Maternal milk feedings and cytomegalovirus infection in preterm infants in Sweden

Abstract

In Sweden, preterm infants are preferably fed human milk. Very preterm infants (< 32 weeks), who are unable to breastfeed, are fed with expressed maternal milk via a nasogastric tube. Mothers of these infants often experience difficulties in establishing and maintaining lactation. The majority of women excrete cytomegalovirus (CMV) in their breast milk. CMV transmitted through maternal milk can cause symptomatic infection in preterm infants presenting as a sepsis like syndrome, pneumonitis, hepatopathy or enterocolitis. Routine freezing of maternal milk decreases the CMV load in breast milk and is used in some neonatal centers to reduce CMV transmission to preterm infants.

The aims of the studies in this thesis were to document existing routines pertaining to breast milk use for preterm infants in Sweden, to investigate predictors of maternal milk feedings in extremely preterm infants (EPIs, < 28 weeks), to evaluate the rate and clinical expression of postnatal CMV infection in EPIs, to evaluate the effect of routine freezing of maternal milk on CMV transmission rate, CMV associated disease and neonatal morbidity and mortality in EPIs and to evaluate the prevalence of CMV infection in intestinal specimens from infants with necrotizing enterocolitis (NEC), spontaneous intestinal perforation (SIP) and related surgical conditions.

In a national cross sectional study in 2006 in Sweden, we found that 27 of 36 (75%) neonatal units had their own milk bank. Milk donors were screened for human immunodeficiency virus, human T-lymphotropic virus, and hepatitis B and C viruses by 27 (100%), 14 (52%), and 22 (81%) of the milk banks, respectively. Bacterial culture was performed on donor milk in 24 (89%) milk banks. Donor milk was pasteurized in 22 (81%) milk banks. In 11 of the 36 (31%) neonatal units maternal milk was frozen to reduce the risk ofCMV transmission. Nutritional analysis of donor and/or maternal milk was performed in 25 (69%) units.

In a prospective cohort study at the neonatal units in Stockholm, including 97 mothers and their singleton EPIs, predictors of maternal milk feedings in EPIs during the first 6 weeks of life and at discharge were evaluated. Favorable predictors of maternal milk feedings the first 6 weeks of life were high maternal milk feedings (>90%) at second week of life, maternal university education and Nordic origin of the mother. The proportion of maternal milk feedings the first 6 weeks of life and maternal age were positively associated to the provision of maternal milk feedings at discharge while maternal overweight was an unfavorable predictor. High maternal milk feedings (>90%) at second week of life, assisted reproduction technology and maternal employment were predictive factors for exclusive maternal milk feedings at discharge.

Ten EPIs and their 6 mothers were included in a pilot study at the neonatal unit, Astrid Lindgrens Children´s hospital to evaluate the rate and clinical expression of breast milk induced CMV infection. Five (83%) mothers were CMV-seropositive; of these, 4 (80%) excreted CMV-DNA in breast milk and 2 (40%) had a positive CMV culture. CMV was detected in the urine of 2/7 (29%) EPIs fed with CMV-positive milk; both were fed with breast milk positive for CMV culture. One EPI, later diagnosed with cystic fibrosis, developed hepatic affection concurrent with CMV urine excretion.

In a randomized study at the neonatal units in Stockholm, evaluating the effect of routine freezing of maternal milk on postnatal CMV infection and neonatal outcome, 140 EPIs were randomized to be fed only freeze-thawed maternal milk (intervention group, IG) or both fresh maternal milk and freeze-thawed maternal milk (control group, CG). Outcome measures were CMV transmission rate and symptomatic infection in EPIs, neonatal mortality and morbidity during hospital stay. Fifty-six EPIs in the IG and 65 EPIs in the CG were included in the final per protocol analysis. We observed an overall low CMV transmission rate (8%) to EPIs from mothers with detectable CMV in breast milk. Routine freezing of maternal milk did not reduce the rate of CMV transmission (9% in IG vs 6% in CG). Congenital CMV infection was detected in 2% of screened infants. No infected EPI presented with clinical symptoms of CMV infection. Mortality rates were similar; 7% in the IG and 6% in the CG. Neonatal morbidity did not differ except for late onset Candida sepsis; the incidence was 12% in the CG while no case was observed in the IG.

In a retrospective observational study, we investigated the occurrence of the CMV in 70 intestinal specimens from 61 infants with NEC, SIP and related surgical conditions at the Karolinska University Hospital Solna and Uppsala University Hospital. Ten intestinal specimens from autopsied infants without bowel disease were controls. By using immunohistochemistry (IHC), we detected the CMV specific proteins CMV-immediate early antigen (CMV-IEA) in 81% (57/70) and CMV-late antigen (CMV-LA) in 64% (45/70) of the intestinal specimens; 2/10 (20%) of the control specimens were positive for both antigens. Although CMV antigens were prevalent irrespective of pathologic diagnosis, they were most frequent in specimens with the pathologic diagnosis NEC and intestinal perforation; 95% and 89% ofthese tissue specimens were positive for CMV-IEA and CMV-LA, respectively. CMV infection was confirmed by CMV-DNA analysis in 4/10 (40%) CMV-IHC-positive intestinal samples using Taqman PCR after laser capture microdissection and in 13/13 (100%) CMV-IHC-positive intestinal samples by in situ hybridization.

To conclude, human milk handling routines vary between neonatal units in Sweden and need to be standardized. Mothers of EPIs should aim for a high breast milk production immediately after delivery to optimize lactation success. Mothers who are young, overweight, ofnon-Nordic origin or without university education may need special lactation support. Postnatal CMV transmission from mothers excreting CMV in breast milk to EPIs was low (8%) and was not reduced by routine freezing of maternal milk. However, congenital CMV infection in EPIs was unexpectedly high (2%). No EPI infected by CMV presented with clinical symptoms. Routine freezing of maternal milk did not affect neonatal death in EPIs although it may have protected against fungal late onset sepsis. CMV infection was prevalent in intestinal specimens from infants with NEC, SIP and related surgical condition implicating a possible role of the virus in disease pathogenesis. More studies are needed to further evaluate the risk/benefit ratio ofmaternal milk feedings in EPIs with regard to the short-term and long-term effects of postnatal CMV infection.