ASSOCIATION BETWEEN FLUID INTAKE AND WEIGHT LOSS DURING THE

FIRST TEN DAYS OF LIFE AND RISK OF BRONCHOPULMONARY

DYSPLASIA IN EXTREMELY LOW BIRTH WEIGHT INFANTS
WILLIAM OH, MD, BRENDA B. POINDEXTER, MD, REBECCA PERRITT, MS, JAMES A. LEMONS, MD, CHARLES R. BAUER, MD,
RICHARD A. EHRENKRANZ, MD, BARBARA J. STOLL, MD, KENNETH POOLE, PHD,
AND LINDA L. WRIGHT, MD, FOR THE NEONATAL RESEARCH NETWORK

Objective To demonstrate the association between fluid intake and weight loss during the first 10 days of life and the risk of
bronchopulmonary dysplasia (BPD) in extremely low birth weight (ELBW) infants.
Study design A retrospective analysis of data from a cohort of ELBW infants enrolled in the Neonatal Research; 1,382
infants with birth weight between 401 and 1,000 g were randomized. The daily fluid intake and weight loss during the first 10
days of life were compared between the infants who survived without BPD and those who either died or developed BPD.
Demographic and clinical neonatal variables were also compared. Multivariate logistic regression was used to analyze the effect
of fluid intake and weight loss on death or BPD, controlling for demographic and clinical factors that are significantly associated
with BPD by univariate analysis.
Results 585 infants survived without BPD and 797 infants either died or developed BPD. Univariate analysis showed that the
daily fluid intakes were higher (day 2-10) and weight loss less (day 6-9) in the group of infants who either died or developed
BPD. In addition, lower birth weight, lower gestational age, male gender, lower 1 and 5-minute Apgar Scores, higher oxygen
requirement at 24 hours of age, longer duration of assisted ventilation, use of postnatal steroids for BPD and presence of severe
intraventricular hemorrhage, proven necrotizing enterocolitis, patent ductus arteriosus, and late onset sepsis, were associated
with higher incidence of death or BPD. The adjusted risk of higher fluid intake and less weight loss during the first 10 days of life
remained significantly related to death or BPD.
Conclusion In this cohort of ELBW infants treated during the post surfactant era, higher fluid intake and less weight loss
during the first 10 days of life were associated with an increased risk of BPD. The finding suggests that careful attention to fluid
balance might be an important means to reduce the incidence of BPD. (J Pediatr 2005;147:786-90)

ronchopulmonary dysplasia (BPD) is the most common morbidity among extremely

B low birth weight (ELBW) survivors.1 The pathogenesis of BPD is multi-factorial,

including immaturity, barotrauma or volutrauma and oxygen toxicity.2 Excessive
fluid and sodium intakes in these high-risk neonates during the early postnatal period has
been suggested as an additional risk factor for the development of BPD.3-6 From the Neonatal Research Center,
Body water content is very high in ELBW infant and a large proportion of it is in Bethesda, MD.
the extracellular fluid (ECF) compartment.7,8 During the first week of life, there is Supported by Grant support: This
study is supported by the following
a physiologic contraction of the ECF with negative fluid and sodium balances.9,10 The grants: HD U10 21397, HD U10
negative fluid balance allows for the physiologic contraction of ECF which is associated 27871, HD U10 27904, HD U10 27851,
HD U10 27856, HD U10 36790.
with weight loss during the early neonatal period. This is achieved by fluid intake that is
Submitted for publication Jan 20,
less than the amount of water excreted through the kidney in the form of postnatal 2005; last revision received May 12,
diuresis10,11 and via insensible water loss. It is postulated that this physiologic process of 2005; accepted Jun 21, 2005.
ECF contraction may not occur if excessive fluid and or sodium is given during the critical Reprint requests: William Oh, MD,
Department of Pediatrics, Women
period. Bell et al. have shown that high fluid intake with persistent expanded ECF is and Infants’ Hospital of RI, 101 Dudley
associated with a higher incidence of symptomatic patent ductus arteriosus (PDA)12 and St, Providence, RI 02905. E-mail:
woh@wihri.org.
0022-3476/$ - see front matter
BPD Bronchopulmonary dysplasia IVH Intraventricular hemorrhage Copyright ª 2005 Elsevier Inc. All rights
ECF Extracellular fluid NEC Necrotizing enterocolitis reserved.
ELBW Extremely low birth weight PDA Patent ductus arteriosus
10.1016/j.jpeds.2005.06.039

