Er zijn onder andere in Nederland verschillende studies gedaan waarin is gekeken naar het effect van inleiden vergeleken met afwachtend beleid bij zwangere vrouwen met een hypertensieve aandoening, maar deze studies zijn nog niet opgenomen in een landelijke richtlijn. Daardoor bestaat er waarschijnlijk praktijkvariatie en mogelijk substandard care.

Maternal outcome measures

Perinatal outcomes

Description of studies

Bernardes (2019) performed an individual participant data meta-analysis to compare immediate delivery with expectant management for prevention of adverse maternal and neonatal outcomes in women with hypertensive disease in pregnancy. CENTRAL, PubMed, MEDLINE and Clinical Trials.gov were searched till 31 December 2017 for RCT’s comparing immediate delivery to expectant management in women presenting with gestational hypertension or pre-eclampsia without severe features from 34 weeks of pregnancy onwards. The primary neonatal outcome was respiratory distress syndrome (RDS) and the primary maternal outcome was a composite of HELLP syndrome (abbreviation for pregnancy complication, characterized by Hemolysis, Elevated Liver enzymes, and a Low Platelet count) and eclampsia. Secondary outcomes were among others pulmonary oedema, renal failure, placental abruption/antenatal haemorrhage, severe postpartum haemorrhage (> 1000 mL), intracerebral haemorrhage and cerebral infarction. Only pooled results were shown for the secondary outcomes. Five RCT’s with a total of 1778 participants were included (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014; van Bulck, 2003). To avoid selection bias, Bernardes (2019) did not use data from the GRIT study (van Bulck, 2003) to calculate preterm delivery rates and median time to delivery because of foetal compromise. Table 1 presents a brief overview of the study characteristics of the included RCTs in Bernardes (2019).

Chappell (2019) performed a parallel-group, non-masked, multicentre, randomised controlled trial to compare initiation of delivery (within 48 hours of randomization) with expectant management (delivery at 37 weeks of pregnancy, or sooner as clinical needs dictated). Pregnant women between 34 to 37 weeks of pregnancy with a diagnosis of pre-eclampsia or superimposed pre-eclampsia with a singleton or dichorionic diamniotic twin pregnancy were included. Women were excluded if the decision to deliver within the next 48 hours had already been made (immediate delivery was indicated according to national guidelines in women with persistent severe features of pre-eclampsia). In total 899 women were eligible for randomization, 448 women (471 infants (5% twins)) were allocated to initiation of delivery and 451 women (475 infants (5% twins)) were allocated to the expectant management group. Groups were comparable at baseline, except that women in the planned delivery group gave birth at 252 days of pregnancy versus 257 days of pregnancy in expectant management, and women in the expectant group were more likely to achieve a spontaneous vaginal delivery (n=2 (<1%) versus n=19 (4%)). Chappell (2019) conducted an intention-to-treat analysis and per protocol analysis (excluding women lost to follow-up (n=1 versus n=2), women who received planned delivery > 48 hours after randomization (n=120 (27%)) in the intervention group, or women who received non-indicated delivery < 37 weeks of pregnancy in the expectant group (n=2 (0.4%)). Results from intention-to-treat analysis and per protocol analysis were similar. Chappell (2019) examined composite score of maternal morbidity and a composite score of perinatal morbidity as co-primary outcome measures.

Table 1 Overview of study characteristics of included RCTs in Bernardes (2019)

Results

It should be noted at forehand that none of the included studies reported the effect of immediate delivery compared to expectant management on perinatal outcomes in women suffering from hypertensive disorders of pregnancy after 32 weeks of pregnancy. Every included study reported on women suffering from hypertensive disorders of pregnancy after 34 weeks of pregnancy. Meta-analysis with crude data were performed where possible. Subgroup analyses exploring differences between studies including patients with or without suspected foetal growth restriction were performed where possible.

Maternal outcomes

1. Pre-eclampsia

All women in the study by Chappell (2019) had pre-eclampsia at inclusion. Five trials included in the review by Bernardes (2019) reported the number of pregnant women with pre-eclampsia, which was defined as hypertension (blood pressure (BP) levels ≥ 140 mmHg systolic or ≥ 90 mmHg diastolic) and proteinuria (300 mg or more total protein in a 24-h urine sample, recurrent positive protein dipstick test or protein/creatinine ratio of 30 mg/mmol or more) (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014; van Bulck, 2003). Bernardes (2019) only reported the pooled results.

