Formula Feeding Is Independently Associated With Acute Kidney Injury in Very Low Birth Weight Infants

Abstract

Background:

Successful strategies to prevent neonatal acute kidney injury are lacking. Nevertheless, it is well known that in breastfed babies the excretory needs of the kidney are low because the intake of most nutrients is just above the nutritional requirement.

Objectives:

This study aimed to determine whether feeding type predicts acute kidney injury in the very low birth weight infant.

Methods:

One hundred and eighty-six infants were enrolled in this pre-post cohort study (114 infants were included in the only human milk-fed group and 72 in the formula-fed group). Routine biological markers of acute kidney injury were collected in both groups from birth to discharge.

Results:

Compared with formula feeding, human milk feeding was associated with almost 80% lower odds of acute kidney injury (odds ratio [OR] = 0.2; 95% confidence interval [CI], 0.05-0.77). After confounding variables had been controlled for, formula feeding was independently associated with acute kidney injury in very low birth weight infants.

Conclusion:

The study showed that, at our institution, acute kidney injury in the neonatal period is frequently associated with the avoidable procedure of formula feeding. Further prospective multicenter studies are needed to determine the generality of this association.

Keywords

acute kidney injury breastfeeding milk bank newborn infant

Well Established

Previous studies have compared the biochemical status of newborn infants fed human milk with those of similar infants fed formula. Serum urea and uric acid levels are significantly lower in the breastfed group, whereas infants receiving formula are prone to metabolic acidosis or hyponatremia.

Newly Expressed

This study reports an inverse association between human milk feeding and acute kidney injury among preterm infants. The robustness of this association was confirmed by multivariate analysis. We suggest that biomarkers of renal function reflect the detrimental effect of formula feeding.

Background

We have made progress in understanding the effects on the kidneys of early neonatal disorders as well as those of the drugs used to treat them.¹ Conversely, studies to determine whether early feeding type (human milk [HM] versus formula) affects renal balance in the preterm infant are scarce. To date, no study has been carried out to determine specifically the effect of feeding type on the kidney function of preterm neonates. In breastfed infants, intake of most nutrients is just above the nutritional requirement. Thus, in breastfed babies, the excretory needs of the kidney are low.² We hypothesize that this optimality in the composition of HM during the first days of life protects very low birth weight (VLBW) infants against acute kidney injury (AKI).

In our neonatal intensive care unit (NICU), we recently implemented the strategy to use donor milk (DM) for preterm infants whose mothers suffered from breast milk shortage. This enabled us to examine the possible contribution of feeding type onto AKI among VLBW infants.

Methods

Overview

We undertook a single centre pre-post cohort study to assess the reported incidence of AKI in the NICU, as defined by an increase in serum creatinine (Scr >1.5 mg/dL; i.e., >132.6 umol/L) and a decrease in urine output (oligoanuria; i.e., <0.5 mL/kg/h) among VLBW infants fed HM versus formula diets. During the period over which this study was conducted there were no changes to the creatinine assay. Measurement of serum creatinine in neonates is based on an isotope dilution–mass spectrometry–traceable Jaffe assay.

The NICU team became aware of university hospitals nearby successfully using DM. Cost was a barrier to implementation of a DM program. The hospital’s program was initially funded by a grant and later supported by the hospital. Hospital administration was supportive of the use of pasteurized donor milk after a presentation of the results of our pilot study. A number of neonatal nurses have qualified as lactation consultants. Our unit has incorporated the Spanish guidelines of DM storage into its practice.

This study was a secondary analysis of infants cared for in our NICU using an existing database, SEN1500,³ which includes all VLBW Spanish infants. We conducted an analysis of available data from all VLBW infants admitted to our NICU, before (October 1, 2005 to March 31, 2009) and after (January 1, 2010 to May 31, 2014) initiating the DM policy. Infants with major congenital anomalies, outborn babies older than 48 hours of age, and neonates who deceased before day 3 of life were excluded. Before the introduction of DM, feedings consisted of preterm formula (PF) when there was an insufficient mother’s own milk (MOM) supply. On April 1, 2009, our NICU officially adopted a policy of providing DM as the preferred alternative when MOM is not available in sufficient quantity to meet her infant’s need. Our NICU infants are eligible to receive DM if they are born <1500 g or <32 weeks gestation.

