The Relationship Between the Amount of Breastfeeding and Late-onset Sepsis in Very Low Birth Weight Infants Within 4 Weeks After Birth


 Objective: To assess the relationship between the amount of breastfeeding and late-onset sepsis in very low birth weight infants within 4 weeks after birth.Study Design: In this retrospective study, we analyzed the effect of the breastfeeding ratio of 305 very-low-birth-weight infants on late onset sepsis.Results: The incidence of late-onset sepsis in the high-breastfeeding group, low-breastfeeding group, and the formula-fed group was 0.49% (1/203), 7.41% (6/81), and 4.76% (1/21), respectively (P < 0.05). Single-factor logistic regression analysis showed that, in the low-breastfeeding group, the risk of late-onset sepsis higher than that in the high-breastfeeding group, with an OR value of 16.16 (95% CI: 1.91, 136.47) (P < 0.05). Multivariate logistic regression analysis also confirmed that, compared with the high-breastfeeding group, the low-breastfeeding group had an increased risk of late-onset sepsis, with an OR value of 18.86 (95% CI: 1.91, 136.47) (P < 0.05). Conclusion: The breastfeeding amount of very-low-birth-weight newborns accounting for more than 60% of the total feeding amount within four weeks after birth may reduce the risk of late-onset sepsis.


Introduction
Neonatal sepsis is a clinical syndrome of neonatal infection 1 . According to the time of its occurrence, sepsis can be divided into early-onset sepsis (EOS) and late-onset sepsis (LOS). EOS is related mostly to intrauterine infection, whereas LOS generally occurs 72 hours after birth, and most of the cases are related to the operation, environment, and a low autoimmune function of the child during or after childbirth. The incidence of sepsis in premature infants is 20-38%, and the mortality Breast milk is the preferred feeding rate is 13-19% 2 , which is one of the common causes of death in premature infants 3 .
Premature infants have immature development of various organs and low immunity, which is easy to cause various complications. The earlier the newborn is born, the greater the risk of LOS 4 . About 60% of neonatal LOS cases come from VLBWI 5 . Compared with Gram-positive bacteria, Gram-negative bacteria such as Klebsiella and Escherichia coli are more likely to develop LOS from VLBWI 6 .LOS exerts many adverse effects to newborns, such as necrotizing enterocolitis (NEC), bronchopulmonary dysplasia (BPD), and neurodevelopment 5 , and can affect the growth and development of the newborns and their neurodevelopment 6 . Breast milk is the preferred feeding for premature infants, which can meet their growth and development needs. It also has a certain immunological protection effect on premature infants. Compared with breast milk of full-term mothers, the colostrum of mothers of premature babies contains higher protein, fat, free amino acids and sodium contents 7 , which can promote the development of newborns and prevent the occurrence of diseases. The essential amino acids provided by breast milk can reduce the risk of neonatal diseases 8 . The study have shown that breastfeeding in premature infants can decrease the occurrence of late-onset sepsis, but the protective effect of breast milk is dose-dependent 9 . Highbreastfeeding for 28 days after birth can reduce the risk of VLBWI LOS and other diseases 10 . Our study analyzed the relationship of breastfeeding intake and the incidence LOS within four weeks after birth for VLBWI.

Materials And Methods
From July 2018 to June 2019, 305 cases of VLBWI who were born in the Obstetrics Department of Shenzhen Maternity and Child Health Hospital a liated to Southern Medical University (Shenzhen, Guangdong, China) and were admitted to the Neonatal Intensive Care Unit within 24 hours were the subjects of the study.
The following inclusion criteria were applied: (1) Birth weight < 1500 g; (2) Birth age After the child was admitted to the Neonatal Intensive Care Unit, the medical staff instructed the family members about the bene ts of breastfeeding as soon as possible. The speci c content included the following main points: (a): On the day of admission, the doctor explained to the family of the child the bene cial effects of breastfeeding on the growth and development of VLBWI, and encouraged the family to prefer breastfeeding. Admission nurses taught family members the correct methods and precautions for breast milk collection, storage and transportation; (b) Breast milk management had dedicated staff. The breast milk nurses called the family members of the child on the day after the child was admitted to the hospital to once again educate the family on the importance of breastfeeding and the methods of collecting, storing, and transporting breast milk; (c) Open classes for parents were held every Monday and Thursday to ensure that the family members of the children admitted to the Neonatal Intensive Care Unit can receive breastfeeding education again and obtain accurate, relevant knowledge of breastfeeding. The basic data of VLBWI were collected in this study, including gestational age, age, gender, mode of delivery, an in vitro fertilization (IVF) or not, Apgar score, intrauterine distress, double entry, and consistency check.
According to the ratio of the breast milk intake to the total intake four weeks after VLBWI birth, they were divided into a high-breastfeeding group (breast milk intake > 60%, 203 cases), a low-breastfeeding group (breast milk intake accounting for a ratio ≤ 60%, 81 cases), and a formula-fed group (no breast milk intake, 21 cases).
The observation indicators of this study included the time of feeding, breast milk intake, total intake, the ratio of breast milk intake to the total intake; and the incidence of LOS. The diagnostic criteria of the complications were in compliance with the guidelines of "Practical Neonatology";. The following diagnostic criteria for LOS were used: isolate pathogenic bacteria from sterile cavity uids such as blood culture, cerebrospinal uid, and urine culture. The basis for the diagnosis of LOS was on the premise of abnormal clinical manifestations, meeting any of the following conditions: (1) Non-speci c blood test ≥ 2 items were positive; (2) Cerebrospinal uid examination was abnormal; (3) DNA or antigen of special bacteria was detected in the blood.
Statistical analysis SPSS 24.0 software (IBM, Armonk, NY, USA) was used to perform statistical analysis, and normally distributed data were expressed as mean ± standard deviation (χ 2 ±s). One-way analysis of variance (ANOVA) was used for making comparison among multiple groups, and count data were presented as frequency and percentage. The χ 2 test was employed for making comparison between two groups. The binary logistic regression analysis was used to analyze the relationship between breast milk volume and incidence of LOS; the single factor logistic regression analysis was employed to analyze the relationship between related variables and incidence of LOS; and the multivariate logistic regression analysis was utilized to analyze the relationship between breastfeeding volume and the incidence of LOS. P < 0.05 was considered statistically signi cant.

