person-weeks). Women with cracks or nipple sores in both breasts in the same week as mastitis experienced the highest mastitis rate (88.7/1,000 person-weeks). TABLE 3. Unadjusted rate ratios for the association of breastfeeding and other health behaviors with lactation mastitis by history of mastitis, Michigan and Nebraska, 1994–1998 Characteristic  No previous mastitis   Mastitis history   Person-weeks exposed  Mastitis rate/1,000  RR*  95% CI*  Person-weeks exposed  Mastitis rate/1,000  RR  95% CI  Nipple cracks or sores†                   Same week as mastitis                1.38, 6.84    Yes  591  30.5  5.6  3.21, 9.77  146  54.8  3.1      No  7,353  5.4      1,345  17.8       Week before mastitis                0.46, 3.14    Yes  849  20.0  3.5  1.99, 6.18  204  24.5  1.2      No  7,000  5.7      1,280  20.3      Any ointment use                0.26, 4.50    Yes  545  11.0  1.6  0.67, 3.64  87  23.0  1.1      No  7,384  7.0      1,404  21.4      Antifungal ointment                0.54, 29.12    Yes  63  47.6  6.8  2.13, 21.78  12  83.3  4.0      No  7,866  7.0      1,479  21.0      Frequency of feeding (times/day)                   Same week as mastitis                   <6  1,937  2.1  0.2  0.06, 0.46  339  14.7  0.4  0.15, 1.11    7–9  3,189  7.2  0.6  0.32, 0.98  654  15.3  0.4  0.20, 0.94   ≥10  2,183  12.8      476  35.7       Week before mastitis                   <6  1,694  2.4  0.2  0.07, 0.55  292  13.7  0.4  0.14, 1.19    7–9  3,211  6.5  0.53  0.03, 0.93  642  14.0  0.4  0.19, 0.92    ≥10  2,440  12.3      530  34.0      Duration of feeding (minutes)†                   Same week as mastitis                    <10  1, 037  5.8      284  31.7        11–20  3,605  8.0  1.4  0.58, 3.35  825  17.0  0.5  0.23, 1.24    21–30  1,836  7.1  1.2  0.47, 3.22  281  24.9  0.8  0.29, 2.11    ≥31  850  9.4  1.6  0.56, 4.69  78  25.6  0.8  0.17, 3.74   Week before mastitis                    <10  1,051  7.6      285  31.6        11–20  3,586  7.0  0.9  0.41, 2.03  809  18.5  0.6  0.26, 1.34    21–30  1,831  7.1  0.9  0.39, 2.25  279  17.9  0.6  0.19, 1.69    ≥31  896  11.2  1.5  0.58, 3.71  90  22.2  0.7  0.15, 3.26  Pain when latching on                    Yes  3,084  12.6  3.2  1.86, 5.58  588  35.7  2.9  1.41, 6.08    No  4,845  3.9      903  12.2      Pain when nursing                    Yes  2,979  14.4  4.8  2.65, 8.57  597  36.9  3.3  1.56, 6.96    No  4,950  3.0      894  11.2      Pain when not nursing                    Yes  1,938  17.0  4.1  2.43, 6.86  372  43.0  3.0  1.50, 6.01    No  5,991  4.2      1,119  14.2      Pain all the time                    Yes  381  21.0  3.2  1.50, 6.67  126  71.4  4.2  1.96, 2.91    No  7,548  6.6      1,365  1.7      Any breast pain in the week before mastitis                    Yes  1,540  8.4  1.2  0.65, 2.25  277  18.1  0.8  0.32, 3.18    No  6,305  7.0      1,205  21.6      No. of positions used†                    1  2,022  5.4      288  6.9        2  3,678  6.0  1.1  0.52, 2.22  690  15.9  2.3  0.51, 10.36    ≥3  4,433  5.6  1.0  0.51, 2.11  516  36.8  5.3  1.24, 22.76  Pumping‡                   Same week as mastitis                    Any  3,346  10.8  2.2  1.32, 3.81  478  20.9  1.0  0.46, 2.03    Electric pump  1,948  9.2  1.9  1.03, 3.59  198  25.3  1.2  0.44, 3.07    Manual pump  956  15.7  3.3  1.70, 6.30  195  15.4  0.7  0.21, 2.37    By hand  722  8.3  1.7  0.70, 4.27  126  15.9  0.7  0.17, 3.11    No pumping  4,583  4.8  Reference    1,013  21.7       Week before mastitis                    Any  3,265  7.7  1.1  0.65, 1.85  469  21.3  1.0  0.46, 2.19    Electric pump  1,850  6.5  0.9  0.48, 1.81  187  21.4  1.0  0.35, 3.01    Manual pump  949  11.6  1.7  0.84, 3.30  187  16.0  0.8  0.23, 2.60    By hand  750  5.3  0.8  0.27, 2.16  134  22.4  1.1  0.32, 3.63    No pumping  4,587  7.0      1,015  20.7      Bottle fed                    Yes  3,273  6.7  0.9  0.51, 1.48  445  18.0  0.8  0.35, 1.74    No  4,653  7.7      1,046  22.9      Engorgement                    Yes  564  12.4  1.8  0.81, 3.94  89  22.5  1.0  0.25, 4.26    No  7,350  6.9      1,402  21.4      Washed nipples before feeding                    Yes  198  