Neonatal Mastitis and Concurrent Serious Bacterial Infection

Using the framework of the Pediatric Emergency Medicine Collaborative Research Committee of the American Academy of Pediatrics, we conducted a 28-site retrospective, cross-sectional study of infants clinically diagnosed with mastitis in the emergency department (ED). The participating sites included 26 pediatric EDs in the United States, 1 in Canada, and 1 in Spain. The study was approved by the Steering Committee of the Pediatric Emergency Medicine Collaborative Research Committee and the research ethics boards of all participating hospitals with a waiver of informed consent. Data sharing agreements were obtained between the lead site, all participating institutions, and the centralized data center at Baylor College of Medicine.

We included infants aged 90 days and younger diagnosed with mastitis, who presented to a participating ED between January 1, 2008, and December 31, 2017. We identified potentially eligible infants by searching the electronic health record for International Classification of Diseases (ICD) (9th or 10th revisions) diagnosis codes for neonatal mastitis (Supplemental Table 7). To ensure accuracy, once potential subjects were identified, charts were reviewed to confirm a clinical diagnosis consistent with mastitis, which required evidence of at least one of the following: surrounding erythema, induration, fluctuance, or purulent drainage. Patients were excluded if chart review suggested benign diagnoses, such as physiologic gynecomastia or galactorrhea.

We defined and described study variables a priori in a shared manual of operations. A standardized data collection form was used for all sites, and all investigators reviewed these forms for clarity and consistency before study commencement. Investigators entered deidentified data electronically into a Research Electronic Data Capture tool (Vanderbilt University, Nashville, TN), a secure, Web-based electronic database. Data collected included demographic information, medical history, presenting symptoms, ED vital signs and physical examination findings, laboratory values, and ED and inpatient management (if hospitalized). We defined history of fever at home as any measured temperature ≥38°C or use of terms such as “felt warm,” “tactile temperature,” or any other subjective term of fever. These variables were classified as yes, no, or unclear or not documented.

For more potentially subjective descriptions from the electronic health record, we provided restrictive key words to define general appearance on examination, clinical diagnoses of sepsis or shock, and the presence of necrotizing or extensive soft tissue infection. Infants were classified as well if documentation included the terms “well appearing,” “no apparent distress,” “alert,” “normal mental status,” “interactive,” “playful,” or other similar terms and unwell if documentation included “sick,” “toxic,” “shocky,” “decreased mental status,” “lethargic,” “unresponsive,” “irritable,” “fussy,” “inconsolable,” “not looking well,” “poor or decreased pulses,” or other similar terms. General appearance was categorized as well, unwell, or unclear or not documented. A clinical diagnosis of sepsis or shock was abstracted if shock was clearly stated in the ED attending note or ICU team note, if vasopressor agents were required (eg, parenteral dopamine, dobutamine, epinephrine, norepinephrine, and vasopressin), if there was evidence of low blood pressure (<60 mmcHg systolic), or if hypotension or sepsis was clearly stated in the chart and the patient was treated with isotonic fluid boluses of ≥40 mL/kg. The presence of necrotizing or severe soft tissue infection was determined by the documentation of those specific terms, “extensive,” “significant,” or other similar terms describing severe local infection.

Our primary outcomes were the presence of SBI and adverse outcomes. We defined SBI as bacteremia, bacterial meningitis, or urinary tract infection (UTI). We defined bacteremia and bacterial meningitis by the growth of a pathogen (Supplemental Table 8) in cultures of blood or cerebrospinal fluid (CSF), respectively, and UTI by growth of ≥50 000 colony-forming units per mL of a single pathogen (Supplemental Table 9) or ≥10 000 colony-forming units per mL in association with positive urinalysis results⁸  from a catheterized specimen. We defined adverse outcomes as ICU admission, a clinical diagnosis of sepsis or shock, severe or necrotizing soft tissue infection, endotracheal intubation, administration of vasopressors, or death. The presence of adverse outcomes was categorized as yes, no, or unclear or not documented. Additionally, we defined surgical treatment of infection as bedside incision and drainage, intraoperative debridement, or other surgical procedure. For infants discharged from the ED, we reviewed the medical records to identify ED return visits within 7 days.

Age is presented with medians and interquartile ranges (IQRs), and categorical variables (demographics, historical and clinical findings, use of cultures, prevalence of SBI, and adverse events) are presented with proportions and 95% binomial confidence intervals (CIs).

We conducted statistical analyses using Microsoft Excel (version 16.0; Microsoft Corporation, Redmond, WA).

