Bij vrouwen met diabetes type 1 en 2 zonder complicaties wordt vaak in de a terme periode een inleiding van de baring aangeboden om de kans op onverwachte intra-uteriene vruchtdood te verkleinen. In hoeverre verhoogt echter een vroeg a terme inleiding (37 tot 39 wk) van de baring de kans op een mislukte inleiding of op neonatale problemen? Bij een afwachtend beleid zou juist de kans op macrosomie, schouderdystocie of keizersnede verhoogd zijn. Mogelijk geldt ditzelfde ook voor vrouwen met GDM die insuline of andere glucoseregulerende medicatie gebruiken. Deze richtlijnmodule beoogt om tot een afgewogen advies te komen over het inleiden van de baring bij vrouwen met pre-existente diabetes mellitus type 1 en 2 en GDM die behandeld worden met insuline of andere glucoseverlagende medicatie.

6.1.2 Timing of birth

6.1.2.1 Review question: What is the gestational age-specific risk of intrauterine death in pregnancies with type 1, type 2 or gestational diabetes, and the optimal timing of birth?

6.1.2.3 Description of included studies

Six studies were identified for inclusion in this review (Rosenstein, 2012; Holman, 2014; Eidem, 2011; Kjos, 1993; Lurie, 1996; Alberico, 2010). Four studies included women with gestational diabetes (Rosenstein, 2012; Kjos, 1993; Lurie, 1996; Alberico, 2010), one study included women with type 1 diabetes (Eidem, 2011) and the sixth study included women with Type 1 or Type 2 diabetes (Holman, 2014).

One large retrospective study from the USA (Rosenstein, 2012) used record linkage to examine mortality rates in babies of women with gestational diabetes compared with those with no diabetes. Three mortality outcomes were examined in the study and were stratified by gestational age: stillbirth, neonatal death and infant death. A second study (Holman, 2014) examined data on stillbirth incidence collected from four audits of women with pre-existing diabetes conducted in England over two time periods and compared these data to stillbirth incidence in all births in England and Wales collected by the Office of National Statistics.

A third study from Norway (Eidem, 2011) also used record linkage to examine perinatal mortality rates in babies of women with pregestational type 1 diabetes compared with those without type 1 diabetes. There were no corresponding studies that could be included that examined mortality rates with different severities of diabetes (such as those estimated by HbA1c).

Three further studies compared morbidity and mortality outcomes in women with gestational diabetes who underwent elective delivery or expectant management (Kjos, 1993; Lurie, 1996; Alberico, 2010). One was a randomised controlled trial (RCT) from the USA (Kjos, 1993), one a prospective cohort study with historical controls from Israel (Lurie, 1996) and one was a retrospective cohort from Italy (Alberico, 2010).

Most women (93.5%) in the RCT (Kjos, 1993) and all women in the prospective cohort study (Lurie, 1996) had insulin treated gestational diabetes (A2). The retrospective cohort study (Alberico, 2010) included women with insulin treated and non-insulin treated gestational diabetes (A1 and A2) and foetal growth acceleration diagnosed at 38 weeks of gestation.

The elective delivery method was active induction of labour in the 38th or 39th gestational week in three studies (Kjos, 1993; Lurie, 1996; Alberico, 2010). Two studies required evidence of fetal lung maturity before induction was started (Kjos, 1993; Lurie, 1996). Up to three applications of vaginal prostaglandin were used for cervical ripening in women with low Bishop scores before treatment with intravenous oxytocin to induce labour in the RCT (Kjos, 1993). In the prospective cohort study, induction of labour was performed by either intracervical balloon catheter or by prostaglandin gel placement if the cervix was unfavourable or by intravenous oxytocin followed by amniotomy if the cervix was favourable. If fetal weight was estimated to be 4500 g or more, then a caesarean section was performed (Lurie, 1996). In the retrospective cohort study, prostaglandin gel was placed vaginally every six to eight hours until labour started, and if induction did not succeed after 5 attempts, or if fetal distress was suspected, then a caesarean section was performed (Alberico, 2010).

Expectant management of pregnancy awaiting spontaneous onset of labour was followed until gestational week 42 in the RCT (Kjos, 1993) and beyond reassessment at gestational week 40 (Lurie, 1996) or week 40 to 41 (Alberico, 2010) if there were no complications. If spontaneous labour had not occurred by these gestational age limits elective delivery was undertaken. In all three studies induction of labour was undertaken if complications arose or, in one study, if the pregnancy exceeded 42 weeks of gestation or if the estimated fetal weight was 4200 g or more (Kjos, 1993). Caesarean section was performed if the estimated fetal weight was more than 4250 g in the prospective cohort study (Lurie, 1996) and was performed electively if the estimated fetal weight was 4500 g or more in the retrospective cohort study (Alberico, 2010).

Data were available for the following outcomes prioritised by the guideline development group:

• mode of birth (Kjos, 1993; Lurie, 1996; Alberico, 2010);
• stillbirth/neonatal death/ perinatal death/infant death (Rosenstein, 2012; Kjos, 1993; Lurie, 1996; Alberico, 2010);
• admission to NICU (Kjos, 1993; Lurie, 1996; Alberico, 2010);
• macrosomia (Kjos, 1993; Lurie, 1996; Alberico, 2010);
• shoulder dystocia (Kjos, 1993; Lurie, 1996; Alberico, 2010).