15.2  2.1  0.67, 6.81  15  0§  1.5  0.09, 24.72    No  7,731  7.1      1,476  21.7      Washed nipples after feeding                    Yes  152  13.2  1.8  0.45, 7.49  14  0§  1.6  0.10, 26.50    No  7,777  7.2      1,477  21.7      Used nursing pads                    Yes  4,767  8.6  1.7  0.87, 3.16  927  22.7  1.1  0.51, 2.19    No  2,319  5.2      513  21.4      Wore a nursing bra                    Yes  6,396  7.8  1.6  0.70, 3.80  1,272  22.8  1.4  0.42, 4.52    Some  276  7.2  1.5  0.31, 7.52  39  0.0  0.7  0.03, 12.55    No  1,257  4.8      180  16.7      Took daytime naps                   Same week as mastitis                    Yes  4,358  9.4  1.8  1.01, 3.13  771  23.3  1.2  0.75, 1.90    No  3,214  5.3      713  19.6       Week before mastitis                    Yes  4,681  8.5  1.4  0.82, 2.54  833  22.8  1.2  0.59, 2.52    No  2,862  5.9      644  18.6      Characteristic  No previous mastitis   Mastitis history   Person-weeks exposed  Mastitis rate/1,000  RR*  95% CI*  Person-weeks exposed  Mastitis rate/1,000  RR  95% CI  Nipple cracks or sores†                   Same week as mastitis                1.38, 6.84    Yes  591  30.5  5.6  3.21, 9.77  146  54.8  3.1      No  7,353  5.4      1,345  17.8       Week before mastitis                0.46, 3.14    Yes  849  20.0  3.5  1.99, 6.18  204  24.5  1.2      No  7,000  5.7      1,280  20.3      Any ointment use                0.26, 4.50    Yes  545  11.0  1.6  0.67, 3.64  87  23.0  1.1      No  7,384  7.0      1,404  21.4      Antifungal ointment                0.54, 29.12    Yes  63  47.6  6.8  2.13, 21.78  12  83.3  4.0      No  7,866  7.0      1,479  21.0      Frequency of feeding (times/day)                   Same week as mastitis                   <6  1,937  2.1  0.2  0.06, 0.46  339  14.7  0.4  0.15, 1.11    7–9  3,189  7.2  0.6  0.32, 0.98  654  15.3  0.4  0.20, 0.94   ≥10  2,183  12.8      476  35.7       Week before mastitis                   <6  1,694  2.4  0.2  0.07, 0.55  292  13.7  0.4  0.14, 1.19    7–9  3,211  6.5  0.53  0.03, 0.93  642  14.0  0.4  0.19, 0.92    ≥10  2,440  12.3      530  34.0      Duration of feeding (minutes)†                   Same week as mastitis                    <10  1, 037  5.8      284  31.7        11–20  3,605  8.0  1.4  0.58, 3.35  825  17.0  0.5  0.23, 1.24    21–30  1,836  7.1  1.2  0.47, 3.22  281  24.9  0.8  0.29, 2.11    ≥31  850  9.4  1.6  0.56, 4.69  78  25.6  0.8  0.17, 3.74   Week before mastitis                    <10  1,051  7.6      285  31.6        11–20  3,586  7.0  0.9  0.41, 2.03  809  18.5  0.6  0.26, 1.34    21–30  1,831  7.1  0.9  0.39, 2.25  279  17.9  0.6  0.19, 1.69    ≥31  896  11.2  1.5  0.58, 3.71  90  22.2  0.7  0.15, 3.26  Pain when latching on                    Yes  3,084  12.6  3.2  1.86, 5.58  588  35.7  2.9  1.41, 6.08    No  4,845  3.9      903  12.2      Pain when nursing                    Yes  2,979  14.4  4.8  2.65, 8.57  597  36.9  3.3  1.56, 6.96    No  4,950  3.0      894  11.2      Pain when not nursing                    Yes  1,938  17.0  4.1  2.43, 6.86  372  43.0  3.0  1.50, 6.01    No  5,991  4.2      1,119  14.2      Pain all the time                    Yes  381  21.0  3.2  1.50, 6.67  126  71.4  4.2  1.96, 2.91    No  7,548  6.6      1,365  1.7      Any breast pain in the week before mastitis                    Yes  1,540  8.4  1.2  0.65, 2.25  277  18.1  0.8  0.32, 3.18    No  6,305  7.0      1,205  21.6      No. of positions used†                    1  2,022  5.4      288  6.9        2  3,678  6.0  1.1  0.52, 2.22  690  15.9  2.3  0.51, 10.36    ≥3  4,433  5.6  1.0  0.51, 2.11  516  36.8  5.3  1.24, 22.76  Pumping‡                   Same week as mastitis                    Any  3,346  10.8  2.2  1.32, 3.81  478  20.9  1.0  0.46, 2.03    Electric pump  1,948  9.2  1.9  1.03, 3.59  198  25.3  1.2  0.44, 3.07    Manual pump  956  15.7  3.3  1.70, 6.30  195  15.4  0.7  0.21, 2.37    By hand  722  8.3  1.7  0.70, 4.27  126  15.9  0.7  0.17, 3.11    No pumping  4,583  4.8  Reference    1,013  21.7       Week before mastitis                    Any  3,265  7.7  1.1  0.65, 1.85  469  21.3  1.0  0.46, 2.19    Electric pump  1,850  