 |
Introduction |
Previous studies of newborn encephalopathy have focused almost
exclusively on the intrapartum causes of "hypoxic ischaemic encephalopathy."1-7 The contribution of intrapartum
events to newborn encephalopathy remains unclear. We report the
intrapartum findings from the Western Australian case-control study of
newborn encephalopathy.8
Subjects and methods
The subjects and methods are as reported in the accompanying
paper.8
 |
Results |
Intrapartum period
Maternal pyrexia, a persistent occipitoposterior position,
and an acute intrapartum event were all labour related events
associated with a significantly increased risk of newborn encephalopathy (table 1). Only nine of the 18 affected infants and none
of the nine control infants whose mothers had experienced pyrexia had a
pathogenic organism isolated from mother or baby. A prolonged interval
from rupture of membranes to delivery, abnormalities in blood
pressure, a nuchal cord, cord prolapse, and shoulder dystocia
were associated with a non-significantly increased
risk.
View this table:
[in this window]
[in a new window]
|
Table 1.
Risk factors for newborn encephalopathy
present in intrapartum period and adjusted for factors before birth and
antepartum
|
|
Onset of labour and final mode of delivery
The final mode of delivery is determined by the delivery plan and
response to intrapartum events. As the delivery plan could not be
determined onset of labour was investigated as a surrogate (table 1).
The same proportion of cases and controls had spontaneous onset of
labour. More case infants than control infants, however, were induced
and fewer case infants were delivered by caesarean sections without
labour.
Overall, a similar proportion of case and control infants were
delivered by caesarean sections (23% (38) and 24% (96),
respectively). Relative to spontaneous vaginal delivery, instrumental
vaginal delivery and emergency section were associated with over a
twofold increased risk of encephalopathy. Only 2.4% (four) affected
infants compared with 14.5% (58) of control infants were delivered by elective section, defined as one planned at least 24 hours before the
procedure (adjusted odds ratio relative to spontaneous vaginal delivery
0.17; 95% confidence interval 0.05 to 0.56). This inverse relation was
not explained by social factors, including health insurance status, as
these had been adjusted for. The documented indications for elective
sections among case and control infants are shown in table 2; previous
caesarean section was the most common.
View this table:
[in this window]
[in a new window]
|
Table 2.
Indications for elective caesarean section
documented by midwife according to whether baby had newborn
encephalopathy (cases) or not (controls)
|
|
To ascertain whether different risk factor profiles explained the
differences in proportion of emergency and elective caesarean sections,
14 practising consultant obstetricians from Perth were asked to develop
a set of criteria which would lead them to recommend an elective
section at term in the interest of the baby. The consensus, which was
developed without knowledge of the study results, comprised intrauterine growth restriction, malpresentation, abnormal antepartum cardiotocography, two previous sections, macrosomia with diabetes or
gestational diabetes, active herpes, and a previous difficult labour.
When we applied these consensus criteria to mothers of case and control
infants (table 3) eligible mothers of case infants were 24 times less
likely (unadjusted odds ratio relative to spontaneous vaginal delivery
24.2; 6.61 to 90.1) than eligible mothers of control infants to have
been sectioned electively. Nearly 40% of the eligible case infants
were eventually delivered by an emergency section and nearly 20% were
delivered instrumentally or by vaginal breech delivery. The consensus
criteria met by eligible mothers are summarised in table 4. This shows
that even in the group that met the consensus criteria there was a
difference in antepartum risk factor profiles between cases and
controls.
View this table:
[in this window]
[in a new window]
|
Table 3.
Details of onset of labour and final mode of
delivery in cases (babies with newborn encephalopathy) and controls by
eligibility for elective caesarean section according to consensus
criteria.* Values are numbers (percentages) of subjects
|
|
View this table:
[in this window]
[in a new window]
|
Table 4.