786
necrotizing enterocolitis (NEC).13 There is also suggestive Table I. Demographic and clinical characteristics
evidence that PDA is associated with an increased incidence of the study subjects
of BPD.14 It is possible that, with the retention of ECF and
the presence of PDA with left-to-right shunt, there may be a BPD-free
higher fluid content in the pulmonary interstitial tissue lead- Death or BPD Survivors
ing to decreased lung compliance and the need for greater (n = 797) (n = 585) P value
respiratory support in the form of oxygen administration and
mechanical ventilation. The latter may result in lung injury Birth weight (g) 736 ± 138* 815 ± 128 ,.001
and BPD. Gestational age (wks) 25.4 ± 1.8 26.7 ± 2.0 ,.001
The current study was undertaken to evaluate the Male 418 (52) 246 (42) ,.001
association between fluid intake and postnatal weight loss and Race NS
the development of BPD among ELBW survivors treated in Black 364 (46) 276 (47)
the post-surfactant era. We hypothesized that high fluid in- White 301 (38) 213 (36)
take and less weight loss during the first ten days of life are Hispanic 116 (15) 87 (15)
associated with an increased risk of BPD in ELBW infants. Other 16 (2) 9 (2)
Hypertension/ 166 (21) 188 (32) ,.001
preeclampsia
METHODS Antenatal antibiotics 559 (70) 382 (65) NS
The Neonatal Research Network recently completed Antenatal steroids 627 (79) 476 (81) NS
a randomized controlled trial (October,1999-August,2001) to 1-min. Apgar ,.001
evaluate the efficacy of parenteral glutamine supplementation #3 331 (42) 190 (33)
to reduce the incidence of late onset sepsis in ELBW infants 4-5 222 (28) 134 (23)
who survived beyond the first 12 hours of life.15 During the $6 244 (31) 261 (45)
course of the trial, the data on daily fluid intake and weight 5-min. Apgar
changes of study subjects were collected prospectively. Mor- #3 90 (11) 27 (5)
tality and incidence of BPD were also documented. 4-5 121 (15) 70 (12)
A total of 1433 infants weighing between 401 and 1,000 $6 586 (74) 488 (83)
grams at birth were randomized in the trial; 51 of these infants FiO2 @ 24 hours 0.37 ± 0.19 0.32 ± 0.17
were excluded from analysis because data (maximum weight Days on assisted 49.4 ± 31.0 23.1 ± 20.1
loss, Apgar score, intraventricular hemorrhage [IVH] and ventilation
late onset sepsis) were unavailable for multiple logistic regres- Postnatal steroid 377 (47.3) 85 (14.5)
sion analysis. Thus, 1,382 infants constituted the cohort for Grade 3 and 4 IVH 192 (24) 59 (10)
analysis. Daily fluid intake and sodium supplementation were Proven NEC 99 (12) 36 (6)
prescribed at the discretion of the clinicians. No consistent PDA 436 (55) 219 (37)
policies were in place in the participating units. The parenteral Late onset sepsis 367 (46) 193 (33)
fluid intake was recorded by the research team until the *M ± SD () = % NS not significant.
infant reached full enteral feeding (110 Kcal/kg/day). The
parenteral fluid intake included intravenous fluid administered
We performed statistical analysis using Chi-square tests
as maintenance, as well as those given in the form of blood
for categorical data and Kruskal-Wallis non-parametric tests
transfusion or fluid used to flush intravascular lines after blood
on continuous co-variates. Multivariate logistic regression
sampling or administration of medications. Sodium intakes
analysis was performed to identify risk factors for BPD by
were not recorded. Daily enteral feeding volumes were also
modeling with factors that were shown to be statistically
recorded. Body weights were measured using electronic scales
associated with death or BPD by univariate analysis. The
as part of routine care. Weighing was done daily but waived if
factors found insignificant by univariate analysis were not
infants were considered too sick to tolerate the procedure as
included in the model. Analyses were performed at the
determined by clinician or nursing staff.
Research Triangle Institute, using SAS software.
Maternal and neonatal demographic and clinical char-
acteristics of the subjects were abstracted from the medical
record. Outcome variables recorded included mortality, BPD RESULTS
(defined as oxygen therapy at 36 weeks post-menstrual age to Of the 1,382 infants in the study cohort, 58% (n = 797)
maintain an adequate range of oxygen saturation as deter- either died (n = 224) or developed BPD (n = 573), while 42%
mined by each center), surfactant use for respiratory distress survived without BPD (n = 585). Compared to survivors
syndrome, occurrence of grade 3-4 IVH (by head ultrasonog- without BPD, infants who died or developed BPD had
raphy), NEC, (Bell’s classification $stage 216), PDA deter- significantly lower birth weight and gestational age, more male
mined by echocardiography, and blood culture proven late infants, lower 1 and 5-minute Apgar scores, higher FiO2
onset sepsis. Data on daily electrolyte and serum electrolyte requirement at 24 hours of age,longer duration of assisted
intake were not collected. ventilation, more use of postnatal steroids for BPD and higher
Association Between Fluid Intake And Weight Loss During The First
Ten Days Of Life And Risk Of Bronchopulmonary Dysplasia In
Extremely Low Birth Weight Infants 787
Table II. Daily total fluid intake (parenteral and enteral) in the study groups (mL/kg)
Death or BPD BPD-free Survivors
Age
(day) N Mean SD Range N Mean SD Range P value