6.5  0.9  0.48, 1.81  187  21.4  1.0  0.35, 3.01    Manual pump  949  11.6  1.7  0.84, 3.30  187  16.0  0.8  0.23, 2.60    By hand  750  5.3  0.8  0.27, 2.16  134  22.4  1.1  0.32, 3.63    No pumping  4,587  7.0      1,015  20.7      Bottle fed                    Yes  3,273  6.7  0.9  0.51, 1.48  445  18.0  0.8  0.35, 1.74    No  4,653  7.7      1,046  22.9      Engorgement                    Yes  564  12.4  1.8  0.81, 3.94  89  22.5  1.0  0.25, 4.26    No  7,350  6.9      1,402  21.4      Washed nipples before feeding                    Yes  198  15.2  2.1  0.67, 6.81  15  0§  1.5  0.09, 24.72    No  7,731  7.1      1,476  21.7      Washed nipples after feeding                    Yes  152  13.2  1.8  0.45, 7.49  14  0§  1.6  0.10, 26.50    No  7,777  7.2      1,477  21.7      Used nursing pads                    Yes  4,767  8.6  1.7  0.87, 3.16  927  22.7  1.1  0.51, 2.19    No  2,319  5.2      513  21.4      Wore a nursing bra                    Yes  6,396  7.8  1.6  0.70, 3.80  1,272  22.8  1.4  0.42, 4.52    Some  276  7.2  1.5  0.31, 7.52  39  0.0  0.7  0.03, 12.55    No  1,257  4.8      180  16.7      Took daytime naps                   Same week as mastitis                    Yes  4,358  9.4  1.8  1.01, 3.13  771  23.3  1.2  0.75, 1.90    No  3,214  5.3      713  19.6       Week before mastitis                    Yes  4,681  8.5  1.4  0.82, 2.54  833  22.8  1.2  0.59, 2.52    No  2,862  5.9      644  18.6      * RR, relative risk; CI, confidence interval. † Rate ratios for no previous mastitis and mastitis history were significantly different for nipple cracks and sores, feeding for 11–20 minutes, and number of positions used. ‡ Categories are not mutually exclusive. All comparisons are to no pumping. § For tables with zero cells, odds ratios were calculated by adding 1/2 to all cells. View Large Cracks and nipple sores may be treated with ointments; however, there was little association with any ointment use during the same week as mastitis. There was a twofold increase (RR = 2.0, 95 percent CI: 1.00, 4.17) for any ointment use in the week before mastitis among women with no history of mastitis. Using an antifungal ointment during the same week as mastitis, but not during the previous week, was statistically significantly associated with mastitis among women with no mastitis history (RR = 6.8, 95 percent CI: 2.13, 21.78). We do not know whether the antifungal creams were self- or physician-prescribed. Average frequency of feeding per day, but not duration of feeding, was associated with mastitis rate; this was true whether or not the feeding habits were in the same week or the week before mastitis. Women without a history of mastitis who fed six or fewer times a day had a rate of mastitis five times lower than those who fed 10 or more times a day (same week as mastitis: RR = 0.2, 95 percent CI: 0.06, 0.46; week before mastitis: RR = 0.2, 95 percent CI: 0.07, 0.55); for women with a history of mastitis, the rate was 2.5 times lower (same week as mastitis: RR = 0.4, 95 percent CI: 0.15, 1.11; week before mastitis: RR = 0.4, 95 percent CI: 0.14, 1.19) (figure 3). FIGURE 3. View largeDownload slide Three-month incidence of mastitis, with the upper limits of the 95 percent confidence intervals shown by “whisker”, by average number of feedings per day and mastitis history, Michigan and Nebraska, 1994–1998. FIGURE 3. View largeDownload slide Three-month incidence of mastitis, with the upper limits of the 95 percent confidence intervals shown by “whisker”, by average number of feedings per day and mastitis history, Michigan and Nebraska, 1994–1998. Breast pain when latching on, nursing, not nursing, or all the time also were associated with an increased rate of mastitis; however, if lagged by 1 week, they were not associated, suggesting that the breast pain may have been a symptom of mastitis rather than a risk factor of acquisition. Similarly, women who reported using more nursing positions had a higher rate of mastitis among those with a history of mastitis, but an increased number of positions likely reflects a response to nipple cracks or sores, breast pain, or mastitis. This