Keywords preeclampsia - prematurity - maternal health
Schlüsselwörter Präeklampsie - Frühgeburt - maternale Gesundheit
Introduction
Pregnancy-related hypertensive disorders (PIH), in particular preeclampsia, is one
of the most common causes of maternal and fetal morbidity and mortality [1 ]. Preeclampsia is defined by elevated blood pressure with the newly involvement of
at least one other organ [2 ]. Though any organ can be affected, the kidneys, liver, and the placenta are most
commonly involved [2 ]. This can acutely lead to short-term maternal complications, such as cerebrovascular
bleeding, retinal detachment, HELLP syndrome and eclampsia [3 ].
Impaired placental development early in pregnancy and consequent release of several
placenta-derived factors into the maternal circulation that lead to generalized maternal
endothelial dysfunction is the leading hypothesis of pathogenesis [4 ]
[5 ]
[6 ]. However, there is increasing evidence that suboptimal maternal cardiovascular performance
is likely to be causative for placental dysfunction in preeclampsia [7 ]. This may be one explanation for the known long-term consequences for women with
preeclampsia especially regarding the cardiovascular system [8 ]
[9 ], the
renal system [10 ], thromboembolism [11 ], and neurocognitive impairment [12 ]
[13 ]. Yet, it is controversial whether preeclampsia can be regarded as the pathological
insult that causes the long-term consequences or whether the development of preeclampsia
reveals a (previously unknown) risk condition prior to pregnancy [4 ]
[14 ].
Early-onset preeclampsia before 32 weeks of pregnancy accounts for 0.7% to 1.0% of
all births [15 ]. The timing of delivery demands balancing maternal risks of acute or chronic disease,
intrauterine fetal demise, and sequelae of neonatal preterm birth. International guidelines
differ in their recommendations for the delivery of early-onset preeclampsia before
32 weeks’ gestation. The Swiss-Austrian-German Guideline for the treatment of women
with preeclampsia states: “Before 34 weeks’ gestation, a primarily conservative approach
can be considered depending on the severity of preeclampsia.” However, the definition
of “severe preeclampsia” is not only inconsistent in the literature [16 ]
[17 ]
[18 ]
[19 ]
[20 ]
[21 ]
[22 ], international societies like the International Society for the Study of Hypertension
in Pregnancy (ISSHP), Society of Obstetric Medicine of Australia and New Zealand (SOMANZ),
European Society of Cardiology (ESC), and the Swiss-Austrian-German Guideline on Hypertension
in Pregnancy (German Society of Gynecology and Obstetrics [DGGG]) refrained from defining
“severe preeclampsia” in their guidelines [2 ]
[23 ]
[24 ]. Furthermore, the decision to deliver
is complex and depends on many aspects, especially as the presentation of preeclampsia
and its complications is heterogeneous. The indication for delivery therefore remains
an individual decision, and the term “severe preeclampsia” is very often used in clinical
language to justify an indication for delivery without a clear definition of the term.
The aim of the present study was to define “severe early-onset preeclampsia” based
on consent among obstetricians from Germany, Switzerland, and Austria about the indication
for mandatory preterm birth for imminent maternal complications. This can be considered
a thought experiment. We have assumed that prolongation of pregnancy up to 32 weeks’
gestation to improve fetal maturity is justified until acute or imminent complications
threaten the mother. Consequently, since delivery is the only cure, we define a finding
that forces the obstetrician to suggest delivery as “severe preeclampsia”. With this
in mind, we asked obstetricians for their opinion on various scenarios and constellations
of preeclampsia. By grading these findings for necessary delivery on a Likert scale
from 0 to 4, results of this study may also be used in preparation of future observational
studies to challenge a severity score of preeclampsia.
Methods
Delphi study design
To address the study aim, we used the classic asynchronous Delphi consensus methodology.
This is an iterative technique based on evaluating a series of structured statements.
After revision, this is communicated to the participants and repeated in increasing
detail in several rounds until a consensus can be reached [25 ]
[26 ]. The classic asynchronous Delphi design anticipates the development of unwanted
group dynamics. This prevented the development of trends and tendencies in opinions
that could potentially interfere with good estimates. Furthermore, the individual
online surveys made it easier and safer to reach the experts. This made it possible
to address a larger group of experts.