We defined “exclusive diet of human milk” (HM) as feeding babies with no other milk, apart from MOM alone, DM alone, or DM plus MOM. In addition, we defined “any formula” (PF) as feeding infants with any amount of PF (PF alone or PF plus MOM). The same type of PF (Alprem®) was used for all VLBW infants during the study period. A higher concentration of protein (2.3 vs 1 g/dL), fat (4.1 vs 3.5 g/dL), carbohydrates (8.5 vs 7 g/dL), calcium (54 vs 15 mg/dL), phosphorus (54 vs 15 mg/dL), sodium (51 vs 35 mg/dL), and potassium (108 vs 48 mg/dL) has been reported in Alprem compared to term milk. We compared AKI of VLBW infants who received PF (pre-DM policy) with those who received HM (post-DM policy). Institutional review boards of Santa Creu i Sant Pau University Hospital reviewed and approved the consent process and the study protocol.

Feeding Policy

Trophic feeding (unfortified HM or PF) was usually introduced within the first 24-48 hours of life. Mothers were also encouraged to provide their own milk by initiation of breast milk pumping in the first 12 hours. After 3 to 4 days of gut priming, feedings were advanced at a rate of 15 to 30 mL/kg/day depending on the size and feeding tolerance of the infant. When enteral feeds reached 80 mL/kg per day, HM was fortified. When infants tolerated 150 mL/kg/day of enteral feeds, parenteral nutrition was discontinued. Replacement of fortified DM with PF was considered for preterm infants weighing more than 1500 grams and already on full enteral feeds.

Data Collection

Major morbidities were collected. Weight, head circumference, and length were converted to standard deviation and corrected for postnatal age using the validated Fenton growth chart.

Data Analysis

The t test was used to analyze continuous data, and the chi-square analysis and odds ratio were used for categorical data. Results were considered significant at P < .05. The main outcome variable was AKI before neonatal discharge. Logistic regression was used to control the effects of possible confounding variables on the association between AKI and feeding type. Data were analyzed using the IBM-SPSS (V22.0) Package.

Results

In the present study, 114 HM-fed and 72 PF-fed VLBW infants were enrolled, of which 172 infants survived to discharge (105 HM and 67 PF). Neonatal survival was not affected by diet (P = 1). At birth, HM and PF groups had similar baseline characteristics. Common neonatal morbidities or length of hospital stay did not differ significantly by milk type. Only AKI was significant. More infants in the PF group satisfied the criteria for AKI than HM infants (12.5% and 3.5%, respectively, P = 0.035) (Tables 1 -3).

Table 1.

Patient Characteristics by Feeding Type (N = 186).

Variable	Mother’s Milk and/or Donor Milk (n = 114)n (%)	Any Formula (n = 72)n (%)	P
In vitro fertilization	19 (17)	15 (21)	0.56
Antenatal steroids	99 (87)	57 (79)	0.22
Antenatal antibiotics	56 (50)	29 (41)	0.22
Vaginal delivery	38 (33)	26 (36)	0.75
Female	59 (52)	39 (54)	0.76
Twin pregnancy	32 (28)	29 (40)	0.10
Small for gestational age	12 (10)	8 (11)	1
Oxygen resuscitation	84 (74)	49 (68)	0.41
Delivery room intubation	39 (34)	28 (39)	0.53
Epinephrine	3 (3)	3 (4)	0.7
Chest compressions	4 (3)	3 (4)	1
Surfactant use	31 (27)	24 (33)	0.41
Transferred from other Hospitals	11 (10)	12 (17)	0.18

Table 2.

Hospital Outcomes by Feeding Type (N = 186).