We analyzed 657 infants diagnosed with mastitis in the ED who met study criteria. The median age was 21 days (IQR 14–35 days). Most (94%) were born at term and 86% had no significant medical history. Fever before or during the ED visit was reported in 21%, and fussiness or poor feeding was reported by the caregivers of 10% of subjects. Providers described 98% of infants as well appearing on presentation. Complete blood cell count was obtained in 579 (88%) infants, with a median peripheral white blood cell (WBC) count of 15.6 × 10³/µL (IQR 12.9–19.5). C-reactive protein was obtained in 148 (23%), with a median of 0.9 mg/dL (IQR 0.5–2.2) (Table 1).

Table 2 summarizes the results of blood, urine, and CSF cultures. Of the 657 subjects, 581 (88%) had blood cultures obtained, with bacteremia identified in 2 infants (0.3% [95% CI 0.04–1.2]): 1 with GBS and 1 with methicillin-resistant S aureus (MRSA). Both were afebrile and described as well appearing. Both also grew MRSA from site cultures, and both received intravenous (IV) antibiotics in the hospital for 4 days and were discharged on oral antibiotics without complications. Growth of a contaminant occurred in 25 (4.3% [95% CI 2.8–6.3]) blood cultures. Urine studies were obtained in 274 (42%) infants, of which 3 were consistent with a UTI (1.1% [95% CI 0.2–2.3]), representing 0.4% (95% CI 0.1–1.3) of all patients. CSF cultures were obtained in 216 (33%) infants, 1 of which had positive results (0.4% [95% CI 0.01–2.5]), representing 0.15% (95% CI 0.0–0.8) of all patients. The one infant with a positive CSF culture result grew the same pathogen as the site of infection (MRSA). This infant had a history of fever but was described as well appearing on presentation; the CSF sample had 15 600 red blood cells per mm³ and 44 WBCs per mm³. This infant received IV antibiotics for 14 days in the hospital and was discharged without complication. None of the infants with SBI had clinical diagnoses of sepsis or shock. The characteristics of infants with SBI are shown in Table 3. All the positive culture results occurred in term infants without significant past medical history. Cultures from the site of infection were obtained in 335 (51%) infants, with growth of a pathogen in 258 (77% [95% CI 72–81]). The most common organism isolated from the site of infection was MRSA (54%), followed by methicillin-susceptible S aureus (29%) and unspecified S aureus (8%) (Table 4). Most (591, 90%) infants were admitted to the hospital, 22 (3.7%) of whom were admitted to an ICU. A bedside incision and drainage, intraoperative debridement, or other surgical procedure was performed in 143 (22%) infants. Of all patients, 10 (1.5% [95% CI 0.7–2.8]) had a clinical diagnosis of sepsis or shock (Table 5); only 3 (0.5% [95% CI 0.1–1.5]) had both a diagnosis of sepsis or shock and an ICU admission. Two infants (0.3% [95% CI 0.04–1.1]) had documented severe cellulitis or necrotizing soft tissue infection. No infants received vasopressor medications, and none required endotracheal intubation other than for an operative procedure. Of the 66 patients discharged at the initial ED visit, 15 (23% [95% CI 13–35]) returned to the ED within 7 days. Of these, 10 (67%) patients were admitted at their return visit with an admission diagnosis of mastitis, with 2 requiring surgical treatment. There were no adverse outcomes among the infants admitted at their return visit. There were no deaths in the cohort (Table 6).

Although all febrile infants aged 28 days and younger require evaluation with cultures of blood, urine, and CSF, we sought to evaluate the utility of cultures in afebrile, well-appearing infants with mastitis. In this large, multicenter cohort of infants with mastitis, SBI and adverse outcomes were rare, with bacteremia occurring in 0.3%. Although nearly all infants (98%) were described as well appearing and only one-fifth were febrile, 88% had blood cultures performed and one-third underwent lumbar puncture. Growth of a contaminant in blood culture occurred 10 times more frequently than growth of a true pathogen, which has significant implications in terms of further testing, medical interventions, and financial costs.^9,10  All the infants with SBI were aged <28 days, whereas the 1 infant with meningitis and 2 with UTI had fever, and the 2 with bacteremia and 1 with a UTI were afebrile and described as well appearing. The 2 with bacteremia did not have clinical sepsis or shock, and both were treated in a manner consistent with the localized infection with 4 days of IV antibiotics, were discharged on oral antibiotics, and did not return after discharge. Although the presence of a UTI did not seem related to the mastitis, the frequency of UTI in our cohort was less than that seen in infants with bronchiolitis.¹¹  In addition, the frequency of UTI was lower than that seen in febrile infants without obvious source.¹²  Given this low rate of SBI, our study suggests that routinely obtaining cultures of blood, urine, and CSF may not be indicated in well-appearing, afebrile infants with mastitis.