For the development of questionnaires international clinical practice guidelines were
reviewed for clinical features, symptoms, and laboratory findings related to preeclampsia.
In detail, the practice guidelines of the International Society for the Study of Hypertension
in Pregnancy (ISSHP) 2018, the American College of Obstetricians and Gynecologists
(ACOG) 2013, the Society of Obstetricians and Gynaecologists of Canada (SOGC) practice
guideline on hypertensive disorders in pregnancy 2014, the National Institute for
Health and Care Excellence (NICE) 2019, and the German Society of Gynecology and Obstetrics
(DGGG) 2019 was used [2 ]
[23 ]
[24 ]
[27 ]
[28 ]
[29 ].
Identified parameters were structured into four categories: (i) 19 maternal parameters
that can be measured and ordered on a metric scale, like oxygen saturation, blood
pressure, weight gain, creatinine concentration, proteinuria, platelet count, and
others; (ii) eight maternal clinical diagnoses that can be determined objectively
by the physician but can be coded binary. This includes pulmonary edema, any heart
disease (e.g. low ejection fraction, myocardiac ischemia), hepatic hematoma, hyperreflexia
and other; (iii) 13 maternal symptoms that are subjective to the patient’s perception
but impair the patient’s well-being, and that can be categorized in binary terms,
like nausea, headache, dizziness, abdominal or thoracal pain. It has been specified
that symptoms last over several hours; (iv) nine fetal parameters assessed by ultrasound
or cardiotocography.
A case was then constructed for all Delphi rounds to query delivery criteria and their
weighting: a pregnant woman between 24 + 0 and 31 + 6 weeks’ gestation with diagnosed
pre-eclampsia according to the AWMF criteria (German Guideline for gynecology and
obstetrics) [2 ]. The treatment with antenatal corticosteroids for fetal lung maturity has already
been completed [30 ]
[31 ]. In the first round of the Delphi procedure, participants were asked about their
confidence in suggesting delivery for this case and for each of the 49 parameters
in categories (i) to (iv). For metric scales, obstetricians were asked to indicate
a specific cut-off value at which they would recommend delivery. For the binary parameters,
participants were asked to rate on a Likert scale
according to whether they considered it as no (0 points), mild, moderate, severe,
or absolute delivery criterion (4 points), where “absolute” refers to delivery without
any delay. During each round, participants could provide feedback on existing items
or suggest additional ones for each category. Finally, participants were asked for
demographic characteristics, clinical and academic background. The questionnaires
were initially critically examined within the section “Hypertensive Pregnancy Diseases
and Fetal Growth Restriction” of the German Society of Gynecology and Obstetrics.
The predefined cut-off for group consensus on an item or group of related answers
was ≥ 70%.
Panel selection
The selection of potential panel members was based on their recognized expertise as
heads of the tertiary perinatal centers in Germany, Austria, and Switzerland. In addition,
clinical experts were identified who attended the congress of the German-Austrian-Swiss
branch of the International Society for the Study of Hypertension in Pregnancy in
Bern, Switzerland, in 2019. In total 126 participants were invited to participate.
In the Delphi process, the votes of all committee members are weighted equally. Experts
who did not participate in a particular round were not invited to participate in subsequent
rounds. Ethical approval was obtained from University Medical Centre of Kiel (D 404/21).
All participants provided informed consent before commencing the first round, and
they were reminded of their right to withdraw before each subsequent round.
Data collection and analyses
The Delphi process started with round one in 11/2019 followed by round two from 12/2020
to 02/2021, round three from 02/2021 to 04/2021, round four from 05/2021 to 08/2021,
and round five from 09/2021 to 11/2021. The first round of the Delphi process involved
a questionnaire sent by post. The remaining four rounds were conducted online. The
questionnaires were completed using the online tool SurveyMonkey (https://de.surveymonkey.com ). In each round, a unique link (token-secured) to the questionnaire was sent to the
panel members via email. The results of the questionnaires for each round were shared
with the participants in the next round. The results were presented anonymously at
the group level. Those who did not respond received reminder emails every 10 to 14
days until 8 weeks. Withdrawal from the proceedings was offered at any time.