There was no difference in the number of women with pre-eclampsia between the immediate delivery group (392/861) and the expectant management group (378/863) (RR 1.04 (95%CI 0.94 to 1.15)).

2. Eclampsia and/or HELLP syndrome

Two studies examined the outcome measure eclampsia and/or HELLP syndrome (Bernardes, 2019; Chappell, 2019). Five trials included in the review by Bernardes (2019) reported the outcome measure eclampsia (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014; Van Bulck, 2003), which was not defined in text. In addition, Bernardes (2019) reported a composite outcome of HELLP syndrome and/or eclampsia (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014; Van Bulk, 2003). Chappell (2019) also reported HELLP syndrome, but it is unclear if these data overlapped with women with eclampsia.

There was no difference in the risk of eclampsia between immediate delivery (3/1309) compared to expectant management (7/1314) (pooled RR 0.53 (95%CI 0.16 to 1.75); Figure 1). Subgroup analysis stratified for study populations with or without suspected foetal growth restriction showed similar effects.

When a composite outcome of HELLP syndrome and/or eclampsia was studied, immediate delivery reduced the risk of HELLP syndrome or eclampsia with 53% (pooled RR 0.47 (95% CI 0.27 to 0.82); Figure 2). The calculated number needed to treat (NNT) is 63. When this analysis was repeated without the number of HELLP cases reported by Chappell (2019), to prevent possible duplication of patients with both HELLP and eclampsia diagnosis, the overall effect was the same (pooled RR 0.39 (95%CI 0.19 to 0.78)). Subgroup analysis stratified for study populations with or without suspected foetal growth restriction showed similar effects.

Bernardes (2019) analysed the composite outcome HELLP syndrome and/or eclampsia stratified for weeks of gestation. Chappell (2019) did not report this subgroup analysis and was therefore not included. There was no difference in the outcome HELLP syndrome and/or eclampsia stratified for weeks of gestation (Table 2). When stratified based on type of hypertensive disorder of pregnancy, Bernardes (2019) reported an increased risk of HELLP syndrome and/or eclampsia in women with pre-eclampsia (RR 0.39 (95%CI 0.15 to 0.98)). For women with gestational hypertension or chronic hypertension it is unclear whether there is an increased risk for HELLP syndrome and/or eclampsia, because the reported 95% confidence intervals contain the value of no (clinically relevant) effect (Table 3).

Figure 1 Eclampsia comparison immediate delivery versus expectant management

Figure 2 HELLP syndrome and/or eclampsia comparison immediate delivery versus expectant management

Table 2 HELLP syndrome and/or eclampsia comparison immediate delivery versus expectant management, subgroup analysis by Bernardes (2019) by weeks of gestations (2019)

Table 3 HELLP syndrome and/or eclampsia comparison immediate delivery versus expectant management, subgroup analysis by Bernardes (2019) by type of hypertensive disorder (2019)

3. Pulmonary oedema

Two studies reported the outcome measure pulmonary oedema (Bernardes, 2019; Chappell, 2019. The outcome was not defined in text. Three trials included in the review by Bernardes (2019) reported the outcome (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009), but only the pooled results were reported by the review.

There was no difference in the outcome pulmonary oedema between the immediate delivery group (1/1201) and the expectant management group (4/1227) (RR 0.36 (95% CI 0.05 to 2.34) Figure 3).

Figure 3 Pulmonary oedema comparison immediate delivery versus expectant management

4. Hepatic haemorrhage

No study reported the outcome hepatic haemorrhage specifically. One study reported the outcome hepatic haematoma or rupture (Chappell, 2019). Four trials included in the review by Bernardes (2019) reported the outcome liver failure, which was not defined in text (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014).

No patients in either the immediate delivery group (0/1309) or expectant management group (0/1314) developed liver failure, hepatic haematoma or hepatic rupture, whereby the RR could not be calculated for this outcome.

5. Renal insufficiency

Chappell (2019) reported the outcome renal insufficiency (defined as creatine > 150 µmol/L; no pre-existing renal disease). Four trials included in the review Bernardes (2019) did not report the outcome renal insufficiency, but did report the outcome renal failure, which was not defined in text (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014). Chappell (2019) also reported renal failure, which was defined as creatinine >200 µmol/L; pre-existing renal disease.

No women in either the immediate delivery group (0/448) or expectant management group (0/451) developed renal insufficiency, whereby the RR could not be calculated for this outcome.

No women in either the immediate delivery group (0/1309) or expectant management group (0/1314) developed renal failure, whereby the RR could not be calculated for this outcome.