Consensus criteria met by mothers of
cases (babies with newborn encephalopathy) and controls eligible for
elective caesarean section.* Values are numbers (percentages) of
subjects
|
|
Other intrapartum factors
The presence of an abnormal intrapartum cardiotocogram, meconium
stained liquor, and fetal distress are usually considered to reflect
intrapartum hypoxia and were not included in the adjusted analyses as
they were likely to be along a causal pathway for, or the first signs
of, newborn encephalopathy or were markers of encephalopathy. Inclusion
of these variables in the adjusted analysis would have masked the
effects of other variables that were working through them. Half the
affected infants had intrapartum cardiotocography performed compared
with 30% of control infants. The cardiotocogram was described as
abnormal in 61% of affected infants compared with 37% of control
infants (unadjusted odds ratio 4.43; 1.81 to 10.85). Meconium was
described more commonly in case infants than control infants (33%
v 12%; 3.72; 2.33 to 5.95) and grade III meconium in
particular was much more common in case infants (13% v
1.0%; 16.7; 5.76 to 50.0). Finally, fetal distress during labour was
recorded by the midwife more often in case infants than control infants
(21% v 8%; 3.16; 1.84 to 5.43). For the same reason we
did not include immediate characteristics of the newborn (table 5) in
the adjusted analysis.
View this table:
[in this window]
[in a new window]
|
Table 5.
Immediate characteristics of babies with
encepalopathy (cases) and controls. Values are numbers (percentages) of
subjects
|
|
Contribution of possible intrapartum hypoxia
In an attempt to estimate the proportion of infants who had been
exposed to possible intrapartum hypoxia we used the following modified
criteria: presence of an abnormal intrapartum cardiotocogram or
abnormal fetal heart rate on auscultation or fresh meconium in labour,
or both, together with a 1 minute Apgar score of less than 3 and a 5 minute Apgar score of less than 7.9 Cord pH measurements
were not included because they were performed so infrequently. Thirty
one affected infants (19%) and two control infants (0.5%) fulfilled
these criteria. A further 16 cases did not strictly fulfil the
definition, but there was evidence that they had experienced a
significant intrapartum event which may have been associated with
intrapartum hypoxia (for example, breech presentation, birth before
arrival at hospital, head stuck, Apgar scores not measured). Therefore,
a total of 47 case infants (29%) had evidence of having experienced
intrapartum hypoxia. Only seven of these (4% of all cases), however,
fulfilled the criteria of possible intrapartum hypoxia in the absence
of preconceptional or antepartum abnormalities. Four case infants (2%)
had no recognised antepartum risk factors or evidence of intrapartum
hypoxia and 113 (69%) had only antepartum factors identified (figure
1). Only 15 of these 47 case infants met the consensus eligibility
criteria for an elective caesarean section.
 |
Discussion |
Our results indicate that intrapartum hypoxia alone accounts for
only a small proportion of cases of newborn encephalopathy, and
elective caesarean section had an unexpected inverse association with
newborn encephalopathy.
Role of intrapartum hypoxia
Although 29% of affected infants experienced events traditionally
indicative of birth asphyxia, it does not necessarily follow that
asphyxia was the primary cause of the encephalopathy. While some
intrapartum factors may be single causes
that is, a previously normal
baby who becomes encephalopathic in labour (fig 2, pathway 1)
this was
an uncommon scenario in our study (see fig 1). Other factors may be on
a causal pathway that starts before birth but which includes
intrapartum hypoxia as a contributor (figure 2, pathway 2). For
example, growth restriction alone is associated with newborn
encephalopathy8 and exposure to labour may compound that
damage.10 A further possibility is that the intrapartum
factors are merely markers of damage associated with adverse events
before birth (fig 2, pathway 3). Abnormality on a cardiotocogram,
meconium stained liquor, low Apgar scores, or the need for active
resuscitation may simply reflect previous neurological
compromise.11
A very small proportion of infants had no recognised antepartum risk
factors nor evidence of intrapartum hypoxia, and it remains unclear as
to when their encephalopathy started and what caused it. Over two
thirds of affected infants had only antepartum factors identified.
Together these two groups represent over 70% of cases among which
there was no evidence of adverse intrapartum events. This points to the
antepartum period being of prime aetiological importance in most cases
of newborn encephalopathy.