General situation
The study nally included 305 newborns who met the inclusion criteria. There were no statistically signi cant differences in the basic characteristics of the three groups of VLBWI, such as gestational age, age, gender, and delivery method (P > 0.05), (Table 1). Note: a is expressed as x±s, and the statistical value of t is the F value; b is expressed as an example (%), and the statistical value is χ 2 ; c is expressed as median(interquartile range), and the statistical value is K-W.

Feeding status and complications
Statistically signi cant differences (P < 0.01) were found in the comparison of the incidence of LOS in the high-breastfeeding group, low-breastfeeding group, and the formula-fed group. However, no statistically signi cant difference was observed in the incidence of BPD, ROP, and IVH (P > 0.05), (Table 2).

Comparison of the risk of late-onset sepsis among the three groups
Single-factor logistic analysis of the breastfeeding amount and LOS shows that taking the highbreastfeeding group as the reference group, the low-breastfeeding group had a signi cantly higher risk of LOS, with OR values of 16.16 (95%CI: 1.91-136.47), and the difference was statistically signi cant (P < 0.05), (Table 3).

Discussion
Due to individual differences and other reasons, mothers of preterm infants cannot produce milk in time after delivery. Such preterm infants are usually fed with donated breast milk or formula milk. In our study, breast milk from their own mothers, and those without breast milk were fed with formula milk (provided uniformly by the hospital).
Boyd et al. 11 found that the use of donated breast milk for feeding of preterm infants reduced their infection rates. However, the growth and development of preterm infants fed by donated breast milk were slower. Furthermore, preterm infants fed with formula milk had signi cantly longer oxygen use and mechanical ventilation times 12 ; importantly, their growth and neurodevelopment were worse than those of the breastfed infants. In our study, breastfeeding was done with mother's milk.
Here, a total number of 305 VLBWI (weight ≤ 1500 g) were included. We analyzed different ratios of breast milk intake to the total intake of VLBW within four weeks after birth. The results showed that when breast milk intake accounted for more than 60% of the total intake, the incidence of LOS was signi cantly reduced. Currently, antibiotics are often used in clinical practice to prevent and treat LOS. Meanwhile, active nursing support treatment is also crucial for the prevention and treatment of LOS in premature infants 13 . A large number of studies 6,14,15 have found that breastfeeding can reduce the incidence of complications in preterm infants such as necrotizing enterocolitis and LOS. In another investigation, the incidence of VLBWI infectious diseases with exclusive breastfeeding was lower than that of the formulafed infants 16 . Furman 17 compared several parameters between VLBWI and non-breastfeeding. Breastfeeding reached or exceed 50 mL/kg/d, the incidence of LOS is reduced. In our study, the incidence of LOS was 2.6%. The univariate analysis showed that the admission weight was related to the occurrence of LOS. Nevertheless, after adjustment for confounding variables related to gender, age, gestational age, and admission weight, the result was still valid, which is consistent with the results of the aforementioned study.
Studies have shown that, generally, after 34 weeks of gestation, the fetus obtains maternal immune protein from placental transfer, but the gestational age of VLBWI is generally less than 32 weeks. Therefore, VLBWI does not have su cient immune protein at birth. Thus, low immunity after birth can facilitate the occurrence and development of various diseases 18 . The American Pediatric Association strongly recommends breastfeeding for preterm infants 19 . Breastfeeding is more conducive to the growth and development of preterm infants than formula feeding and reduces their infection rates. The breast milk of preterm infants contains more secreted immunoglobulin A (IgA), lysozyme, lactoferrin, and interferon than the breast milk of term infants 20 , which provides good conditions for development and protection to premature infants by enhancing their immunity and reducing the risk of infection 21 . A large number of studies 16-18 have revealed that breastfeeding exerts a good protective effect in premature infants after reaching a certain proportion, which is consistent with the results of our study. Unlike previous studies, Our study found that the incidence of LOS in the low-proportion group was higher than that in the formula milk group, which may indicate that low-proportion breast-milk mixed feeding is more likely to occur when VLBWI is fed than full-formula.
In summary, within four weeks after birth of very-low-birth-weight infants, breastfeeding accounted for 60% of the total feeding amount, which reduced the incidence of VLBWI LOS. Formula feeding or mixed feeding should be reduced, and low breast milk mixed formula feeding should be avoided for exerting the bene cial effects of breast milk on premature infants. Therefore, in the early days after birth, highbreastfeeding regimen should be adopted in preterm infants as much as possible. Its protective effect reduces effectively the incidence of LOS in preterm infants, and at the same time promotes their smooth growth and development. The number of the positive cases in this study was too small, which might have led to an excessively wide con dence interval, leading to an unstable outcome. Therefore, a larger sample size needs to be employed in the follow-up study to con rm our ndings.

Conclusion
The breastfeeding amount of very-low-birth-weight newborns accounting for more than 60% of the total feeding amount within four weeks after birth may reduce the risk of late-onset sepsis. Besides, VLBWI intake a small amount of breast milk mixed feeding or formula feeding may increase the incidence of LOS. Declarations