variable could not be lagged because it was not asked on a week-by-week basis. Among women with no history of mastitis, use of a breast pump, particularly a manual pump, resulted in an increased rate of mastitis. A history of using a breast pump for the previous week was also associated with mastitis, but the rate ratio was reduced and was no longer statistically significant; for other breast pumps, use of a pump during the previous week also noticeably reduced the association. For women with a history of mastitis, using a pump had no association with mastitis rate, with the point estimates for manual pumping suggesting a protective effect for both the same and the previous week. Other breastfeeding behaviors, including bottle-feeding, engorgement, washing breasts before or after feeding, using nursing pads, and wearing a nursing bra had little association with mastitis rate; none of the observed associations were statistically significant. Women who reported taking daytime naps had higher rates of mastitis, but, again, this result could be an effect rather than a cause of mastitis. When information from the previous week was used, there was still a modest increase in the rate for women with no history of mastitis, but it was no longer statistically significant. Discrete survival regression analysis To explore further the relations between variables identified in the bivariate and stratified analyses, we fit a time-dependent logistic regression model (table 4), including mastitis history, weeks since birth, and the following variables measured in the same week as mastitis: nipple sores or cracks, antifungal nipple cream, frequency of feeding, and use of a manual breast pump. Duration of feeding and daytime naps were not associated with risk of mastitis after adjustment for other variables and, thus, were not included in the model. After adjustment for covariates and clustering between responses for women who participated more than once, mastitis history, cracks and nipple sores in the same week as mastitis, and use of antifungal nipple cream in the same 3-week interval as mastitis remained positively associated with mastitis, increasing risk more than threefold. Those who fed fewer than 10 times per day in the same week as mastitis were at significantly decreased risk of mastitis, with those feeding fewer than six times per day having a risk 2.5 times lower than women feeding 10 or more times a day. Use of a manual breast pump in women without mastitis history was also significantly associated with mastitis rate. The interaction with mastitis history (p = 0.02) indicates that those with mastitis history do not share the increased risk with manual breast pump use: Using a manual breast pump increased risk of mastitis 2.1 times (95 percent CI: 1.09, 3.86) among women with no mastitis history, but had no association with mastitis among women who had experienced mastitis with a previous baby. When the model was fit using covariates from the previous week, the results were similar, although cracks and nipple sores were only marginally statistically significant (p = 0.06). The interaction between mastitis history and use of a manual breast pump was also no longer statistically significant. In both models, the time variables showed mastitis rates to be significantly higher in weeks 1–6 relative to weeks 10–12. TABLE 4. Odds ratios and 95 percent confidence intervals for the associations of breastfeeding practices and other health behaviors with lactation mastitis: results from a time-dependent logistic regression model using characteristics from the same week as mastitis and from the week before mastitis* for 946 breastfeeding women who participated in a prospective cohort study of lactation mastitis, Michigan and Nebraska, 1994–1998 Week of time-dependent covariates  OR†  95% CI†  p value  Covariates from the same week as mastitis         Mastitis history  4.0  2.64, 6.11  0.0000   Nipple cracks or sores  3.4  2.04, 5.51  0.0000   Antifungal nipple cream  3.4  1.37, 8.54  0.009   Feed (times/day)          <6  0.4  0.19, 0.82  0.01    7–9  0.6  0.41, 1.01  0.06    ≥10  Reference      