Delphi rounds
After the first round, all examined 49 parameters were evaluated descriptively. If
≥ 70% of respondents agreed or rejected a value or parameter as an indication of delivery,
it was accepted or rejected. The parameters without consensus were checked for relevance
and, if necessary, the question was reformulated or made more precise. All further
rounds were conducted via an online survey tool. In the second round, a precision
of category (ii) (newly occurring clinical finding) and category (iii) (newly occurring
clinical symptom) from round one was carried out. The respective finding or symptom
was described in more detail to achieve a better assessment. In round three, parameters
from category (i) have been specified and dynamic changes have been added. By defining
cut-offs based on the participants’ feedback a purely categorical query was carried
out on a 5-point Likert scale. In round four, the parameters from category (ii) and
(iii) which still failed to reach consensus
were examined again. The classification into the respective categorical query, based
on the majority opinion from the previous surveys, was presented. The panel was asked
whether they agreed or disagreed. In round five, categories (i) and (iv) were reassessed.
The classification into the respective categorical query, based on the majority opinion
from the previous surveys, was presented. The panel was asked whether they agreed
or disagreed. For various fetal parameters (category [iv]), it was checked whether
the wording of the current national guideline on fetal growth restriction “delivery
should be considered” was equivalent to a “more serious relative delivery criterion”
in the opinion of the panel.
Within the five Delphi rounds, seven additional parameters were included by clarifying
the parameters and questions and through feedback from the participants, so that a
total of 56 parameters finally were considered as an indication for delivery. A flowchart
of the development, the process and the participation per Delphi round is shown in
[Fig. 1 ].
Fig. 1
The Delphi procedure. A flowchart of the development, the process and the participation
per Delphi round. DGGG = German Society of Gynecology and Obstetrics, ISSHP = International
Society for the Study of Hypertension in Pregnancy, ACOG = American College of Obstetricians
and Gynecologists, NICE = National Institute for Health and Care Excellence
Results
Participants
Sixty-nine of 126 experts (54.8%) responded to the invitation and took part in the
first round. Of these, 54 agreed to be contacted for further Delphi rounds. The response
rates for round two were 53/54, for round three 52/53, for round four 52/52, and round
five 50/52. Thus, 72.2% (50/69) of participants who began the Delphi procedure completed
the entire procedure. Of all initial 69 participants, 97.1% were senior obstetricians,
of whom 53 (76.8%) had a recognized perinatal sub-specialization on maternal-fetal
medicine, an optional additional qualification in the German-speaking area. The majority
of experts regularly deal with the issue in scientific and/or in training aspects;
94.2% of the experts work in a tertiary maternal and perinatal care center or University
hospitals. On average, experts were 48 years of age, 45% were female ([Table 1 ]).
Table 1
Demographic and baseline characteristics of respondents (mean ± SD or n [%]).
Characteristic
Respondents (n = 69)
Age, years
47.6 ± 8.4
Gender
Female
31 (45%)
Male
38 (55%)
Region of practice
Germany
67 (97%)
Austria
1 (1.5%)
Switzerland
1 (1.5%)
Operating level
Head of department
19 (27.5%)
Head of obstetrics
22 (31.9%)
Consultant
26 (37.7%)
Resident physician
2 (2.9%)
Level of experience
Specialist in obstetrics
53 (76.8%)
Consultant
10 (14.5%)
Resident physician
2 (2.9%)
Level of care
Tertiary obstetric center or University
65 (94.2%)
General/routine obstetric center
3 (4.3%)
Primary care
1 (1.5%)
Deliveries per year
> 3000
14 (20.3%)
2000–3000
27 (39.1%)
1000–2000
27 (39.1%)
< 1000
1 (1.5%)
Delphi rounds
In the first round, a new hepatic hematoma or liver capsule rupture as well as disseminated
intravascular coagulation led to a consensus of 100% and 94.2%, respectively, that
these findings represent an absolute indication of delivery. Eclampsia was considered
as an absolute indication of delivery by 87.0% of the panel, as was suspected fetal
status in cardiotocography, evaluated according to FIGO classification (76.1%). The
laboratory parameters albumin (86.8%), leukocytes (86.6%), uric acid (75.0%) and serum
PlGF (93.7%) as well as the symptom weight gain (73.1%) reached a consensus as not
being an indication for delivery. No consensus was reached in the assessment of the
remaining parameters in the first round. The committee proposed additional parameters
and cut-offs for assessing an indication of delivery in preeclampsia, which were included
in the next rounds. Furthermore, the wish was expressed to assume that all conservative
measures had been exhausted for some
criteria without any improvement to be able to assess the indication of delivery more
easily. In the next rounds, individual laboratory parameters and clinical findings
were clustered to form groups (e.g. acute liver failure, acute kidney failure).