Variable	Mother’s Milk and/or Donor Milk (n = 114)n (%)	Any Formula (n=72)n (%)	P
Parenteral nutrition on day 28	9 (9)	5 (7)	1
ROP	27 (5)	24 (35)	1
ROP surgery	1 (1)	3 (4)	0.30
Any NEC	11 (10)	6 (8)	1
NEC surgery	7 (6)	4 (6)	1
Focal intestinal perforation	3 (2)	1 (1)	1
Hyaline membrane disease	41 (36)	29 (40)	0.64
Pneumothorax	2 (2)	0	0,52
Neonatal hypotension	42 (37)	16 (22)	0,05
Antihypotensive inotropes	41 (36)	16 (22)	0,05
Early-onset sepsis (<72 hours)	5 (4)	0	0.16
Red blood cell transfusion	41 (36)	28 (39)	0.76
Acute kidneys insufficiency	4 (3.5)	9 (12.5)	0,035*
Intraventricular hemorrhage	23 (20)	8 (11)	0.16
Periventricular leukomalacia	7 (6)	1 (1)	0.15
Patent ductus arteriosus (PDA)	29 (25)	24 (33)	0.25
Medication use in PDA	23 (20)	23 (32)	0.08
Surgical ligation of PDA	6 (5)	3 (4)	1
CPAP	97 (85)	55 (76)	0.17
Surfactant therapy	46 (40)	36 (50)	0.23
Conventional mechanical ventilation	53 (46)	38 (53)	0.45
Nitric oxide	3 (3)	0	0.28
High frequency oscillatory ventilation	9 (8)	7 (10)	0.79
Weight at discharge < -2 SD	37 (32)	33 (46)	0.09
Oxygen therapy	76 (67)	54 (76)	0.19
Oxygen therapy at 28 days of life	28 (28)	15 (23)	0.59
Steroids for BPD	3 (3)	0	0.28
Death	9 (8)	5 (7)	1
Breastfeeding rate at discharge	84 (78)	47 (69)	0.22

Abbreviations: BPD, bronchopulmonary dysplasia; CPAP, continuous positive airway pressure; NEC, necrotizing enterocolitis; PDA, patent ductus arteriosus; ROP, retinopathy of prematurity; SD, standard deviation.

Table 3.

Comparison of Growth Parameters According to Feeding Type (N = 186).

Variable	Mother’s Milk and/or Donor Milk (n = 114)Mean (SD)	Any Formula (n = 72)Mean (SD)	P
Birth weight, g	1078 (289)	1108 (273)	0.49
Birth weight z score	−0.5 (1.14)	−0.68 (1.26)	0.31
Birth length, cm	36.0 (3.69)	36.5 (3.21)	0.27
Birth length z score	−0.66 (1.41)	−0.67 (1.33)	0.95
Birth Ponderal index	2,27 (0.33)	2,23 (0.28)	0.32
Birth head circumference, cm	25.8 (2.7)	26.1 (2.5)	0.33
Birth head circumference z score	−0.43 (1.31)	−0.44 (1.32)	0.97
Gestational age, days	204 (20.3)	207 (20.3)	0.31
Weight at discharge, g	2484 (684)	2253 (615)	0.02*
Weight at discharge z score	−1.61 (1.3)	1.8 (1.5)	0.26
Length at discharge, cm	44.7 (3.5)	4.2 (2.9)	0.39
Length at discharge z score	−1.9 (1.4)	−2.0 (1.8)	0.73
Head circumference at discharge, cm	33.1 (2.8)	32.8 (2.0)	0.28
Head circumference at discharge z score	−0.5 (1.2)	−0.7 (1.7)	0.23
Ponderal index at discharge	2.7 (0.32)	2.6 (0.33)	0.15
Days of life at discharge	65 (34)	58.5 (33)	0.19
Postconceptional age at discharge, days	204.3 (20.3)	207.4 (20.3)	0.30

Abbreviations: cm, centimeters; g, grams; SD, standard deviation.

Premature and LBW infants begin life with an incomplete complement of immature nephrons.⁴ They are then exposed to a variety of external stressors that can hinder ongoing kidney development or cause additional nephron loss. In addition to the two independent variables of interest, weeks of gestation and birth weight, several variables were identified in the literature as predictors of AKI such as hemodynamic alterations and infection. We eliminated variables when a bivariate analysis resulted in P values of greater than .20 (Tables 1 -3). Consequently, only weeks of gestation and birth weight were included in the regression model. When these two variables are analyzed together in a regression analysis, feeding type remains significant. HM feeding was associated with almost 80% lower odds of AKI (OR = 0.2; 95% CI, 0.05-0.77). OR did not increase when confounding variables were adjusted (adjusted OR 0.2; 95% CI, 0.06-075). Hence, multivariable stepwise logistic regression confirmed PF feeding as an independent predictor of AKI (Table 4).