Using a manual breast pump          No mastitis history  3.3  1.92, 5.62  0.0000    Mastitis history  0.9  0.21, 4.04     Manual breast pump* mastitis history (weeks)      0.02    1–3  3.3  1.41, 7.74  0.006    4–6  3.0  1.31, 6.89  0.009    7–9  2.4  1.02, 5.73  0.05    10–12  Reference      Covariates from the week before mastitis         Mastitis history  3.3  2.12, 5.16  0.0000   Nipple cracks or sores  1.7  0.96, 2.86  0.07   Antifungal nipple cream‡  3.4  1.06, 10.80  0.04   Feed (times/day)          <6  0.4  0.20, 0.89  0.02    7–9  0.6  0.40, 0.99  0.05    ≥10  Reference       Using a manual breast pump          No mastitis history  2.1  1.09, 3.86  0.03    Mastitis history  0.5  0.12, 1.65  0.23   Manual breast pump* mastitis history (weeks)          1–3  3.7  1.53, 8.96  0.004    4–6  3.0  1.22, 7.23  0.02    7–9  2.3  0.93, 5.86  0.07    10–12  Reference      Week of time-dependent covariates  OR†  95% CI†  p value  Covariates from the same week as mastitis         Mastitis history  4.0  2.64, 6.11  0.0000   Nipple cracks or sores  3.4  2.04, 5.51  0.0000   Antifungal nipple cream  3.4  1.37, 8.54  0.009   Feed (times/day)          <6  0.4  0.19, 0.82  0.01    7–9  0.6  0.41, 1.01  0.06    ≥10  Reference      Using a manual breast pump          No mastitis history  3.3  1.92, 5.62  0.0000    Mastitis history  0.9  0.21, 4.04     Manual breast pump* mastitis history (weeks)      0.02    1–3  3.3  1.41, 7.74  0.006    4–6  3.0  1.31, 6.89  0.009    7–9  2.4  1.02, 5.73  0.05    10–12  Reference      Covariates from the week before mastitis         Mastitis history  3.3  2.12, 5.16  0.0000   Nipple cracks or sores  1.7  0.96, 2.86  0.07   Antifungal nipple cream‡  3.4  1.06, 10.80  0.04   Feed (times/day)          <6  0.4  0.20, 0.89  0.02    7–9  0.6  0.40, 0.99  0.05    ≥10  Reference       Using a manual breast pump          No mastitis history  2.1  1.09, 3.86  0.03    Mastitis history  0.5  0.12, 1.65  0.23   Manual breast pump* mastitis history (weeks)          1–3  3.7  1.53, 8.96  0.004    4–6  3.0  1.22, 7.23  0.02    7–9  2.3  0.93, 5.86  0.07    10–12  Reference      * Adjusted for clustering between the 43 women who participated twice, using generalized estimating equations. † OR, odds ratio; CI, confidence interval. ‡ This variable was not lagged, as measured in 3-week intervals. View Large DISCUSSION In this prospective cohort study of 946 breastfeeding women followed for the first 3 months postpartum, 9.5 percent of the population self-reported health care provider-diagnosed lactation mastitis at least once. The strongest risk factors were history of mastitis with a previous child, cracks and nipple sores in the same week, using an antifungal nipple cream (presumably for nipple thrush) in the same week, feeding the baby more frequently and, for women with no prior mastitis history, using a manual breast pump in the same week. The incidence was essentially the same by sociodemographic variables, study site, sleeping habits, and breastfeeding practices other than frequency of feeding. This result is in contrast to a study in southwest Finland that found an increased incidence among women aged 21–35 years relative to younger and older women (4). The incidence of mastitis reported here is approximately half that reported from prospective cohorts in Australia (27 percent, 95 percent CI: 22, 32 percent, n = 306 (5)), New South Wales (20 percent, 95 percent CI: 18, 22 percent, n = 1,075 (17)), or southwest Finland (24 percent, 95 percent CI: 21, 27 percent, n = 664) (4)), but was more than three times the incidence reported in a United States cohort (2.5 percent, 95 percent CI: 1.8–3.0 percent, n = 2,534) (1) . This may be explained by differences in case definition, population, temporal variation, selection biases, or other unmeasured factors. Nevertheless, there is general agreement that mastitis is a common bacterial infection among breastfeeding women. One of the enduring myths of mastitis risk is that it results from inexperience with breastfeeding. However, in our study, women with a history of mastitis—by definition, previous breastfeeders—had a greater