In the second round, consensus was reached regarding new-onset headache with pathologic
cCT/MRI, leading to an immediate indication of delivery. 82.4% of the panel agreed.
Furthermore, 74.0% of the panel were in favor of immediate delivery if there was retrosternal
pain in combination with a pathological finding in ECG.
In the third round, therapy-resistant hypertension of 220/140 mmHg or higher achieved
a consensus for immediate indication of delivery (71.2%) after no agreement could
be reached in the previous rounds at lower blood pressure values. The same applied
to acute liver failure with fulminant coagulation disorder (80.8%), while liver failure
with mild changes in coagulation did not achieve consensus.
In round four and five, the parameters which still failed to reach consensus were
examined again. The classification into the respective categorical query, based on
the majority opinion from the previous surveys, was presented. The panel was asked
whether they agreed or disagreed. For various fetal parameters, it was checked whether
the wording of the current national guideline on fetal growth restriction “delivery
should be considered” was equivalent to a “more serious relative delivery criterion”
in the opinion of the panel.
There was consensus that an oxygen saturation < 90%, an acute severe renal insufficiency,
a moderately severe to severe oliguria (AKI stage 2 and 3), a pulmonary edema, a pleural
effusion, a hyperreflexia, visual complaints, and fetal parameters as low short-time-variability,
increased pulsatility in ductus venosus flow, and reverse end-diastolic flow of A.
umbilicalis are severe delivery criterions. In this round, no consensus was reached
on indication of delivery for pain in the lower and middle abdomen, therapy-resistant
hypertension of 160/110 mmHg, mild oliguria, and absent a-wave in the ductus venosus.
The results of the individual rounds can be reviewed as supplements (Supplemental
Table S1–S5 , online).
All delivery criteria for each organ system, ordered according to the urgency of the
delivery indication from mild (1) to absolute (4) are shown in [Fig. 2 ], [Table 2 ] (ordered by organ systems) and [Table 3 ] (ordered by relevance of delivery). Consensus and assessment of relevance regarding
an indication of delivery presented by domains is shown in Supplemental Table S6 .
Fig. 2
Results of delphi consensus presented by organ systems and delivery criterion. The
delivery criteria for each organ system which reached a consensus are ordered according
to the urgency of the delivery indication from mild (1) to absolute (4).
Table 2
Results of delphi consensus presented by organ systems.
System/organ
Parameter
Consensus
Results of delphi consensus presented by organ systems. Consensus = ≥ 70% of panelists
agree to absolute, severe, moderate or no delivery criterion.