2 795 136 41 46-332 581 118 32 66-297 ,.001

3 788 158 51 48-466 584 134 34 55-309
4 766 170 46 60-390 584 147 33 51-326
5 763 171 40 67-329 583 154 31 63-294
6 760 169 37 56-345 583 157 30 84-344
7 756 165 35 46-389 582 156 29 66-261
8 751 163 33 41-339 582 153 27 69-259
9 745 159 35 56-362 583 152 26 86-315
10 736 158 35 56-391 581 150 25 74-260
Day 1 data not presented because it only represents partial intake for the day.
Discrepancy in numbers of subject in each day is due to either death or lack of data recorded because the weighing was waived for clinical reasons.

Table III. Enteral intakes in the study Multivariate logistic regression analysis adjusting for
subjects (mL/kg) factors that were shown to be associated with the development
of BPD by univariate analysis demonstrated that higher fluid
Age Death BPD-free intake and lack of weight loss remains significantly associated
(day) n or BPD n Survivors P value with higher risk of death or BPD (p < 0.001 and 0.006
respectively) (Table V). Higher birth weight was associated
2 797 0±1 583 1 ±3 ,.001
with lower risk for death or BPD. Male gender, severe IVH,
3 789 1±4 585 3 ±8
NEC, PDA, longer duration of assisted ventilation and more
4 777 3±7 584 6 ± 13
frequent use of postnatal steroids were associated with higher
5 767 5 ± 11 584 10 ± 18
risk for death or BPD. There is a also signifcant center
6 762 7 ± 14 583 15 ± 25
variation on the incidence of death or BPD (p < 0.001).
7 759 10 ± 18 583 21 ± 30
Factors not significantly associated with death or BPD
8 755 13 ± 23 582 26 ± 34
included gestational age, maternal hypertension, use of
9 750 16 ± 26 583 31 ± 38
antenatal antibiotics, Apgar Score at 1 minute, higher FiO2
10 741 19 ± 30 581 37 ± 41
@ 24 hours, and late onset sepsis.
Day 1 data not presented because it only represents partial intake
for the day.
Discrepancy in numbers of subject in each day is due to either death of
lack of data recorded. DISCUSSION
This retrospective analysis of prospectively collected
incidence of severe IVH, proven NEC, PDA and late onset data confirmed our hypothesis that higher fluid intake and
sepsis (Table I). Maternal history of hypertension and pre- lack of postnatal weight loss during the immediate postnatal
eclampsia was associated with lower incidence of death or period predisposed this high-risk group of infants to BPD.
BPD. Infants who died or developed BPD had significantly Death was included as a primary outcome because it is a
higher total fluid intakes (parenteral and enteral) from the competing outcome with BPD. There was a concern that
2nd to the 10th day of life (Table II), despite the fact that the including death as variable may include those infants who died
enteral fluid intake is higher among survivors without BPD because of physicians’ decision to withdraw care due to
(Table III). Because packed red blood cells used in transfu- extreme immaturity and serious morbidity, such as severe IVH.
sion (not much water content) is included in the total fluid However, many of these deaths will likely occur during the
intake, we recalculated the data listed in table 2 without the 12 hours of life and would have been ineligible for participa-
transfusion volume. The recalculation shows that the differ- tion in the trial. Furthermore, we analyzed the cause of death
ence in fluid intake between the 2 groups remains significant of those who died, only 14 of 223 had extreme immaturity thus
(data not shown). Weight loss (expressed as percent of birth accounting for 6% of the cohort which will likely not affect
weight) in infants who died or developed BPD were consis- the outcome of analysis even if all of these deaths were due
tently less than those who survived without BPD and was to phycicians’ decision to withdraw care.
statistically significant between day 6-9 (Table IV). The nadir The pathogenesis of BPD is multifactorial involving
of weight loss occured on day 5 for both groups. The in- prematurity, respiratory failure requiring oxygen administra-
creasing number of available weight data with increasing age tion and mechanical ventilation.2 The significant association
reflects the fact that more weighings were waived in the first between death or BPD and duration of assisted ventila-
few days of life because of high acuity of illness. tion, and use of postnatal steroid confirm this concept.