risk. Adjustment for the observed breastfeeding practices did not remove the significant effect of mastitis history. Standard breastfeeding literature recommends frequent, short feedings to build milk supply (18). While each additional 10 minutes per feeding increased risk of developing nipple fissures, cracks, or sores, longer feedings were not associated with increased incidence of mastitis. More frequent, rather than longer, feedings were associated with nipple fissures, cracks, or sores and mastitis incidence. Nipple trauma may result from latching the baby onto and removing the baby from the nipple. It is also possible that the act of latching on or removing the baby from the nipple results in increased exposure to potential mastitis pathogens on the mother's hands. Interestingly, washing the nipple either before or after feeding was not associated with mastitis incidence. We did not evaluate hand-washing practices. Nipple cracks and sores have been identified previously as mastitis risk factors (1, 4); cracks and sores enhance the entry of bacteria into breast tissue. Nipple cracks and sores were reported by more than one third of all participants in the first week postpartum. The observed threefold increase in mastitis incidence with mastitis history is similar to the threefold increase found among 664 women in southwest Finland who were also followed for 12 weeks. Women with a history of mastitis may have a breastfeeding style that puts them at greater risk of mastitis; indeed, they had significantly different breastfeeding practices than do women without mastitis history. Alternatively, the increase in risk with mastitis history may reflect persistent colonization with a potential pathogen. Breast anatomy may also predispose to mastitis or skin type prone to nipple cracks or sores. Siblings may nurse with similar frequency either due to genetic predisposition or to learned behaviors on the part of the mother. Mastitis caused considerable pain, worry, days in bed, and physician visits. Further, mastitis can lead to early weaning (5, 19). It was frequently diagnosed and treated over the telephone. As antibiotic resistance is an increasing concern, better algorithms to distinguish between milk stasis and mastitis without the benefit of physical examination are required. Moreover, the field would benefit from controlled studies of mastitis treatment; such studies are virtually nonexistent in the general medical literature. Our sample size was large, with very stable estimates. We had frequent, regular contact (every 3 weeks) with study participants to assess breastfeeding practices and health behaviors. We had a high initial response (99 percent) and follow-up rate (89 percent). Study participants were recruited from two sites. While Michigan women had much higher rates of breastfeeding and tended to breastfeed longer, there was no difference in mastitis incidence rate or the incidence of nipple cracks, fissures, or sores by site. Although the women in our study are not representative of all breastfeeding women because they are primarily White and middle class, the associations should be broadly generalizable, especially since we did not find any difference between sites in mastitis incidence or any associations with sociodemographic characteristics. Mastitis is a multiagent syndrome. To the best of our knowledge, the role of specific agents has yet to be described. The most common etiologic agents, S. aureus, Staphylococcus epidermidis, and Streptococcus species, inhabit either the skin or nasopharynx but are not uniformly present among all persons. For example, an estimated 20 percent of the general population are chronic carriers of S. aureus (20). Given that risk of mastitis is independent of sociodemographic variables and site, that mastitis history is a strong predictor, and that nipple sores occur frequently in the absence of infection, agent characteristics must be an important determinant of mastitis risk. Unfortunately, this study was not designed to examine the etiologic agents. Published articles on the bacteria-causing lactation mastitis are few, and most