Infection
Maternal pyrexia in labour was a significant risk factor,
confirming our previous finding.12 Prolonged interval between rupture of membranes and delivery, a risk factor for ascending infection, was more common in cases compared with controls but not
significantly so. Chorioamnionitis is of current interest as a
cause of cerebral palsy in both term13 and
preterm14 infants. The mechanisms of fetal damage,
however, are not known but could include cerebral sepsis, hyperthermia,
or action via inflammatory mediators.15
Caesarean section
The most striking finding relates to mode of delivery. These data
suggest an important inverse association between elective caesarean
section and newborn encephalopathy. There are several possible
explanations for this finding. Chance alone is an unlikely explanation,
as shown by the 95% confidence interval, although mode of delivery was
not one of the initial study hypotheses.12 The results are
also unlikely to be due to biased selection of control subjects. The
control subjects were randomly selected and their final mode of
delivery and all 21 other characteristics of pregnancy, labour, and
infant available for comparison were the same as for all term live
births in Western Australia during the study period.16
There was no evidence of case selection bias as all affected infants
were included and none died before transfer.8 We therefore
conclude that our findings are
real.
A vital distinction, not made in most other studies, is the
differentiation between elective and non-elective
sections.
2 12 17 18
Had we failed to make this
distinction we would have concluded that caesarean section had no
effect on the risk of newborn encephalopathy. When we applied the
eligibility criteria for elective sections we found that eligible case
infants were more than 20 times less likely to be delivered by elective
section than eligible control infants. The reasons for the apparent
differences in the management of labour in the cases and controls are
undoubtedly complex and may reflect genuine differences (see table 4).
Unrecognised high risk features, alternatives to the consensus view,
women's choice of vaginal delivery, or perhaps some undefined factors
which led a pregnancy to result in a baby with encephalopathy may also
have operated to affect the management of delivery. As the definition of an elective caesarean section was one in which there were 24 hours
between the decision and delivery, it is also possible that some of
these women had been booked for an elective section which they did not
receive because they went into labour. On close review of the eligible
cases, however, a maximum of only 20% could possibly fall into this
category.
It is of note that even in those women not meeting the consensus
criteria for elective section, mothers of control infants were
electively sectioned much more commonly than mothers of case infants.
Furthermore, eligible mothers of case infants did not avoid operative
and instrumental delivery but had emergency rather than elective
procedures. Non-elective sections involve inherently more operative and
postoperative risk, reflected in the lower maternal morbidity after
elective sections.19 In addition, the baby delivered by a
non-elective section has usually been exposed to the stresses of
labour, and this may have an independent impact on outcome.
Elective caesarean sections may exert their apparent beneficial effects
by avoiding some of the intrapartum risk factors for encephalopathy.
For example, elective sections prevent exposure to post-maturity,
persistent occipitoposterior position, intrapartum maternal pyrexia,
and catastrophic events in labour. It may be the avoidance of these
factors other than caesarean section per se which contributes to its
apparent benefit.
We readily recognise that there is no "correct rate" of elective
caesarean sections, but it is pertinent to ask whether women who would
benefit most are being identified and given access to this method of
delivery. It is not possible to say from this observational study
whether elective section would have actually changed the outcome in any
of the cases, but it is an obvious question and one worthy of further
investigation. As a trial to answer this question is unlikely ever to
be performed,20 however, observational studies such as
this would probably be our only source of information. It is, however,
pertinent to note that our findings cannot be used to argue on a very
wide basis that disability can be prevented by elective caesarean
section.
Increasingly, the debate about the aetiology of perinatal brain injury
emphasises the relatively small contribution of the intrapartum period.
The presence of antepartum events does not mean that the intrapartum
course did not contribute to the final outcome. Nevertheless, even with
the best care not all potentially damaging intrapartum events are
avoidable. It seems likely, however, that many babies already have
encephalopathy before labour and others, whose reserve is diminished at
the onset of labour, may have less capacity to cope with hypoxia when
it occurs during labour. Elucidating these multiple pathways will
be the only way we can go forward in the prevention of newborn
encephalopathy.
For acknowledgements and details of contributors, funding, and
competing interests please see the accompanying
paper.8