Keywords
Small for gestational age - growth restricted - extremely low birth weight - premature
- mortality - morbidity
Introduction
Predicting outcomes for growth-restricted extremely low birth weight (ELBW) infants
remains difficult, even as survival of babies at the “threshold of viability” improves.[1]
[2]
[3]
[4]
[5]
[6] Accurate prognostic information is vital to inform shared decision-making regarding
the care of ELBW babies who are also small for gestational age (SGA). Although evidence
supports that growth restriction is associated with increased mortality and morbidity
for term babies,[7] it is unclear whether this holds true for ELBW newborns, who have inherent increased
risks.[5] Few studies specifically investigate outcomes for growth-restricted ELBW babies.
Conflicting outcome data for growth restriction in very low birth weight (VLBW) and/or
extremely preterm babies confuses informed counseling of expectant parents regarding
whether their baby will be at higher risk for mortality and excess morbidity. While
some researchers have found growth restriction is associated with increased mortality
in preterm babies,[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19] others have shown that the neonatal mortality rate is no higher.[20]
[21]
[22] Although traditional thinking may be that growth restriction is associated with
poorer outcomes at any gestation, contrary evidence suggests that being SGA may not
be an important independent prognostic variable for the smallest preterm neonates.
Most previous works suggest no increased risk of intraventricular hemorrhage (IVH)[9]
[11]
[15]
[16]
[21] nor periventricular leukomalacia (PVL)[11]
[15]
[16]
[18]
[19] in growth-restricted preterm babies. Contradictory findings have been reported.[14] Whether gastrointestinal complications such as necrotizing enterocolitis (NEC) or
gastrointestinal perforation are more common[10]
[14]
[18]
[21]
[22] or not[16]
[19] for growth-restricted ELBW babies is also unclear. Perhaps most controversially,
there has been a debate in the literature regarding the respiratory complications
of preterm SGA babies. Controversy remains as to whether respiratory distress syndrome
(RDS) is increased[10] or unaffected/lowered[12]
[13]
[15]
[16]
[20]
[22] with growth restriction, but most authors agree that growth restriction increases
the risk of chronic neonatal lung disease (CNLD).[11]
[13]
[14]
[16]
[23]
[24]
[25]
Yamakawa et al reported that the risk of specific complications is dependent upon
the deviation of the birth weight from the mean and that each complication has a different
threshold for increased risk[18] suggesting that applying findings from studies of bigger babies to ELBW newborns
may not be appropriate. There is a paucity of evidence for the growth-restricted ELBW
group, warranting further investigation. Babies born <1,000 g have increased risks
of neonatal morbidity and mortality, even when compared to babies in the 1,000 to
1,500 g cohort. In 2017, Hasthi et al reported that ELBW babies had a higher risk
than VLBW babies of all studied neonatal morbidities (including a significant increase
in continuous positive airway pressure support) and significantly higher neonatal
death.[12] Given that ELBW babies are known to have increased complication rates, our study
investigated whether growth restriction has an additional detrimental influence on
mortality and short-term morbidity for babies born <1,000 g.
The aim of this study was to determine whether growth restriction is associated with
increased mortality (primary outcome) and short-term morbidity in babies born <1,000 g
and treated at Townsville University Hospital (TUH).
Methods
This study was conducted at TUH, a tertiary perinatal center uniquely positioned to
care for most neonates born extremely preterm and with low birth weight in North Queensland.
The unit provides care for a region with 10,000 births per year.[26] This paper describes the outcomes for ELBW babies treated between January 1, 2010
and January 1, 2021. The hospital has on-site maternal–fetal-medicine specialists
and a level six neonatal intensive care unit. All newborns with birth weight of <1,000 g
admitted to the neonatal unit between January 1, 2010 and January 1, 2021were identified
for inclusion in this study. Data were collected from the neonatal unit's NeoDATA
Microsoft Access database for retrospective analysis. Information is entered into
this database contemporaneously during care by the neonatal team. Patient charts and
imaging reports were accessed when data were not recorded in the database and to confirm
data accuracy for a selection of babies. Only ELBW newborns treated by the neonatal
team were included and those with syndromes incompatible with life were excluded.
In line with previous researchers, we have used birth weight below the 10th centile
for gestational age to define SGA.[7]
[27] Fenton Preterm Growth Charts were used[28]
[29] with birth weight plotted according to gestation (as per obstetric records) and
sex. Babies were categorized as SGA or non-SGA (birth weight ≥10th centile). In this
manuscript, the term “growth restricted” refers to an infant being SGA.
