BREAST MILK AS A SOURCE OF LATE ONSET NEONATAL SEPSIS

Abstract

Neonatal sepsis can be a life-threatening complication in preterm neonates. We present the clinical course of 3 preterm neonates, 1 with recurrent sepsis and 2 with late onset sepsis attributed to ingestion of breast milk containing pathogenic organisms. Breast milk should be considered as a potential source of infection in neonates with recurrent infections or when infections occur simultaneously in siblings.

Breast milk, apart from being the optimal source of nutrition, has immunologic properties that prevent infection in neonates.¹ It contains lysozyme that acts directly against bacteria and indirectly potentiates the bactericidal activity of antibodies.² Secretory IgA, lactoferrin and lysozyme in breast milk are important for prevention of gastrointestinal infection and necrotizing enterocolitis.³

In this article, we report the clinical course of 3 preterm neonates, 1 with recurrent sepsis and 2 with late onset sepsis attributed to ingestion of breast milk containing pathogenic organisms.

CASE 1

A 30-week gestational age infant was born vaginally to a gravida 3, para 1, Caucasian woman. Fetal membranes had ruptured at 27 weeks gestational age, and the mother had received a course of amoxicillin and antenatal steroids. Vaginorectal swab and urine culture done before delivery were negative for group B Streptococcus (GBS). After birth, the infant developed respiratory distress syndrome. After a blood culture was obtained, the infant was treated with ampicillin and gentamicin intravenously. Lumbar puncture was performed at 20 hours of age when the blood culture was known to be positive for GBS and showed 4320 × 10⁹/L white blood cells with polymorphonuclear cell predominance (87%) in the cerebrospinal fluid (CSF). The CSF culture was positive for GBS. Blood and CSF GBS bacteria were susceptible to ampicillin, clindamycin, erythromycin and penicillin. The minimum inhibitory concentration for ampicillin reported was 0.12 μg/mL. The patient received 3 weeks of treatment with ampicillin, with gentamicin during the first 2 weeks. Repeat blood and CSF cultures 5 and 10 days after starting therapy were sterile. Enteral feedings with breast milk was started on day 3 of life with full feeding achieved within 1 week. Seizures, confirmed by electroencephalogram, were controlled with phenobarbitone on the first day. There was no evidence of ventriculitis or abscess on cranial ultrasound examination. Nine days after discontinuation of antibiotics (day 30 of life), the infant developed lethargy, irritability, poor feeding, increased apneic spells and bradycardia. After blood, CSF and urine cultures were obtained, the infant's treatment was commenced with ampicillin and cefotaxime. Blood culture was again positive for GBS, but CSF and urine cultures were negative. Abdominal ultrasound, echocardiography and radiographic skeletal survey examination failed to identify the source of reinfection. Culture of fresh breast milk on the day of clinical deterioration was positive for GBS, which were susceptible to ampicillin, clindamycin, erythromycin and penicillin. On the basis of this finding, the mother decided to discontinue breast-feeding. Antibiotics were changed to penicillin, and the infant was treated for 2 weeks. Repeat blood cultures on the 2nd and 14th days after second presentation were negative. No further episodes of infection were observed. Neurologic examination including hearing test was normal at discharge (day 42 of life). In this case, a maternal vaginorectal swab for GBS was negative at the time of delivery, and the mother did not have any symptoms of mastitis in the postpartum period.

CASES 2 AND 3

Preterm triplets were born at 25 weeks gestational age to a gravida 2 para 1 mother of Asian origin. At 20 weeks gestation, she had been diagnosed with urinary tract infection, but urine culture was sterile. Apart from preterm birth, no risk factors for sepsis were noted. Each triplet was intubated after birth for assisted ventilation. A blood culture was obtained, and intravenous ampicillin and gentamicin therapy was started. Antibiotics were discontinued at 48 hours of age when the blood culture had no growth.