Central nervous system
Headache with pathological cCT/MRI findings
Absolute delivery criterion
Eclampsia
Absolute delivery criterion
Somnolence
Absolute delivery criterion
Hyperreflexia/widened reflex zones
Severe delivery criterion
Dizziness and/or balance problems
Moderate delivery criterion
Headache with non-pathological cCT/MRI findings
Mild delivery criterion
Cardiovascular system
Retrosternal pressure/pain/thoracic pain with pathological electrocardiogram and/or
echocardiography
Absolute delivery criterion
Any pathological cardiac event
Absolute delivery criterion
Therapy-resistant hypertension of 220/140 mmHg
Absolute delivery criterion
Retrosternal pressure/pain/thoracic pain with non-pathological electrocardiography
and/or echocardiography
Moderate delivery criterion
Therapy-resistant hypertension of 160/110 mmHg
No consensus
Respiratory system
Oxygen saturation < 90%
Severe delivery criterion
Pulmonary edema
Severe delivery criterion
Pleural effusion
Severe delivery criterion
Dyspnea
Moderate delivery criterion
Haematological system
Disseminated intravascular coagulation
Absolute delivery criterion
Platelets < 50000 cells/dl or 50 Gpt/l
Absolute delivery criterion
Platelets < 100000 cells/dl or 100 Gpt/l
Moderate delivery criterion
Serum sodium value < 120 mmol/l
Moderate delivery criterion
Leukocytes
No delivery criterion
Serum sodium value < 135 mmol/l
No delivery criterion
Serum potassium
No delivery criterion
Renal and excretion system
Acute severe renal insufficiency (AKI stage 3)
Severe delivery criterion
Moderately severe oliguria (AKI stage 2)
Severe delivery criterion
Severe oliguria to anuria (AKI stage 3)
Severe delivery criterion
Acute mild renal insufficiency (AKI stage 1)
Moderate delivery criterion
Acute moderately severe renal insufficiency (AKI stage 2)
Moderate delivery criterion
Proteinuria
No delivery criterion
Serum uric acid
No delivery criterion
Edema/swelling of the hands/face
No delivery criterion
Weight gain
No delivery criterion
Mild oliguria (AKI stage 1)
No consensus
Metabolic system
Liver hematoma, liver capsule rupture
Absolute delivery criterion
Acute liver failure (transaminase value > 500 U/l + coagulation disorder (detectable
by INR and/or aPTT))
Absolute delivery criterion
Acute liver failure (e.g. very rapid within 24 hours, increase of transaminases and
bilirubin + collapse in coagulation (detectable with INR and/or aPTT))
Absolute delivery criterion
Acute liver dysfunction (transaminase value > 70 and < 500 U/l)
Moderate delivery criterion
Nausea
Moderate delivery criterion
Vomiting
Moderate delivery criterion
Pain in the upper abdomen/epigastrium
Moderate delivery criterion
LDH serum concentration
No delivery criterion
Bilirubin serum concentration
No delivery criterion
Serum albumin
No delivery criterion
Serum haptoglobin
No delivery criterion
Pain in the lower and middle abdomen
No consensus
Sensory system
Papilla congestion/retinal detachment
Absolute delivery criterion
Visual complaints/vision problems
Severe delivery criterion
Hearing problems
Moderate delivery criterion
Uteroplacental and fetal
CTG pathology evaluated according to FIGO classification
Absolute delivery criterion
Vaginal bleeding with abnormal sonography
Absolute delivery criterion
Short-time variability < 2.6 ms (26+0–28+6 weeks of gestation) or < 3.0 ms (29+0–31+6
weeks of gestation)
Severe delivery criterion
Ductus venosus: increased pulsatility (PIV > 95th percentile)
Severe delivery criterion
Ductus venosus: absent a-wave or a reverse flow of the a-wave
Severe delivery criterion
A. umbilicalis: reverse end-diastolic flow
Severe delivery criterion
Vaginal bleeding with normal sonography
Moderate delivery criterion
Serum sFlt-1/PIGF ratio
No delivery criterion
Serum PlGF
No delivery criterion
Table 3
Results of delphi consensus presented by relevance of delivery criterion.
Consensus
Parameter
Results of delphi consensus presented by relevance of delivery criterion. Consensus
= ≥ 70% of panelists agree to absolute, severe, moderate or no delivery criterion.