are at least 10 years old. The frequency of these bacteria in mothers' milk and on their breasts and in the nasopharynx among women with and those without mastitis deserves attention. Mastitis is common but poorly understood. Our study highlights how little is known about the epidemiology and pathogenesis of this common condition. The popularly reported risk factors explain very little. Future studies should describe the role of nasal and skin flora on mastitis risk. Editor's note: An invited commentary on this article appears on page 115. Reprint requests to Dr. Betsy Foxman, Department of Epidemiology, University of Michigan School of Public Health, 109 Observatory Street, Ann Arbor, MI 48109-2029 (e-mail: [email protected]). Supported by National Institute of Child Health and Human Development grant HD30866 to Dr. Betsy Foxman. The authors thank the Providence Family Birthing Center and the Mutual of Omaha for allowing us to identify study participants. Telephone interviews were conducted by the Survey Unit at the Institute of Social Research. REFERENCES 1. Marshall BR, Heppa JK, Zirbel CC. Sporadic puerperal mastitis: an infection that need not interrupt lactation. JAMA  1975; 233: 1377–9. Google Scholar 2. Devereux WP. Acute puerperal mastitis: evaluation of its management. Am J Obstet Gynecol  1970; 108: 78–81. Google Scholar 3. Lawrence RA. Breastfeeding: a guide for the medical profession. St Louis, MO: C. V. Mosby, 1989. Google Scholar 4. Jonsson S, Pulkkinen MO. Mastitis today: incidence, prevention and treatment. Ann Chir Gynaecol Suppl  1994; 208: 84–7. Google Scholar 5. Fetherston C. Characteristics of lactation mastitis in a Western Australian cohort. Breastfeed Rev  1997; 52: 5–11. Google Scholar 6. Fulton AA. Incidence of puerperal and lactational mastitis in an industrial town of some 43,000 inhabitants. Br Med J  1945; May 19: 693–6. Google Scholar 7. Matheson I, Aursnes I, Horgen M, et al. Bacteriological findings and clinical symptoms in relation to clinical outcome in puerperal mastitis. Acta Obstet Gynecol Scand  1988; 67: 723–6. Google Scholar 8. Niebyl JR, Spence MR, Parmley TH. Sporadic nonepidemic puerperal mastitis. J Reprod Med  1978; 20: 97–100. Google Scholar 9. Thomsen AC, Espersen T, Maigaard S. Course and treatment of milk stasis, noninfectious inflammation of the breast, and infectious mastitis in nursing women. Am J Obstet Gynecol  1984; 149: 492–5. Google Scholar 10. Dilts CL. Nursing management of mastitis due to breastfeeding. J Obstet Gynecol Neonatal Nurs  1985; 14: 286–8. Google Scholar 11. Riordan JM, Nichols FH. A descriptive study of lactation mastitis in long-term breastfeeding women. J Human Lact  1990; 6: 53–8. Google Scholar 12. d'Arcy H, Gillespie B, Foxman B. Respiratory symptoms in mothers of young children: Association with outside childcare. Pediatrics  2000; 106: 1013–16. Google Scholar 13. Rothman KJ, Greenland S. Modern epidemiology. Second ed. Philadelphia, PA: Lippincott-Raven, 1998. Google Scholar 14. Sullivan KM, Foster DA. dEPID A Program for stratified and standardized analysis. Version 2.1. October 30 (shareware), 1987. Google Scholar 15. Allison PD. Logistic regression using the SAS System. Cary, NC: SAS Institute, Inc, 1999. Google Scholar 16. SAS Institute, Inc. SAS/STAT user's guide. Version 8. Cary, NC: SAS Institute Inc, 1999. Google Scholar 17. Kinlay JR. O'Connell DL. Kinlay S. Incidence of mastitis in breastfeeding women during the six months after delivery: a prospective cohort study. Med J Aust  1998; 169: 310–12. Google Scholar 18. Klaus MH. The frequency of suckling. A neglected but essential ingredient of breast-feeding. Obstet Gynecol Clin North Am  1987; 14: 623–33. Google Scholar 19. Schwartz K, d'Arcy HJ, Gillespie B, et al. Factors associated with weaning in the first three months post-partum. (In press). Google Scholar 20. Kluytmans J, van Belkum A, Verbrugh H. Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin Microbiol Rev  1997; 10: 505–20. Google Scholar

Journal

American Journal of Epidemiology – Oxford University Press

Published: Jan 15, 2002

There are no references for this article.