On day 17 of life, triplet C became ill with multiple episodes of apnea and increased oxygen requirements needing assisted ventilation. After a blood culture was obtained, therapy was commenced with cloxacillin and gentamicin. Lumbar puncture was deferred for 48 hours because of thrombocytopenia. Blood culture was positive for Klebsiella pneumoniae, susceptible to ceftriaxone, cefazolin, cefoxitin, cefuroxime and gentamicin. CSF and urine cultures were sterile. Antibiotic treatment was changed to cefotaxime and gentamicin. A swelling in the right axilla measuring 0.8 × 0.6 cm was detected. Five days later, another swelling was noted in the left axilla. An ultrasound scan of these swellings documented an isoechoic mass of nonspecific origin. Therefore the lesions were not drained, and they resolved with 3 weeks of antibiotic therapy. Repeat blood culture was sterile.

On day 20 of life, triplet B developed temperature instability, poor peripheral perfusion, metabolic acidosis, hyponatremia and thrombocytopenia, prompting resuscitation with normal saline, red blood cell and platelet transfusion. After blood was obtained, CSF and urine cultures, therapy was started with cloxacillin and gentamicin. Blood culture grew K. pneumoniae with susceptibilities the same as for the organism cultured in the sibling. CSF and urine cultures were negative. Antibiotic treatment was changed to cefotaxime and gentamicin administered for 3 weeks. Skin abscesses over the nasal bridge, right wrist and neck were noted during the next few days. Pus drained from a right wrist abscess grew K. pneumoniae, whereas pus from the right neck abscess was sterile. Repeat blood cultures were negative.

The third triplet became unwell during the same time as the other two, but blood culture was negative. The infant was empirically treated for 5 days with cefotaxime and gentamicin.

The triplets were receiving breast milk as the enteral source of nutrition at the time of deterioration. Because 2 of the siblings developed K. pneumoniae sepsis, an attempt was made to identify a common source of infection. Abdominal ultrasound, echocardiography and radiographic skeletal survey examination were normal. Thawed and freshly expressed breast milk cultures were positive for heavy growth of Klebsiella pneumoniae susceptible to ceftriaxone, cefazolin, cefoxitin and cefuroxime. Until the results of breast milk culture were available, the infants were given formula feeding with parental consent. The mother was counseled by the lactation consultant and advised of the importance of hand washing and breast pump hygiene. A new breast pump was provided, and breast milk cultures sent a week later revealed no bacterial growth. The infants were restarted on breast milk after 2 sterile breast milk cultures and remained infection-free for the rest of their hospitalization.

DISCUSSION

We attributed recurrent and late onset sepsis to ingestion of breast milk containing pathogenic organisms in these babies. None of the mothers had evidence of mastitis nor were the infants being breast-fed. They received gavage feedings of either fresh or thawed breast milk expressed by a breast pump. We speculate that the mother of case 1 was colonized with GBS during the postpartum period leading to contamination of the breast milk during use of the breast pump. We postulate that infection in the triplets was linked to the breast pump given that replacement with a new pump led to eradication of the organism in the milk; however, a culture from the breast pump was not sent to confirm this. Molecular typing of the organisms from the milk and blood of all 3 cases would have provided additional evidence of this association.

There have been several reports of late onset and recurrent GBS disease and of bacteremia with Klebsiella spp. related to human milk contamination.^4,5 The mechanism by which late onset sepsis after ingestion of infected breast milk infection occurs is not certain. It has been postulated that the organisms initially colonize the infant's oropharyngeal mucosa from perinatal sources. During breast-feeding, the mammary ducts become infected. As the microbial concentration increases in the milk, the infant is reinfected during breast-feeding.⁴ This explanation is not applicable to our cases, because none of the infants was put to the mother's breast. In a randomized controlled trial, Boo et al⁶ reported significantly higher contamination rate when milk was expressed by breast pump than manually (86.3% versus 61%, P = 0.005) and higher bacterial contamination with Gram-negative organisms in the breast pump group expressed at home than in the hospital (52.6% versus 17.1%, P = 0.001).

Despite the benefits of breast milk, it should be considered as a potential source of infection in neonates presenting with recurrent infections or when infections occur simultaneously in siblings. Emphasis should be placed on educating mothers regarding hand washing techniques before handling the breast pump equipment and regarding care during pumping, storage and transport of milk for their infants. Both verbal and written information should be provided to all mothers and should be reviewed on an ongoing basis.