6. Cerebral haemorrhage

Five trials included in the review by Bernardes (2019) reported the outcome intracerebral haemorrhage, which was not defined in text (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014; van Bulck, 2003).

No women in either the immediate delivery group (0/886) or expectant management group (0/892) developed an intracerebral haemorrhage, whereby the RR could not be calculated for this outcome.

7. Placental abruption

Two studies reported the outcome placental abruption (Bernardes, 2019; Chappell, 2019). Four studies reported the outcome placental abruption. Three trials included in the review by Bernardes (2019) reported the outcome (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009), but only the pooled results were reported by the review.

There was no difference in the outcome placental abruption between the immediate delivery group (4/1204) and the expectant management group (6/1227) (RR 0.77; 95% CI 0.22 to 2.69; Figure 4).

Figure 4 Placental abruption comparison immediate delivery versus expectant management

8. Caesarean section

Two studies reported the outcome caesarean section (Bernardes, 2019; Chappell, 2019). Five trials included in the review by Bernardes (2019) reported the outcome (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014, Van Bulck, 2003), but only the pooled results were reported by the review.

Caesarean sections were reported in 493/1357 women undergoing immediate delivery compared to 534/1367 women treated with expectant management (pooled RR 0.92 (95%CI 0.85 to 1.01) figure 5). This was not a clinically relevant difference, as the 95%CI crosses the line of no clinically relevant effect (and the line of no statistically significant effect).

Figure 5 Caesarean section comparison immediate delivery versus expectant management.

9. Postpartum haemorrhage

One study reported the outcome postpartum haemorrhage (Bernardes, 2019). Four trials included by the review of Bernardes (2019) reported the outcome severe postpartum haemorrhage (defined as > 1000 ml blood loss) (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014), but only the pooled results were reported by Bernardes (2019).

Postpartum haemorrhage was reported in 69/861 women undergoing immediate delivery compared to 90/863 women in expectant management (pooled RR 0.77 (95%CI 0.57 to 1.04)). It is uncertain whether this is a clinically relevant difference, as the 95%CI interval includes the limit of no clinically relevant effect (and no statistically significant effect).

10. Patient satisfaction

The outcome measure patient satisfaction was not reported in the included studies.

11. PTSD (post-traumatic stress disorder)

The outcome measure PTSD was not reported in the included studies.

12. Bonding (mother-child)

The outcome measure bonding (mother-child) was not reported in the included studies.

Perinatal outcomes

13. Perinatal death (intra-uterine death and neonatal death)

The outcome measure perinatal death was reported by three definitions: perinatal death, intra-uterine death and neonatal death, reported respectively in paragraphs 13.1, 13.2, 13.3. Most studies did not define the outcomes specifically.

13.1. Perinatal death

Five studies in the review by Bernardes (2019) reported the outcome perinatal mortality (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014, Van Bulck, 2003). Chappell (2019) reported the number of stillbirths and neonatal deaths. Both studies did not define the outcomes in text.

There was no difference in the outcome perinatal death between the immediate delivery group (1/1357) and the expectant management group (1/1367) (pooled RR 1.01 (95% CI 0.06 to 16.07); Figure 6).

Figure 6 Perinatal mortality comparison immediate delivery versus expectant management

13.2. Intra-uterine death

One study reported the outcome intra-uterine death (Chappell, 2019), which was reported as stillbirth, not further defined in text.

There were no stillbirths in either the immediate delivery (0/471 infants) compared to expectant management groups (0/475 infants), hence no RR could be calculated.

13.3. Neonatal death

One study reported the outcome neonatal death (Chappell, 2019), which was not defined in text.

There were no reports of neonatal death in either the immediate delivery group (0/471) or expectant management groups (0/475), whereby the RR could not be calculated for this outcome.

14. Infant respiratory distress syndrome

Five studies in the review by Bernardes (2019) reported the outcome infant respiratory distress syndrome (IRDS) (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014, Van Bulck, 2003) (Figure 7). Bernardes (2019) also reported the pooled RR stratified for gestational age (Table 4).

There was an increased risk of IRDS for neonates in the immediate delivery group (29/886) compared to the expectant management group (14/892) (pooled RR 1.94 (95% CI 1.05 to 3.59); Figure 7).

Figure 7 IRDS comparison immediate delivery versus expectant management.