Absolute delivery criterion
Headache with pathological cCT/MRI findings
Eclampsia
Somnolence
Retrosternal pressure/pain/thoracic pain with pathological electrocardiogram and/or
echocardiography
Any pathological cardiac event
Therapy-resistant hypertension of 220/140 mmHg
Disseminated intravascular coagulation
Platelets < 50000 cells/dl or 50 Gpt/l
Liver hematoma, liver capsule rupture
Acute liver failure (transaminase value > 500 U/l + coagulation disorder (detectable
by INR and/or aPTT))
Acute liver failure (e.g. very rapid within 24 hours, increase of transaminases and
bilirubin + collapse in coagulation (detectable with INR and/or aPTT))
Papilla congestion/retinal detachment
CTG pathology evaluated according to FIGO classification
Vaginal bleeding with abnormal sonography
Severe delivery criterion
Hyperreflexia/widened reflex zones
Oxygen saturation < 90%
Pulmonary edema
Pleural effusion
Acute severe renal insufficiency (AKI stage 3)
Moderately severe oliguria (AKI stage 2)
Severe oliguria to anuria (AKI stage 3)
Visual complaints/vision problems
Short-time variability < 2.6 ms (26+0–28+6 weeks of gestation) or < 3.0 ms (29+0–31+6
weeks of gestation)
Ductus venosus: increased pulsatility (PIV > 95th percentile)
Ductus venosus: absent a-wave or a reverse flow of the a-wave
A. umbilicalis: reverse end-diastolic flow
Moderate delivery criterion
Dizziness and/or balance problems
Retrosternal pressure/pain/thoracic pain with non-pathological electrocardiogram and/or
echocardiography
Dyspnea
Platelets < 100000 cells/dl or 100 Gpt/l
Serum sodium value < 120 mmol/l
Acute mild renal insufficiency (AKI stage 1)
Acute moderately severe renal insufficiency (AKI stage 2)
Acute liver dysfunction (transaminase value > 70 and < 500 U/l)
Nausea
Vomiting
Pain in the upper abdomen/epigastrium
Hearing problems
Vaginal bleeding with normal sonography
Mild delivery criterion
Headache with non-pathological cCT/MRI findings
No delivery criterion
Leukocytes
Serum sodium value < 135 mmol/l
Serum potassium
Proteinuria
Serum uric acid
Edema/swelling of the hands/face
Weight gain
LDH serum concentration
Bilirubin serum concentration
Serum albumin
Serum haptoglobin
Serum sFlt-1/PIGF ratio
Serum PlGF
No consensus
Mild oliguria (AKI stage 1)
Therapy-resistant hypertension of 160/110 mmHg
Pain in the lower and middle abdomen
Discussion
The term “severe preeclampsia” is not universally defined, nor is the indication for
a mandatory delivery in preeclampsia. Particularly in the case of very early pre-eclampsia
before 32 weeks’ gestation, the decision to deliver is often made by an interdisciplinary
team based on individual factors and circumstances, including fetal maturity, maternal
symptoms and clinical picture, but also logistical aspects and personal opinions.
The term “severe preeclampsia” is then often used as a justification for delivery
without a more precise definition. The aim of the present study is to reach a consensus
among German-speaking obstetricians experienced in the management of women with preeclampsia
before 32 weeks’ gestation to define “severity” by the need for delivery for acute
maternal reasons with impending complications. To this end, a Delphi procedure was
conducted in which participants were asked for their opinion on when delivery should
take place in a particular case with
different constellations of pre-eclampsia. Fifty experts finished all five Delphi
rounds. Consensus was reached on 53 parameters. Of these 14 were considered absolute
criteria, i.e. rapid preparation for delivery without any delay, 26 parameters were
rated as factors that should be considered in the decision without being absolute,
and 13 parameters should have no influence on the decision to deliver. On three parameters
a consensus was not reached. Most interesting, “blood pressure values” was one of
these. Depending on the height of blood pressure, obstetricians either agreed that
this only mildly influences decision to deliver, or it was difficult to get any consent.
In contrast to the expected threshold of 160/110 mmHg, which generally defines severe
hypertension in national guidelines, consent was only reached at therapy-resistant
blood pressure values that permanently exceed 220/140 mmHg.