Neonatal particulars were collected for comparison between the two groups including
birth weight, birth gestational age, sex, antenatal steroid administration (one or
more doses), and inborn versus retrieved to TUH. The primary outcome variable was
survival to discharge. Morbidity outcomes collected (and defined below) were the need
for intubation and mechanical ventilation, total duration of respiratory support,
CNLD, home oxygen, IVH, PVL, significant retinopathy of prematurity (ROP), NEC, sepsis
(including all early and late-onset sepsis), time to full enteral feeds, and total
duration of admission.
The need for intubation and positive pressure ventilation was considered an important
surrogate marker for the severity of early respiratory disease. The definition of
CNLD was any form of respiratory support (supplemental oxygen and/or ventilation)
at 36-weeks postmenstrual age, as used by the Australian and New Zealand Neonatal
Network (ANZNN).[30] As home oxygen is a significant outcome for families, it was included in analysis.
All neonates born <32-weeks gestation routinely receive at least two head ultrasound
scans at this center, including an early scan (in the first week of life) and a later
scan (at approximately 5 weeks) for all surviving babies. IVH was classified using
Papile's description.[31] All grades of IVH were considered significant in this study, and any evidence of
PVL was noted. All babies had regular ophthalmology reviews commencing at 30 to 32
weeks corrected age or 28 to 35 days of life. ROP requiring treatment was considered
significant. Any clinical, radiological, and/or surgical diagnoses of NEC of any grade[32] were noted as an outcome. Sepsis was defined as babies with a positive blood culture
who were managed as infected by the treating team.
The Townsville Hospital and Health Service Human Research Ethics Committee (reference
HREC/QTHS/74970) and the Research Governance Officer provided ethical approval for
this study, including a waiver of consent. Analysis was performed using MedCalc Version
18.11.6. Babies were separated into two groups: SGA versus non-SGA. Normally distributed
variables are expressed as mean (standard deviation), nonnormally distributed data
have been expressed as median (interquartile range), and categorical data are presented
as a number and percentage. Chi-square analysis was performed for categorical variables
and Mann–Whitney u-test for continuous variables, to determine significant associations
between SGA and studied outcomes. Regression analysis was used to identify whether
SGA was independently associated with mortality and each short-term morbidity outcome.
Potential confounding factors included in the analysis were birth weight, gestational
age, sex, antenatal corticosteroids, and birth at this center. The impact of being
SGA on each categorical and continuous outcome has been reported as the odds ratio
and beta coefficient, respectively, with 95% confidence interval. An association was
considered significant when p < 0.05. The analysis was repeated for the subgroup of babies born less than 28-weeks
gestation.
Results
A total of 461 ELBW babies met the study inclusion criteria, of whom 62 were SGA and
354 were not. Group demographics are shown in [Table 1]. The overall median birth weight was 800 (684 to 880) g ([Graph 1]) and median gestation 25.6 (24.5 to 27.2) weeks ([Graph 2]). There were comparable numbers of male and female babies in each group. Babies
in the SGA group were significantly smaller than those in non-SGA group: 714 (580
to 850) versus 810 (700 to 885) g (p < 0.001). Conversely, the gestational age at birth for the SGA babies was higher:
28.6 (26.6 to 30.2) versus 25.4 (24.4 to 26.6) weeks (p < 0.001). Antenatal steroid administration of 95% of SGA and 89% of non-SGA babies
was not significantly different (p = 0.124).
Graph 1 Birth weight.
Graph 2 Gestational age at birth.
Table 1
Population characteristics
|
Patient characteristics
|
SGA
|
Non-SGA
|
p-Value
|
|
Total babies
|
62
|
354
|
N/A
|
|
Birth weight (g)
(median [IQR])
|
714 (580–850)
|
810 (700–885)
|
<0.001
|
|
Birth gestational age (weeks)
(median [IQR])
|
28.6 (26.6–30.0)
|
25.4 (24.4–26.6)
|
<0.001
|
|
Sex
Female (n, %)
Male (n, %)
|
27, 14%
35, 16%
|
168, 86%
186, 84%
|
0.570
|
|
Antenatal corticosteroids (n, %)
|
59, 95%
|
314, 89%
|
0.124
|
|
Born at TUH (n, %)
|
51, 82%
|
285, 81%
|
0.747
|
Abbreviations: IQR, interquartile range; n = number to the abbreviation list; SGA, small for gestational age; TUH, Townsville
University Hospital.