Table 4.

Logistic Regression Analysis of Risk Factors for Neonatal Acute Kidneys Insufficiency.

	Bivariate Analysis			Multivariate Analysis
	P	OR	CI 95% OR	P	OR	CI 95% OR
Human milk	0.028*	0.26	0.075–0.860	0.015*	0.213	0.060–0.752
Birth weight	0.017*	0.99	0.996–1.001	0.011*	0.997	0.995–0.999
Weeks of gestation	0.055*	0.97	0.939–1.001	—	—	—

Abbreviations: CI, confidence interval; OR, odds ratio.

P ≤ 0.05.

Discussion

AKI in a critically ill neonate is often multifactorial, and its risk increases with the number of prevalent risk factors.⁵ Common causes of renal hypoperfusion in the NICU include fluid losses, excessive diuretic use, heart failure, and medications that cause reduction in renal blood flow. The descriptive nature of this study limited our ability to document all covariates involved in AKI pathogenesis. Nevertheless, this study does succeed in recording information on most of the risk factors associated with AKI (Tables 1 -3). In this study, we have not collected information on biological values other than serum creatinine and urine output of our patients, but we have recently reported no significant biochemical differences between the two groups (HM vs PF) before starting trophic feeds in a sub-sample of 56 very preterm infants.⁶ The reported incidence of AKI in the neonatal NICU varies widely depending on the patient sample and AKI definition used. Before 2008, most neonatal AKI studies used arbitrary definitions of AKI frequently defined by an absolute SCr > 1.5 mg/dL.⁷ Our patient recruitment started in 2005. We applied the 2005 standard for identifying AKI events in critically ill infants (Scr >1.5 mg/dL; oligoanuria; ie, <0.5 mL/kg/h) throughout this research. In our series, AKI was diagnosed in 13 cases (6.9%), including 4 HM-fed and 9 PF-fed infants, compared to worldwide frequencies ranging from 3.9% to 10.8%.^8,9 In worldwide studies, the cause of AKI is pre-renal in more than 95% of the cases. In most cases, AKI is associated with a primary condition such as sepsis, metabolic diseases, or perinatal asphyxia. Further, the course of illness and a number of therapeutic interventions are known risk factors associated with development of AKI in the neonatal period.⁸ This pre-post cohort study investigated the impact of feeding type on the development of AKI, lacking until now.

From pioneer studies to current knowledge of neonatal renal physiology, researchers have realized that when newborn infants are fed with milk containing higher concentrations of electrolytes than their mothers’ milk, renal dysfunction is more likely. McCance et al.² reported that the breastfed neonate does not need to have and indeed lacks kidneys as efficient at excretion of mineral salts as an adult, because most of those substances are used for growth. Previous studies have been designed to compare the biochemical status of newborn infants fed HM with that of similar infants fed formula. Serum urea^10-12 and uric acid¹² levels were significantly lower in the breastfed group compared with the formula-fed group. The HM group had more serum bicarbonate values <20 mmol/L than the PF group,¹³ and whereas hyponatremia developed in 20% of the infants receiving formula, it developed in 15% of those receiving preterm HM (P < .05).¹⁴ Urinary sodium concentration in preterm infants on HM was significantly low when compared to infant formula.¹⁵ Recent evidence suggests that artificial feeding exerts an independent detrimental effect on the neonate.¹⁶ Even nutritionally insignificant volumes designed to stimulate the developing gastrointestinal system (gut priming or trophic feeding) are reflected by different serum biochemistries.⁶ Even more, kidney growth is significantly increased in partially or fully formula-fed 3-month-old infants.¹⁷ Whether there are any long-term effects of early formula feeding on later kidney function remains to be seen. As a final point, in renal failure there is enhanced generation of reactive oxygen during kidney disease initiation and progression.¹⁸ Many aspects of oxidative stress still require clarification of the mechanisms that prompt renal failure. However, previous data^19,20 and our data suggest that breast milk, not artificial formula, acts as an antioxidant during early life in premature infants.