Most international and national guidelines provide indications for delivery, independent
of gestational age. Uncontrollable or worsening hypertension and eclampsia were most
advocated [2 ]
[19 ]
[32 ]. A specific upper limit of blood pressure at which delivery should take place is
not specified in the current guidelines, which apply to most of the experts involved
[2 ]. However, there is a general agreement that severe hypertension at or above 160/110 mmHg
is associated with increased risk for the mother to suffer from stroke and other acute
cardiovascular or neurologic complications [2 ]
[33 ]
Antihypertensive medication should therefore be considered at the latest if blood
pressure consistently exceeds this threshold [2 ]
[34 ]. It has been suggested if severe hypertension persists despite adequate antihypertensive
therapy, delivery should be attempted after maternal stabilization [35 ]. It was therefore rather surprising that no consensus could be reached in favor
of a delivery in this situation. Since delivery is the only causative cure of preeclampsia,
the decision to end pregnancy aims at reducing short-term morbidity of the mother
including stroke and eclampsia. However, one may also consider delivery at a certain
point of persistent high blood pressure values since preeclampsia is associated with
long-term risks on maternal health [8 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ]. Whether these long-term consequences may be influenced by timely delivery (i.e.
by shortening the exposure to a preeclamptic condition) is a matter of debate.
In contrast to blood pressure, other parameters were clearly regarded by the panel
as absolute indications for delivery. Most parameters belong to the group of HELLP
syndrome, which is excluded in most published studies on expectant management of severe
preeclampsia, as these abnormalities are generally considered an indication for delivery
[34 ]
[36 ]
[37 ]. A large systematic review by Magee et al. examined the frequency of expectant management
complications in HELLP syndrome before 34 weeks of gestation [38 ]. Complications included severe hypertension, abruptio placentae and eclampsia, subcapsular
hepatic hematoma, stroke, stillbirth, and neonatal death. The panel’s assessment of
these
parameters as an absolute indication for delivery is therefore plausible. Likewise,
acute moderate to severe renal insufficiency (AKI stage 2 and 3) is regarded as absolute
indication for delivery. This is compatible with the available studies, which compared
conservative treatment to immediate delivery [34 ]
[36 ]
[37 ]
[39 ]. A high risk of increasing severe renal insufficiency was found for expectant vs.
interventionist care (RR 3.33) [38 ]. In addition, prolonged exposure to placental ischemia impairs postpartum maternal
renal function [40 ]
[41 ]. Beside maternal complications, non-assurance of fetal well-being by Doppler ultrasound
or cardiotocogram is generally accepted for timing of birth. Several randomized multicenter
studies in recent years have addressed the management and monitoring of early severe
fetal growth restriction, which is often a concomitant condition of preeclampsia.
For pregnancies before 32 weeks of gestation, the TRUFFLE study showed that a combination
of monitoring using computerized CTG and Doppler of the ductus venosus is most suitable
for estimating the correct time of delivery [42 ]
[43 ]. However, in a third of the women included in the TRUFFLE study, maternal factors
such as severe preeclampsia, rather than the results of fetal monitoring, led to the
decision to deliver. This
emphasizes the need for clear maternal delivery indications regardless of the fetal
situation. However, a fetal assessment should be carried out and fetal well-being
should be considered during management [39 ]
[44 ]
[45 ]
[46 ].
A limitation of the study is that the invited senior obstetricians primarily derive
from German-speaking countries (Germany, Austria, Switzerland). The answers given
must therefore be interpreted in the context of the respective national healthcare
system and cannot necessarily be generalized. Moreover, though skilled for the management
of high-risk pregnancies in a clinical setting the participants overall may not necessarily
have a scientific focus in the treatment of preeclampsia. Therefore, the consent achieved
may rather reflect clinical practice than expert opinion. Additionally, the Delphi
output quality reflects the contemporary interpretation of the actual state of the
art, which may rapidly change over time. Another potential bias is the Delphi consensus
design in which results have been presented to the participants in subsequent rounds.
Participants might have re-considered their initial views and the views of the majority
are agreed upon [47 ]. A further limitation is that the questionnaire asked about single events or parameters
that occurred in connection with a specific case. In the clinical context, however,
the decision is usually based on a summation of several factors. For example, severe
hypertension, even if resistant to treatment, is often only considered an indication
for delivery if other symptoms such as headaches, nausea, or abdominal tenderness
occur. It therefore remains to be investigated whether a combination of characteristics
or a summation of the mild to moderate indications for delivery identified here can
change the decision.