There was no significant difference in mortality when the two groups were compared,
with 85% of SGA and 84% of non-SGA babies surviving until discharge ([Table 2]). The most common causes of death in order of decreasing frequency were extreme
prematurity, NEC, IVH, respiratory failure/CNLD, sepsis, hypoxic-ischaemic encephalopathy,
multiorgan failure, intractable seizures, and one death from rhinovirus infection
with neurological and respiratory comorbidities ([Table 3]). Some had more than one cause attributed.
Table 2
Incidence of mortality and morbidity outcomes on univariate analysis
|
Outcomes
|
SGA
|
Non-SGA
|
p-Value
|
|
Survival to discharge (n, %)
|
53, 85%
|
296, 84%
|
0.712
|
|
Intubated for mechanical ventilation (n, %)
|
36, 61%
|
301, 86%
|
<0.001
|
|
Total duration respiratory support (hours) (median [IQR])
|
699.5 (121–1,171)
|
1,239 (758–1,628)
|
<0.001
|
|
Chronic neonatal lung disease¶
|
41, 69%
|
201, 57%
|
0.082
|
|
Home oxygen (n, %)
|
16, 26%
|
79, 22%
|
0.546
|
|
Intraventricular hemorrhage (n, %)
|
7, 11%
|
106, 30%
|
0.002
|
|
Periventricular leukomalacia (n, %)
|
0, 0%
|
10, 3%
|
0.181
|
|
Significant retinopathy of prematurity (n, %)
|
1, 2%
|
24, 8%
|
0.120
|
|
Necrotizing enterocolitis (n, %)
|
2, 3%
|
43, 12%
|
0.037
|
|
Sepsis (n, %)
|
15, 24%
|
99, 28%
|
0.773
|
|
Time to full enteral feeds (days)
(median, IQR)
|
13.5 (10–20)
|
16 (11–25)
|
0.165
|
|
Total duration of admission (days)
(median, IQR)
|
82 (60–114)
|
96 (75–119)
|
0.038
|
Abbreviations: IQR, interquartile range; n = number to the abbreviation list; SGA, small for gestational age.
Table 3
Causes of mortality
|
Cause of death
|
Total number of cases
|
Number SGA
|
|
Extreme prematurity
|
30
|
6
|
|
Necrotizing enterocolitis
|
14
|
2
|
|
Intraventricular hemorrhage/intracerebral bleed
|
12
|
0
|
|
Respiratory failure/chronic neonatal lung disease
|
10
|
6
|
|
Sepsis
|
8
|
0
|
|
Hypoxic ischemic encephalopathy/perinatal asphyxia
|
3
|
1
|
|
Multiple organ failure
|
2
|
0
|
|
Intractable seizures
|
1
|
0
|
|
Viral infection
|
1
|
0
|
Abbreviation: SGA, small for gestational age.
Note: Some babies had more than one cause attributed to their death.
Univariate analysis revealed significant associations between growth restriction and
five morbidity outcomes ([Table 2]) and in each instance the morbidity was less frequent in the SGA group. The outcomes
less common in SGA babies were requirement for intubation and mechanical ventilation
(p < 0.001), total duration of respiratory support (p < 0.001), IVH (p = 0.002), NEC (p = 0.037), and total duration of admission (p = 0.038). SGA babies also had a significantly shortened period of respiratory support,
699.5 (121 to 1,171) versus 1,239 (758 to 1,628) hours (p < 0.001). There were higher rates of CNLD in the SGA (69%) compared to the non-SGA
(57%), but this was not statistically significant (p = 0.082). There was no difference in home oxygen prescription: SGA group 26% versus
non-SGA group 22% (p = 0.546). SGA babies had a shorter median duration of admission by 14 days, but the
range of admission duration was broad for both groups: SGA 82 (60 to 114) versus non-SGA
96 (75 to 119) days (p = 0.038). IVH was less common in the SGA (11 vs. 30%, p = 0.002). There was no significant difference in the rates of PVL (0 vs. 3%, p = 0.181). Significant ROP, sepsis, and time to full enteral feeds were similar between
the SGA and non-SGA groups.
Regression analysis demonstrated no significant association between SGA and any outcome
when correcting for birth weight, gestational age, sex, antenatal corticosteroid administration,
and inborn status ([Tables 4] and [5]). Higher birth weight, more advanced gestational age at birth, and receipt of antenatal
steroids significantly improved survival ([Table 6]). When the analysis was repeated for the subgroup of babies born at less than 28-weeks
gestation, the results were similar (data not shown).