Conclusion

We report a 78% reduction in the odds of AKI with HM feeds. This new finding points at the importance of HM to prevent AKI. Further prospective research is needed if we are to obtain a full understanding of this association.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

All authors declare no support from any organization for the submitted work; the work has no financial support.

References

Jain

. Our kidneys: a lot is riding on their shoulders. Clin Perinatol. 2014;41(3):xv-xvii.

McCance

. Advise to a young scientist. In: Ashwell

, ed. McCance & Widdowson: A Scientific Partnership of 60 Years. London: British Nutrition Foundation; 1993:53-571.

Verd

Porta

Botet

. Hospital outcomes of extremely low birthweight infants after introduction of donor milk to supplement mother’s milk. Breastfeed Med. 2015;10(3):150-155.

Carmody

Charlton

. Short-term gestation, long-term risk: prematurity and chronic kidney disease. Pediatrics. 2013;131(6):1168-1179.

Jetton

Askenazi

. Acute kidney injury in the neonate. Clin Perinatol. 2014;41(3):487-502.

Verd

García

Gutiérrez

. Blood biochemical profile of very preterm infants before and after trophic feeding with exclusive human milk or with formula milk. Clin Biochem. 2014;47(7-8):584-587.

Selewski

Charlton

Jetton

. Neonatal Acute Kidney Injury. Pediatrics. 2015;136(2):e463-473.

Agras

Tarcan

Baskin

. Acute renal failure in the neonatal period. Ren Fail. 2004;26(3):305-309.

Youssef

Abd-Elrahman

Shehab

Abd-Elrheem

. Incidence of acute kidney injury in the neonatal intensive care unit. Saudi J Kidney Dis Transpl. 2015;26(1):67-72.

10.

Janas

Picciano

Hatch

. Indices of protein metabolism in term infants fed human milk, whey-predominant formula, or cow’s milk formula. Pediatrics 1985;75(4):775-784.

11.

Davies

Saunders

. Blood urea. Normal values in early infancy related to feeding practices. Arch Dis Child. 1973;48(7):563-565.

12.

T-Z

Huang

I-F

Chen

Y-C

Chen

P-H

Yang

L-Y

. Differences in serum biochemistry between breast-fed and formula-fed infants. J Chin Med Assoc. 2011;74(11):511-515.

13.

Schanler

Shulman

Lau

. Feeding strategies for premature infants: beneficial outcomes of feeding fortified human milk versus preterm formula. Pediatrics. 1999;103(6 Pt 1):1150-1157.

14.

Gross

. Growth and biochemical response of preterm infants fed human milk or modified infant formula. N Engl J Med. 1983;308(5):237-241.

15.

Kalhoff

Manz

Kiwull

Kiwull-Schöne

. Food mineral composition and acid-base balance in preterm infants. Eur J Nutr. 2007;46(4):188-195.

16.

Meier

Engstrom

Platel

Jegier

Bruns

. Improving the use of human milk during and after the NICU stay. Clin Perinatol. 2010;37(1):217-245.

17.

Shmidt

Damgaard

Boisen

. Increased kidney growth in formula-fed versus breast-fed healthy infants. Pediatr Nephrol. 2004;19(10):1137-1144.

18.

Ratliff

Abdulmahdi

Pawar

Wolin

. Oxidant mechanisms in renal injury and disease. Antioxid Redox Signal. 2016 Feb 23. [Epub ahead of print]

19.

Friel

Diehl-Jones

Cockell

. Evidence of oxidative stress in relation to feeding type during early life in premature infants. Pediatr Res. 2011;69(2):160-164.

20.

Korchazhkina

Jones

Czauderna

Spencer

. Effects of exclusive formula or breast milk feeding on oxidative stress in healthy preterm infants. Arch Dis Child. 2006;91(4):327-329.