Another weakness of this Delphi procedure is that we did not incorporate gestational
age as an independent factor. For simplification of the questionnaire, and hence for
feasibility of the Delphi procedure we rather assumed that between 24 to 32 weeks’
gestation decision will not be influenced by dynamics in neonatal outcome. However,
gestational age at preterm birth is inversely correlated to adverse medical and neurodevelopmental
outcomes, including motor, neurosensory, cognitive, and behavioral deficits [48 ]
[49 ]
[50 ]. Therefore, the criteria for delivery are relative since at more than 30 weeks most
clinicians would be more liberal to deliver than at the limit of viability. Likewise,
we specifically asked for criteria for delivery after completion of antenatal
steroid therapy. The reason for this is that when designing the study, we primarily
aimed at defining endpoints for the treatment of women with preeclampsia in multicenter,
randomized, controlled intervention trials. In these, completed antenatal steroid
therapy was considered a compulsory inclusion criterion [51 ]
[52 ]
[53 ]
[54 ]
[55 ]. For general clinical practice it would also be of interest how the experts would
have decided without having completed antenatal steroid therapy. This issue should
be investigated in future observational studies that can now be based on the here
presented findings. The restriction to a clear case of
preeclampsia can also be regarded as an advantage of the study. Preeclampsia is characterized
by many variants in additional risk factors and multivariate expression. It was therefore
important to create a uniform picture of pre-eclampsia that could be presented to
the participating experts. Another strength of the study was the recruitment of a
large and diverse panel of clinicians in the field. Once having completed the first
round of the Delphi procedure, dropout rate was low. As the clinicians were obstetricians
from different perinatal centers all over Germany (and Austria and Switzerland) with
different specialties, local infrastructure, and resources, the data presented here
can be considered a good average of current clinical practice. We are aware that in
other countries the indication for delivery may be made by an interdisciplinary team,
usually obstetrics, neonatology, anesthesia, internal medicine, and others. In Germany,
however, the final decision to recommend
delivery to women and to balance the maternal risks against the fetal benefits of
prolongation lies with the obstetrician (subspecialized in maternal and fetal medicine).
Therefore, only obstetricians have been invited to participate. Finally, an already
established Delphi method was used in this study [56 ]
[57 ]
[58 ]. To maximize the possibilities for consensus, a modification was used to reformulate
questions or define parameters more precisely after rounds had ended. By doing this,
we achieved consensus in 53 of 56 questions and scenarios.
In conclusion, whether to prolong pregnancy in preeclampsia and when to deliver is
not uniformly defined and varies from clinic to clinic and case to case. This makes
it difficult for clinical interventional studies to use the time from admission to
delivery, or the duration of prolongation, respectively, as primary endpoints [52 ]
[54 ]. Since the non-standardized assessment of disease severity and the different indications
for delivery between study sides, involved staff, and other individual patient’s factors
call into question the prolongation of pregnancy until delivery as a primary endpoint,
consensus-based diagnostic criteria may be helpful by integrating into prospective
study designs. By generating a consensus within a panel of clinical experts working
in the field, we stimulated a discussion on how to define “severe
preeclampsia” by the need of delivery for acute or imminent maternal complications
of preeclampsia in Germany. Although our study aimed to develop objective degrees
of severity indicating prompt delivery in pregnancies with early-onset preeclampsia,
the partly high cut-off values for consensus were rather surprising. This might be
because the disease is characterized by many variants in additional risk factors and
multivariate expression. In addition, the population served by the experts might show
such variation in severity that the experience of serious outcomes is rather small
also because long-term follow-up is not realized. As with neonatal outcomes, quality
of life and health of mothers throughout their life cycle are equally important and
must be considered when timing delivery.
Supplementary Material
Supplemental Table S1 . Results of round 1.
Supplemental Table S2 . Results of round 2.
Supplemental Table S3 . Results of round 3.
Supplemental Table S4 . Results of round 4.
Supplemental Table S5 . Results of round 5.
Supplemental Table S6 . Consensus and assessment of relevance regarding an indication of delivery presented
by domains.