Table 4
Impact of growth restriction (small for gestational age) on morbidity outcomes using
multiple logistic regression analysis controlling for birth weight, gestational age,
sex, antenatal steroids, and birth place
|
Outcome
|
OR
|
95 CI
|
p-Value
|
|
Intubated for mechanical ventilation
|
0.839
|
0.250–2.817
|
0.777
|
|
Chronic neonatal lung disease
|
0.607
|
0.238–1.552
|
0.297
|
|
Home oxygen
|
0.502
|
0.174–1.452
|
0.204
|
|
Intraventricular hemorrhage
|
1.774
|
0.533–5.907
|
0.350
|
|
Significant retinopathy of prematurity
|
3.755
|
0.258–54.611
|
0.333
|
|
Necrotizing enterocolitis
|
2.946
|
0.469–18.496
|
0.249
|
|
Sepsis
|
1.140
|
0.410–3.171
|
0.874
|
Abbreviations: CI, confidence interval; OR, odds ratio.
Table 5
Impact of growth restriction (small for gestational age) on total duration of respiratory
support, time to full enteral feeds and total duration of admission using multivariate
linear regression analysis controlling for birth weight, gestational age, sex, antenatal
steroids, and birth place
|
Outcome
|
β coefficient
|
95 CI
|
p-Value
|
|
Total duration respiratory support (hours)
|
15.377
|
−265.281–296.037
|
0.914
|
|
Time to full enteral feeds (days)
|
1.304
|
−5.322–7.930
|
0.699
|
|
Total duration of admission (days)
|
−2.692
|
−24.043–18.698
|
0.806
|
Abbreviation: CI, confidence interval.
Table 6
Factors associated with survival to discharge determined using multiple logistic regression
analysis
|
Variables
|
OR
|
95% CI
|
p-Value
|
|
Birth weight
|
1.003
|
1.000–1.008
|
0.030
|
|
Gestational age at birth
|
1.522
|
1.121–2.068
|
0.007
|
|
Sex
|
1.442
|
0.792–2.625
|
0.231
|
|
Antenatal corticosteroids
|
3.005
|
1.328–6.799
|
0.008
|
|
Birth place (inborn vs. out born)
|
0.963
|
0.470–1.974
|
0.918
|
|
Growth restriction (SGA vs. non-SGA)
|
1.444
|
0.371–5.615
|
0.596
|
Abbreviations: CI, confidence interval; OR, odds ratio; SGA, small for gestational
age.
Discussion
Growth restriction was not an independent predictor of poor prognosis nor increased
mortality for ELBW babies in this series. This supports the findings of recent papers
showing no excess mortality for SGA babies born at moderate-to-late preterm gestations[21]
[22] and at <30-weeks gestation.[20] After correcting for confounders, there were no significant associations between
growth restriction and any measure of neonatal morbidity in our study. On univariate
analysis, SGA babies were significantly less likely to need intubation for mechanical
ventilation and they had a shorter duration of respiratory support, shorter admission,
less IVH, and less NEC. These associations were no longer significant after correcting
for birth weight, gestational age, sex, antenatal steroids, and birthplace. Given
that demographics between the two study groups were otherwise largely comparable,
the most likely explanation is a protective effect from the comparatively advanced
gestational age in the SGA group. With increasing gestational age, the proportion
of growth-restricted babies increases. This effect has been described by Yamakawa
et al although they ultimately concluded that growth-restricted extremely preterm
babies had increased mortality and morbidity risks in their cohort.[18] After controlling for the relatively advanced gestational age of our SGA babies,
the two groups had comparable outcomes for all neonatal morbidity measures.
The need for intubation and mechanical ventilation and duration of respiratory support
was significantly reduced in SGA on univariate analysis. After controlling for known
confounding factors, the SGA did not have an increased rate of severe early respiratory
disease. Our findings cannot be compared directly to studies reporting lower or unchanged
rates of RDS in growth-restricted preterm babies[12]
[13]
[15]
[16]
[20]
[22] as the definitions used are different and the patient weights and ages generally
higher. Our findings complement previous literature that shows growth restriction
is not independently associated with worse early respiratory status in preterm and
small babies. The higher incidence of CNLD in the SGA was interesting, but not statistically
significant in our population. Importantly, although we have defined CNLD as per the
ANZNN, this outcome is analogous to bronchopulmonary dysplasia reported by other researchers.