788 Oh et al The Journal of Pediatrics  December 2005

Table IV. Weight changes during the first ten days Table V. Odds for death or BPD adjusted for
(expressed as percent change of birth weight) selected clinical and demographic variables by
multivariate logistic regression analysis
Age BPD or
(day) n death n Neither P value Risk of death or BPD

2 395 20.1 ± 6.3 341 20.4 ± 6.1 .54 Variable Odds Ratio 95% CI P values
3 480 22.8 ± 8.1 432 23.8 ± 6.6 .10
Higher fluid intake 1.011 1.005-1.017 ,.001
4 514 26.7 ± 8.9 468 27.3 ± 8.4 .24
Maximum weight loss* 1.028 1.008-1.049 .006
5 548 28.1 ± 9.2 482 29.0 ± 8.0 .25
Higher birth weighty 0.840 0.735-0.961 .011
6 516 27.4 ± 9.5 475 28.7 ± 8.3 .05
Male sex 1.411 1.067-1.867 .016
7 529 26.3 ± 9.9 489 28.0 ± 8.5 .01
Grade 3-4 IVH 3.025 1.971-4.642 ,.0001
8 605 24.7 ± 10.6 519 26.4 ± 9.2 .02
NEC 2.900 1.732-4.855 ,.0001
9 588 23.3 ± 10.0 506 24.7 ± 9.1 .04
PDA 1.373 1.010-1.868 .043
10 577 21.8 ± 11.0 514 23.1 ± 10.2 .07
Days on Vent 1.032 1.024-1.040 ,.0001
Day 1 data not presented because it only represents partial intake Postnatal Steroids 2.018 1.418-2.870 ,.0001
for the day.
Gestation 0.966 0.876-1.066 .496
Discrepancy in numbers of subject in each day is due to either death or
lack of data recorded because the weighing was waived for clinical reasons. Maternal hypertension 1.198 0.830-1.731 .334
Antenatal antibiotics 1.257 0.910-1.737 .165
Apgar score @ 1 min 1.012 0.951-1.077 .697
Inappropriately high fluid intake with retention of extracel- FiO2 @ 24 hours 1.559 0.663-3.665 .309
lular fluid may lower lung compliance, exacerbate respiratory Late onset sepsis 1.113 0.822-1.508 .488
compromise and lead to an additional need for oxygen and
*Coded as maximum wt. Loss as percent of birth weight during
mechanical ventilation. The association of higher fluid intake the first 10 days.
and BPD is consistent with previous observations in regards to yCoded as 100-g increment.
excessive fluid intake and development of BPD. Van Marter,
et al.3 demonstrated the association between excess fluid and note that the association of BPD and PDA is consistent with
colloid intake and less weight loss in very low birth weight previous observation.14
infants who were oxygen-dependent at 28 days of age. The Studies by Costarino4 and Hartnoll 5,6 demonstrated
current definition of BPD is oxygen-dependence at 36 weeks the association of excessive sodium intake and BPD. When
completed postmenstrual age.17 We used the latter definition sodium intake is in excess of the requirement, a positive
rather than the 28-day definition. Bell et al18 reviewed the sodium balance occurs, which results in fluid retention in the
results of three randomized trials19-21 on fluid restriction and extracellular space with end results similar to those of excess
BPD, and concluded that although there is a trend of lower fluid intake. Although these studies showed significant
incidence of BPD in preterm infants who received restricted association between sodium intake and BPD (defined as
fluid intake during the first few days of life, the difference is oxygen dependence at 28 days), the sample size for each study
not statistically significant. They recommended further inves- was relatively small. We can not comment on this issue
tigation of this subject to justify wide spread clinicial practice. because of lack of sodium intake and serum electrolyte data in
