621 - Stream of Evidence: Association of Urine Output in the First Week with In-Hospital Outcomes in Very Preterm Neonates

Publication Number: 1598.621

Olivia C. Brandon, University of Washington School of Medicine, Seattle, WA, United States; Krystle Perez, University of Washington School of Medicine, Snohomish, WA, United States; Sandy Huang, University of Washington School of Medicine, Sammamish, WA, United States; Gabriel Byram, Pacific Northwest University of Health Sciences, Yakima, WA, United States; Clara G.. McAdams, University of Washington School of Medicine, Tacoma, WA, United States; Sophia Rodriguez, University of Washington School of Medicine, Kirkland, WA, United States; Virginia Higgins, University of Washington School of Medicine, Edina, MN, United States; Ineeya Senthilnathan Kayal, University of Washington School of Medicine, Seattle, WA, United States; Natalie Chapkis, Seattle Children's/ University of Washington, Seattle, WA, United States; Elizabeth Riffel, St. Joseph's Children's Hospital, Tacoma, WA, United States; John Haddock, St. Joseph Medical Center, Puyallup, WA, United States; Addysen Pulley, University of Washington School of Medicine, Puyallup, WA, United States; Laura A.. Boyce, Seattle Children's, Tacoma, WA, United States; Jill Irvine, university of washington medical center, seattle, WA, United States; Stephen Welty, University of Washington School of Medicine, Tacoma, WA, United States; Robin G. LaGrandeur, University of Washington School of Medicine, Seattle, WA, United States; Sarah Kolnik, University of Washington - Seattle Children's Hospital, Seattle, WA, United States; Rebecca Hoban, Seattle Children's, Seattle, WA, United States; Katie M.. Strobel, University of Washington School of Medicine, Seattle, WA, United States; Gregory C. Valentine, University of Washington School of Medicine, Tacoma, WA, United States

Background: The first week after birth is a critical period of fluid transition for preterm infants. Immature kidneys, high insensible losses, and extracellular fluid shifts lead to wide variation in urine output (UOP). Although UOP is often used as a proxy for fluid status, its relationship to weight loss, fluid intake, and clinical outcomes in preterm neonates remains poorly understood.

Objective: To (a) examine clinical factors associated with UOP and (b) determine whether average UOP in the first 3 days or first week after birth is associated with in-hospital outcomes.

Design/Methods: We conducted a retrospective analysis among very preterm (VP) infants (born ≤32 weeks' gestation) admitted to two level III/IV neonatal intensive care units. Average UOP (mL/kg birthweight/hr) was calculated for the first 3 and 7 days after birth, excluding infants who died before those time points, respectively. Scatterplots with locally estimated smoothing evaluated relationships between UOP and clinical factors. Logistic and linear regression models, adjusted for confounders including fluids, weight loss, and illness severity, were used to evaluate associations between UOP and the following in-hospital outcomes: bronchopulmonary dysplasia, retinopathy of prematurity, necrotizing enterocolitis, patent ductus arteriosus requiring treatment, culture positive late onset sepsis, mortality or transfer to higher level of care.

Results: Among N=270 infants, the median UOP in the first 3 days was 3.4 (IQR: 2.8-4.0) mL/kg/hr (see Table 1 for infant characteristics). UOP during the first 3 days slightly decreased with higher gestational age and greater enteral intake (Figure 1A) but increased with higher sodium variability, intravenous (IV) fluids, and total fluids (Figure 1C-D,F). Maximal weight loss was unrelated to UOP (Figure 1B). In adjusted analyses, small for gestational age status was significantly associated with lower UOP, whereas higher IV and enteral fluid volumes were associated with greater UOP in both 3- and 7-day analyses (Figure 2A-B). Lower 3-day UOP was significantly associated with increased odds of late onset sepsis (aOR 0.49, 95% CI 0.26-0.95) but not with other outcomes (Figure 2C); 7-day UOP was not associated with any outcomes (Figure 2D).

Conclusion(s): In VP infants, UOP is not an accurate proxy for weight loss in the first week after birth. However, higher UOP during the first 3 days after birth is associated with reduced odds of developing late onset sepsis, suggesting it may be a useful indicator of later sepsis risk.

Table 1. Maternal and infant characteristics among those who had average urine output <3.38 and ≥3.38 mL/kg/hr averaged over the first 3 days after birth.
All data presented as n/N (%), median ± IQR. Mann Whitney U-test or Fisher's Exact Test for continuous and categorical variables, respectively. Only infants surviving at least 3 days included. Only babies surviving to 36 weeks postmenstrual age included in ROP and BPD analyses. ROP was defined as at least grade 1 or higher. Sepsis included culture positive sepsis after the first 3 days. Patent ductus arteriosus requiring intervention was defined as medical, surgical, or device closure. Bronchopulmonary dysplasia was classified as Jensen grade 1 or higher. Only babies discharged home included in length of stay analysis. Significant p-values bolded.
ROP: retinopathy of prematurity. NEC: necrotizing enterocolitis. PDA: patent ductus arteriosus. BPD: bronchopulmonary dysplasia.

Figure 1. Unadjusted relationship between average urine output (mL/kg/hr) in the first 3 days after birth and clinical factors including gestational age (A), maximal weight loss (B), sodium coefficient of variation (C), intravenous fluids (D), enteral feeds (E), and total fluid intake (F).
The purple line represents the local mean determined using locally estimated scatterplot smoothing (loess), and the shading represents the 95% confidence interval of the mean. The sodium coefficient of variation is the ratio of standard deviation to the mean multiplied by 100.

Figure 2. Clinical factors associated with urine output in the first 3 days (A) and first week (B) after birth (top row), and adjusted odds ratios showing associations between urine output and in-hospital outcomes during the same periods (C-D, bottom row).
The estimate shows the expected change in average urine output for a one-unit increase in each factor. Estimates from linear regression models were used to examine associations between average urine output (mL/kg/hr) in the first 3 days or first week after birth and clinical factors including SGA, sex, gestational age, maximal weight loss (continuously), sodium coefficient of variation percentage (ratio of standard deviation to the mean multiplied by 100), vasopressor use (only in 7-day analysis), average enteral fluid intake, and average intravenous fluid intake in the first 3 or 7 days.
In the bottom row, forest plots show adjusted odds ratio per 1mL/kg/day increase in urine output averaged over the 3 days or first week after birth and association with in-hospital neonatal outcomes. Logistic regression models adjusted for gestational age, sodium coefficient of variation (variability), SGA status, site, sex, vasopressor use in first week (only in 7-day analysis), Apgar score at 1 minute, antenatal steroids (full course), maximal weight loss <5% in first week, maximal weight loss >15% in first week, average enteral feeds over 3 or 7 days, and average intravenous fluids over 3 or 7 days. Sepsis included culture positive sepsis after the first 3 or 7 days. Patent ductus arteriosus requiring intervention was defined as medical, surgical, or device closure. Bronchopulmonary dysplasia was classified as Jensen grade 1 or higher. NEC was defined as modified Bell's Stage IIA or higher. Significant associations shown in purple.
SGA: small for gestational age based on <10th birthweight percentile. ROP: retinopathy of prematurity. NEC: necrotizing enterocolitis. PDA: patent ductus arteriosus. BPD: bronchopulmonary dysplasia. aOR: adjusted odds ratio.