Given the consensus in the literature that CNLD is more common in growth-restricted
preterm babies,[11]
[13]
[14]
[16]
[23]
[24]
[25] it is possible that this association might reach statistical significance if we
had a larger sample size. An alternative explanation would be that ELBW babies already
have increased CNLD rates compared to the broader population of preterm babies, negating
any additional risk from growth restriction in this group. The rate of CNLD in our
series would support this alternative explanation. Several explanations for the increased
risk of CNLD in growth-restricted babies have been proposed, and animal studies have
shown physical changes in the lungs of growth-restricted animals (including altered
surfactant quantity and activity, altered alveoli size and number, and interstitial
thickening). Long-term implications of newborn growth restriction have been described
including increased asthma, bronchiolitis, and worsened lung function at school age.[7]
[33] Although our data did not show a significant association between growth restriction
and CNLD, further investigation of this is warranted. Parents should be reassured,
however, that discharge on home oxygen is no more likely for a growth-restricted ELBW
baby.
One plausible explanation for why growth restriction is not associated with worse
outcomes in ELBW babies is that obstetric management is different for complicated
pregnancies. The birth plan, including timing of delivery, is influenced by multiple
factors including gestational age and expected fetal growth.[9] A growth-restricted baby is more likely to be semielectively delivered for faltering
growth, or deteriorating maternal health, than a well-grown baby unexpectedly born
preterm. Zeitlin et al have reported 81.8% of babies <10th centile for weight were
born by caesarean section without labor, versus 30.1% of babies in the 50 to 75th
centile for weight. The SGA babies had an increased rate of maternal hypertension/growth
restriction indicated birth and less prolonged rupture of membranes/hemorrhage. Preterm
infants from pregnancies complicated by maternal hypertension have a lower mortality
rate.[11] Infection is a likely confounding factor in the birth of appropriately grown preterm
babies, who may not have been receiving the intensive obstetric care of their growth-restricted
counterparts. Mactier et al recommend a range of interventions to optimize the survival
of extremely preterm babies including antenatal steroids, tocolysis, transfer to a
tertiary obstetric center, magnesium sulfate, deferred cord clamping, intrapartum
fetal heart rate monitoring, and consideration for caesarean section.[1] It is likely that closely monitored growth-restricted babies will have an increased
opportunity to receive these interventions, impacting improved outcomes.
The hypothesis that a growth-restricted fetus will have an adaptive stress response,
increasing endogenous corticosteroids, thereby accelerating their lung maturation
has been proposed.[20]
[22] This hypothesis has come under intense scrutiny, with multiple authors arguing a
lack of evidence.[7]
[33] Our findings suggest that growth restriction in the ELBW cohort is not associated
with worsened early respiratory disease but provides no evidence to support this as
the mechanism.
With a retrospective design, our study has intrinsic limitations. It has been assumed
that correct diagnoses and recording of clinicians occurred at the time of treatment
for each patient, and that this information has been accurately transferred into the
neonatal database. In some instances, data were incomplete or missing, which could
be explained by the transfer of babies back to referring hospitals or incomplete records.
There are differences between authors in how SGA is defined and described,[27]
[34] which limits the external application of study findings. Zaw et al reported an increased
incidence of growth restriction and associations between growth restriction and neonatal
complications when fetal growth standards are utilized.[34] Contradictory evidence is provided by Garite et al who reported adverse outcomes
were only associated with neonatally diagnosed SGA, rather than antenatally diagnosed
growth restriction that was not confirmed after birth.[9] Our study design using postnatal categorization of SGA may underestimate any association
between growth restriction and poorer outcomes. The total number of babies included
in this study was only 461, despite 11 years of recruitment, and the study may not
be powered to detect differences in infrequent outcomes. The scope of our study was
limited to short-term outcomes. Previous research suggests that the implications of
growth restriction may have longer-term effects,[7]
[17] and further study is required in this area.
Conclusion
Growth restriction is not associated with excess mortality nor short-term morbidity
for ELBW babies at our center. This information should aid antenatal parental counseling
for those in the difficult position of expecting an ELBW baby. Clinicians should not
be unnecessarily pessimistic regarding survival when ELBW is complicated by growth
restriction, and the neonatal journey will not necessarily be more complicated for
these babies. This information can reliably inform shared decision-making for the
care of SGA ELBW infants.