Another important message in our data is that the observation our study.
of association between high fluid intake and less weight loss The strength of this research is the large sample size
with death or BPD is made in the era of high surfactant usage. and the fact that the data on fluid intakes were collected
In contrast to previous pre-surfactant era’s observations,3,19,20 prospectively. Another important feature of the current
a large majority of our study subjects received surfactant for report is that the study was conducted in the surfactant treat-
respiratory distress syndrome. ment era in contrast to all other previous reports,3-6 which were
It is of interest that the survivors without BPD have done in the era when surfactant therapy was unavailable. The
higher enteral intakes during the first 10 days, possibly ref- weakness are the retrospective observational data and the lack
lecting healthier clinical status with greater feeding tolerance. of other pertinent data, such as daily sodium intake and serum
The significant association of BPD with higher fluid electrolyte values that may help elucidate the nature of fluid and
intake and less postnatal weight loss suggests that normal electrolyte balance. Observational data are subject to study bias
physiologic transition of extracellular fluid contraction was and other confounding variables.
probably impeded by excess fluid intake, leading to lower lung Compared to the dismal outcome of infants with BPD
compliance (because of fluid accumulation of pulmonary in the 1980’s,22 the outcome of infants with this disorder has
interstitial tissue), with greater barotrauma and or oxygen not improved significantly. Furthermore, the incidence of
toxicity leading to BPD. However, the significant association BPD has not changed, so the number of infants with this
of less weight loss is only on day 6-9 which does not exactly fit disorder is increasing. BPD is associated with other serious
the concept that the physiologic transition of ECF contraction morbidities and abnormal neuro-developmental outcomes,23
occurs during the first few days of life.9,10 It is of interest to therefore any intervention that could reduce its incidence
Association Between Fluid Intake And Weight Loss During The First
Ten Days Of Life And Risk Of Bronchopulmonary Dysplasia In
Extremely Low Birth Weight Infants 789
would be of great clinical importance. Adequately powered
randomized control trials are needed to determine whether the
incidence of BPD can be reduced with management strategies
controlling fluid and sodium administration.
REFERENCES
1. Lemons JA, Bauer CR, Oh W, Korones SB, Papile L, Stoll BJ, Verter J,
Temprosa M, Wright LL, Ehrenkranz RA, Fanaroff AA, Stark AR,
Carlo WA, Tyson JE, Donovan EF, Shankaran S, Stevenson DK.
Very-low-birthweight outcomes of the NICHD Neonatal Research
Network, January 1995 through December 1996. Pediatrics 2001;1:107.
2. Bancalari E, Claure N, Sosenko IR. Bronchopulmonary dysplasia: changes
in pathogenesis, epidemiology and definition. Semin Neonatol 2003;8:63-71.
3. Van Marter LJ, Leviton A, Allred EN, Pagano M, Kuban KC.
Hydration during the first days of life and the risk of bronchopulmonary
dysplasia in low birth weight infants. J Pediatr 1990;116:942-9.
4. Costarino AT, Gruskay JA, Corcoran L, Polin RA, Baumgart SO.
Sodium restriction versus daily maintenance replacement in very low birth