Table 4 IRDS comparison immediate delivery versus expectant management, subgroup analysis by Bernardes (2019) by weeks of gestations (2019)

15 NICU admission

Two studies reported the outcome NICU admissions (Bernardes, 2019; Chappell, 2019). Four studies in the review by Bernardes (2019) reported the outcome (Boers, 2010; Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014), but only the pooled results were reported by Bernardes (2019). Chappell (2019) reported the outcome admission to neonatal unit, which was further specified in categories of care where mother and baby were separated: time in intensive care, time in high dependency care and time in special care. These outcomes were not defined in text by Chappell (2019). For the analysis of the outcome measure NICU admission, the time in intensive care by Chappell (2019) was included in the meta-analysis.

NICU admission was reported in 80/1279 infants in the immediate delivery group compared to 59/1273 infants in the expectant management group (pooled RR 1.35 (95% CI 0.97 to 1.87); Figure 8). It is uncertain whether this is a clinically relevant difference, as the 95%CI interval includes the limit of no clinically relevant effect (and no statistically significant effect).

Figure 8 NICU admission comparison immediate delivery versus expectant management

16. Low birth weight for gestational age

Two studies reported the outcome low birth weight for gestational age (Bernades, 2019; Chappell, 2019). Three trials included in the review by Bernardes (2019) reported the outcome measure as small for gestational age (defined as birth weight < 10^th percentile) (Broekhuijsen, 2015; Koopmans, 2009; Owens, 2014), but only the pooled results were reported by Bernardes (2019). Chappell (2019) reported the outcome as birthweight < 10^th centile.

Low birth weight for gestational age (< 10^th percentile) was reported in 196/1286 women with immediate delivery (15.2%) compared to 241/1273 women in expectant management (18.9%) (pooled RR 0.81 (95% CI 0.68 to 0.96)) (Figure 9). It is uncertain whether this is a clinically relevant difference, as the 95%CI interval includes the limit of no clinically relevant effect.

Figure 9 Low birth weight for gestational age (<10^th percentile) comparison immediate delivery versus expectant management

17. Preterm birth < 37weeks

One study reported the outcome preterm birth < 37 weeks (Chappell, 2019).

There was an increased risk of preterm birth < 37 weeks of pregnancy in women treated with immediate delivery (387/471) compared to expectant management (261/475) (RR 1.50 (95% 1.36 to 1.64) (figure 10).

Figure 10 Preterm birth < 37 weeks comparison immediate delivery versus expectant management

18. Successful breastfeeding at 6 weeks

The outcome measure successful breastfeeding at 6 weeks was not reported in the included studies. Chappell (2019) reported the outcome method of infant feeding 24 hours prior to discharge, which is reported in Table 5 below.

Exclusive breastfeeding at 24 hours prior to discharge from hospital was reported in 112/471 women in the immediate delivery group compared to 139/475 women in the expectant management group (RR 0.81 (95%CI 0.66 to 1.01)). It is uncertain whether this is a clinically relevant difference, as the 95%CI interval includes the limit of no clinically relevant effect (and no statistically significant effect).

Table 5 Method of feeding 24 hours prior to discharge comparison immediate delivery versus expectant management, extracted from Chappell (2019)

Level of evidence of the literature

RCTs start at a GRADE high.

Maternal outcome measures

The level of evidence regarding the outcome measure pre-eclampsia was downgraded by one to a ‘moderate’ level because of risk of bias (impossible to blind participants in Bernardes (2019)).

The level of evidence regarding the outcome measure eclampsia was downgraded by three levels to ‘very low’ because of imprecision (low number of events, wide 95%CI) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure HELLP syndrome and/or eclampsia was downgraded by one level to ‘moderate’ because of risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure pulmonary oedema was downgraded by three levels to ‘very low’ because of imprecision (low number of events, wide 95%CI) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure hepatic haemorrhage defined as either liver failure, hepatic heaematoma or hepatic rupture was downgraded by three levels to ‘very low’ because of imprecision (no events) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measures renal insufficiency and renal failure were both downgraded by three levels to ‘very low’ because of imprecision (no events) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure cerebral haemorrhage was downgraded by three levels to ‘very low’ because of imprecision (no events) and risk of bias (impossible to blind participants in Bernardes (2019)).

The level of evidence regarding the outcome measure placental abruption was downgraded by three levels to ‘low’ because of imprecision (low number of events, wide 95%CI) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure caesarean section was downgraded by one level to ‘moderate’ because of risk of bias (impossible to blind participants in in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure postpartum haemorrhage was downgraded by three levels to ‘very low’ because of imprecision (95%CI exceeds the limit of no clinically relevant effect) and risk of bias (impossible to blind participants in Bernardes (2019)).