The low prevalence of SBI and adverse outcomes is consistent with previous studies. In several case series, the rate of concurrent bacteremia was ∼4%; however, these estimates included potential contaminants.^3,5,7  Similarly, the rates of concurrent meningitis and UTI were low. In a series of 41 cases from 1947 to 1983,³  19 had blood, 9 had CSF, and 6 had urine cultures obtained, all of which had negative results. The largest recent series included 94 patients from 2005 to 2011.⁷  There were no positive blood cultures; results of 2 of 49 urine cultures were positive; and 1 patient was diagnosed with bacterial meningitis. Approximately half had systemic manifestations, such as irritability or decreased oral intake, and 24% were febrile. Previous studies have suggested an excellent prognosis in general, but serious complications, including extensive cellulitis, necrotizing fasciitis, and osteomyelitis, have been reported.^13,14  Our larger, multicenter cohort supports these previous smaller studies and lends credence to the generalizability of their findings. In our cohort, the occurrence of sepsis or shock and severe cellulitis or necrotizing soft tissue infection was also extremely low. Although we provided specific instruction regarding the definition of sepsis and shock, it is possible that these diagnoses were either underestimated or overestimated. In addition, ICU use may reflect local practice patterns; of note, only 3 infants had a diagnosis of sepsis or shock and were also admitted to an ICU. Given the low frequency of adverse outcomes and the retrospective nature of the study, we were not able to evaluate and identify risk factors for the presence of SBI and other adverse outcomes. Similarly, we were not able to assess the utility of inflammatory markers in potential management algorithms.

In a previous series of 18 infants,²  all 5 who had been pretreated with oral antibiotics before presentation developed abscesses that required surgical drainage. This has led to the recommendation that initial antibiotics be administered intravenously in the hospital. Additional evidence suggests that significant variation exists in the evaluation and management of infants with mastitis.¹⁵  The vast majority of patients in our cohort were admitted to the hospital, and 22% had surgical treatment of their infection. Among the 10% of patients discharged from the ED at their initial visit, 23% returned within 7 days of discharge, and 10 were admitted at their return visit, representing 1.5% of the entire cohort. Although 2 of these infants required surgical treatment, there were no adverse outcomes among this group. Although there may be a lower risk population that might not require hospitalization or surgical intervention, our data do not allow evaluation of this process. Further research is warranted to determine if it may be reasonable to consider outpatient antibiotics as an initial approach in some, if strict return precautions are given and adequate follow-up is ensured. Given the isolation of MRSA in more than half of the positive site culture results in our cohort, empirical antibiotic coverage for MRSA should be included, regardless of route. Future studies focusing on identification of lower risk patients may lead to opportunities for quality improvement in reducing unnecessary admissions.

Our study has several limitations. Although the frequency of concurrent SBI was low, the 2 patients with bacteremia were described as well appearing and were afebrile in the ED, so one cannot rely solely on clinical appearance at presentation to help guide the management approach. Because case ascertainment was based on ICD diagnosis codes, potential cases could be missed if the charted diagnosis was something other than mastitis, such as nonspecific cellulitis or abscess or a complication such as meningitis. Given the retrospective nature of the study, we cannot fully confirm the accuracy and completeness of abstracted data. Although we used restrictive guidelines to ensure appropriateness of diagnosis, it is not possible to be sure that all patients truly had confirmed mastitis. Interrater reliability was not assessed. If infants with noninfectious processes, such as galactorrhea or physiologic gynecomastia, were inadvertently included, it could falsely lower the prevalence of SBI and adverse outcomes. In addition, we did not perform a sensitivity analysis to find cases of mastitis that may have occurred in infants with primary SBI diagnosis codes that may not have been coded as mastitis, again potentially underestimating the incidence of concurrent SBI. There were no deaths in our cohort. Although it is possible that fatal cases of mastitis could be missed if they never presented to the hospital, or if they presented to a different hospital in the same geographic area, the likelihood of this seems low. Because we conducted this study principally at children’s hospitals or hospitals with dedicated pediatric EDs, our data may not accurately reflect the presenting signs or symptoms or treatment of infants with mastitis seen at general hospitals, in primary care provider offices, or in clinics. In addition, given the retrospective nature of the study, there is no way to know why some infants had cultures performed and some did not and if this was idiosyncratic and provider-driven or if there were other cues or reasons.

In this large, international cohort of infants with mastitis, the prevalence of concurrent SBI and adverse outcomes was low, and growth of a contaminant in blood culture occurred 10 times more frequently than growth of a pathogen. Routine testing for SBI in most afebrile, well-appearing infants with mastitis is likely unnecessary.

CI

confidence interval

CSF

cerebrospinal fluid

ED

emergency department

GBS

group B Streptococcus

ICD

International Classification of Diseases

IQR

interquartile range

IV

intravenous

MRSA

methicillin-resistant Staphylococcus aureus

SBI

serious bacterial infection

UTI

urinary tract infection

WBC

white blood cell