weight premature neonates: A randomized, blind therapeutic trial. J Pediatr
1992;120:99-106.
5. Hartnoll G, Betremieux P, Modi N. Randomized controlled trial of
postnatal sodium supplementation on ‘‘oxygen dependency and body weight in
25-30 week gestational age infants.’’ Arch Dis Child 2000;85:F29-32.
6. Hartnoll G, Betremieux P, Modi N. Randomized controlled trial of
postnatal sodium supplementation on body composition in 25-30 week
gestational age infants. Arch Dis Child 2000;82:F24-8.
7. Friis-Hansen B. Changes in body water compartment during growth.
Acta Paediatrica Scand 1957;110:S1-68.
8. Friis-Hansen B. Body water compartment in children: Changes during
growth and related changes in body composition. Pediatrics 1961;28:169-81.
9. Stonestreet BS, Bell EF, Warburton D, Oh W. Renal response in low-
birth-weight neonates. Results of prolonged intake of two different amounts
of fluid and sodium. Am J Dis Child 1983;137:215-9.
10. Bauer K, Versmold H. Postnatal weight loss in preterm neonates
less than 1,500 g is due to isotonic dehydration of the extracellular volume.
Acta Pediatrica Scand 1989;369:37-42.
790 Oh et al
11. Bidiwala KS, Lorenz JM, Kleinman LI. Renal function correlates of
postnatal diuresis in preterm infants. Pediatrics 1988;82:50-6.
12. Bell EF, Warburton D, Stonestreet BS, Oh W. Effect of fluid
administration of the development of symptomatic patent ductus arteriosus
and congestive heart failure in premature infants. New Engl J Med 1980;302:
598-604.
13. Bell EF, Warburton D, Stonestreet BS, Oh W. High volume fluid
intake predisposes premature infants to necrotizing enterocolitis. Lancet
1979;2:90.
14. Brown E. Increased risk of bronchopulmonary dysplasia in infants with
patent ductus arteriosus. Pediatrics 1979;95:865-6.
15. Poindexter BB, Ehrenkranz RA, Stoll BJ, Wright LL, Poole WK, Oh
W, et al. Randomized Clinical trial of Parenteral Glutamine Supplementation
Does Not Reduce the Risk of Mortality or Late-Onset Sepsis in Extremely-
Low-Birth-Weight Infants. Pediatrics 2004;113:1209-15.
16. Bell MJ, Ternberg JL, Feigin RD, Keating JP, Marshall R, Barton L,
et al. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon
clinical staging. Ann Surg 1978;187:1-7.
17. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit
Care Med 2001;163:1723-9.
18. Bell EF, Acarregui MJ. Restricted versus liberal water intake for
preventing morbidity and mortality in preterm infants. Cochrane Review. In:
Cochrane Library 2001;3:CD000503.
19. Lorenz JM, Kleinman LJ, Kotagal UR, Reller MD. Water balance in
Very Low-Birth-weight infants: Relationship between water and sodium
intake and effect on outcome. J Pediatr 1982;101:423-32.
20. Tammela OK, Koivisto ME. Fluid restriction for preventing broncho-
pulmonary dysplasia? Reduced fluid intake during the first week of life
improves the outcome of low-birth-weight infant. Acta Paediatrica 1992;81:
207-12.
21. Kavvadia V, Greenough A, Dimitriou G, Hooper R. Randomized trial
of fluid restriction in ventilated very low birth weight infants. Arch Dis Child
Fetal Neonatal 2000;83:F91-6.
22. Vohr BR, Bell EF, Oh W. Infants with bronchopulmonary dysplasia.
Am J Dis Child 1982;136:443-7.
23. Vohr BR, Wright LL, Dusick AM, Mele L, Verter J, Steichen JJ, et al.
Neurodevelopmental and functional outcome of extremely-low-birth-weight
(ELBW) infants. The NICHD Neonatal Research Network Follow-up
Study. Pediatrics 2000;105:1216-26.
The Journal of Pediatrics  December 2005