The level of evidence regarding the outcome measures patient satisfaction, PTSD and bonding (mother-child) could not be assessed because of lack of data (none of the included studies reported on outcomes).

Perinatal outcome measures

The level of evidence regarding the outcome measure perinatal death was downgraded by three levels to ‘very low’ because of imprecision (low number of events, wide 95%CI) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)). The level of evidence regarding the outcome measures intra-uterine death (stillbirths) and neonatal death were downgraded by three levels to ‘very low’ because of imprecision (no events) and risk of bias (impossible to blind participants in Chappell (2019)).

The level of evidence regarding the outcome measure infant respiratory distress syndrome was downgraded by one level to ‘moderate’ because of risk of bias (impossible to blind participants in Bernardes (2019)).

The level of evidence regarding the outcome measure NICU admission was downgraded by two levels to ‘low’ because of imprecision (95%CI includes the limit of no clinically relevant effect) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure low birth weight for gestational age was downgraded by two levels to ‘low’ because of imprecision (95%CI includes the limit of no clinically relevant effect) and risk of bias (impossible to blind participants in Bernardes (2019) and Chappell (2019)).

The level of evidence regarding the outcome measure preterm birth <37 weeks was downgraded by two level to ‘moderate’ because of risk of bias (impossible to blind participants in Chappell (2019)).

The level of evidence regarding the outcome measure successful breastfeeding at 6 weeks could not be assessed because of lack of data (none of the included studies reported on outcomes). Proxy outcome measure exclusive breastfeeding at 24 hours prior to discharge from hospital was downgraded by two levels to ‘low’ because of imprecision (95%CI includes the limit of no clinically relevant effect) and risk of bias (impossible to blind participants in Chappell (2019)).

A systematic review of the literature was performed to answer the following question:

Is induction of labour improving maternal and perinatal outcomes compared to expectant management in pregnant women with hypertensive disorders in pregnancy > GA 32+0?

P: pregnant women with hypertensive disorders in pregnancy > GA 32+0;

I: induction of labour;

C: expectant management;

O: at least one of the following outcome measures:

Maternal outcomes:

• Pre-eclampsia
• Eclampsia and/or HELLP syndrome
• Pulmonary oedema
• Hepatic haemorrhage
• Renal insufficiency
• Cerebral haemorrhage
• Placental abruption
• Caesarean section
• Postpartum haemorrhage
• Patient satisfaction
• PTSD (post-traumatic stress disorder)
• Bonding (mother-child)

Perinatal outcomes:

• Perinatal death (intra-uterine death and neonatal death)
• Infant respiratory distress syndrome (IRDS)
• NICU admission
• Low birth weight for gestational age
• Preterm birth < 37weeks
  • Successful breastfeeding at 6 weeks

Relevant outcome measures

The guideline development group considered eclampsia as a critical outcome measure for decision making; and pre-eclampsia, pulmonary oedema, hepatic haemorrhage, renal insufficiency, cerebral haemorrhage, placental abruption, caesarean section, postpartum haemorrhage, patient satisfaction, PTSD, bonding (mother-child), perinatal death (intra-uterine death and neonatal death), IRDS, NICU admission, low birth weight for gestational age, preterm birth < 37weeks and successful breastfeeding at 6 weeks as important outcome measures for decision making.

A priori, the working group did not define the outcome measures listed above but used the definitions used in the studies.

For the outcome measures eclampsia, pulmonary oedema, hepatic haemorrhage, renal insufficiency, cerebral haemorrhage, placental abruption and perinatal death (intra-uterine death and neonatal death), any statistically significant difference was considered as a clinically important difference between groups. For the outcome caesarean section, the working group defined a difference of 5% in the relative risk as a minimal clinically important difference. For all other outcome measures, the GRADE default - a difference of 25% in the relative risk for dichotomous outcomes (Schünemann, 2013) and 0.5 standard deviation for continuous outcomes - was taken as a minimal clinically important difference.

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms until October 9th, 2019. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 322 hits. Studies were selected based on the following criteria:

• the study population had to meet the criteria as defined in the PICO;
• intervention as defined in the PICO;
• original research or systematic review.

Fifty-five studies were initially selected based on title and abstract screening. After reading the full text, 53 studies were excluded (see the table with reasons for exclusion under the tab Methods) and two studies were included.

Results

One systematic review and one randomized controlled trial (RCT) were included in the analysis of the literature. Important study characteristics and results are summarized in the evidence tables. The assessment of the risk of bias is summarized in the risk of bias tables.