No data were available in the six studies for outcomes regarding maternal complications of delivery, maternal satisfaction/experiences or neonatal intensive care unti (NICU) stay or more than 24 hours.

6.1.2.4 Evidence profile

Table 74: GRADE profile for incidence of stillbirth by gestational age in pregnancies of women with gestational diabetes compared with women who do not have gestational diabetes.

Table 75: GRADE profile for incidence of neonatal death by gestational age in the babies of women with gestational diabetes compared with women who do not have gestational diabetes.

Table 76: GRADE profile for the incidence of infant death by gestational age in the babies of women with gestational diabetes compared with women who do not have gestational diabetes.

Table 77: GRADE profile for incidence of stillbirth in the babies of women with type 1 and type 2 diabetes compared with all women in England and Wales.

Table 78: GRADE profile for incidence of perinatal mortality in the babies of women with type 1 diabetes compared with women who do not have type 1 diabetes.

Table 79: GRADE profile for effectiveness of elective delivery in pregnant women with gestational diabetes compared with expectant management for maternal outcomes.

Table 80: GRADE profile for effectiveness of elective delivery in pregnant women with gestational diabetes compared with expectant management for fetal/neonatal outcomes.

6.1.2.5 Evidence statements

One retrospective cohort study (n=4,190,953 non anomalous deliveries, table 74) reported a trend of falling stillbirth rates from 36 to 40 weeks of gestation and higher rates thereafter in women with and without gestational diabetes. The incidence of stillbirth was higher in babies of women with gestational diabetes compared with those without throughout weeks 36 to 41, but the risk of stillbirth was increased only during weeks 37 (RR 1.34, 95% CI 1.06 to 1.70) and 39 (RR 1.30, 95% CI 1.01 to 1.66) of gestation in women with gestational diabetes compared with those without. In week 42, the incidence of stillbirth was higher in women without gestational diabetes compared with those with gestational diabetes.

The same study found U-shaped trends for the incidence of neonatal (table 75) and infant death (table 76) in the babies of women with and without gestational diabetes, which were highest for babies delivered at 36 weeks and which fell to a nadir at 39 to 40 weeks before rising again at 41 weeks. The neonatal death rate was higher in babies of women with gestational diabetes delivered at weeks 36, 37, 39 and 42 compared with the babies of women without gestational diabetes. However, there were no differences in the risk of neonatal death between the two groups at any time point. Although the infant death rates were higher in babies of women without gestational diabetes compared with those with gestational diabetes from 36 to 42 weeks gestation, there were no differences in the risk of infant death between the 2 groups at any time point.

In the second study (n=3,522,869), the incidence of stillbirth in women with type 1 and type 2 diabetes also demonstrated a U-shaped trend, being highest at 24 to 27 weeks of gestation, decreasing to a minimum at 37 to 38 weeks of gestation and then rising at 39 weeks’ gestation and later (table 77). The incidence of stillbirth in all women in England and Wales similarly decreased from 24 to 27 weeks of gestation but rather than rising, the minimum rate was observed at 39 weeks’ gestation and later. When the groups were compared, there were no differences in the risk of stillbirth between the two groups at any time point except at 39 weeks’ gestation and later, when there was a higher rate in women with type 1 and type 2 diabetes compared with all women in England and Wales (RR 7.2, 95% CI 1.31 to 39.63).

A third retrospective study (n=1,162,399) reported a U-shaped trend for rates of perinatal mortality in babies of women with type 1 diabetes, with the highest risk at a timepoint of 32 to 34 weeks, dipping to a nadir at week 38 before rising again at 39 weeks (table 78). The value at 40 weeks dipped and then rose again at 41 to 45 weeks. Similarly, rates of perinatal mortality in babies of women without type 1 diabetes was also highest at the 32 to 34 weeks timepoint, but thereafter fell and plateaued from week 39. The risk of perinatal mortality was higher in babies of women with type 1 diabetes compared with those without type 1 diabetes at week 39 (RR 4.25, 95% CI 1.38 to 13.11) and weeks 41–45 (RR 12.42, 95% CI 4.06 to 37.93).

The evidence for all the above mortality outcomes was of very low quality.

When women with gestational diabetes who had an elective delivery were compared with those who were managed expectantly, there were no differences in these mode of birth outcomes: vaginal delivery, spontaneous vaginal birth, operative deliveries or caesarean section (table 79):

• vaginal delivery (RR 1.09, 95% CI 0.91 to 1.29, one RCT, n=200);
• spontaneous vaginal birth (two cohort studies: RR 0.92, 95% CI 0.79 to 1.07, n=260; RR 0.98, 95% CI 0.78 to 1.23, n=99);
• operative deliveries (two cohort studies: RR 0.95, 95% CI 0.33 to 2.75, n=260; RR 3.19, 95% CI 0.34 to 29.60, n=99);
• caesarean section (one RCT: RR 1.28, 95% CI 0.70 to 2.37, n=200; and two cohort studies: RR 1.21, 95% CI 0.75 to 1.97, n=260; RR 0.87, 95% CI 0.40 to 1.91, n=99).

One study compared the risk of caesarean section in women having an elective delivery with those managed expectantly according to body mass index (BMI) but found no differences in caesarean section risk between groups for women with a normal (20 to 25) BMI (odds ratio [OR] 0.99, 95% CI 0.2 to 4.91, n=28) or an obese (30 or more) BMI (OR 0.31, 95% CI 0.04 to 2.14, n=78). The evidence for these findings was of low and very low quality.

There were no differences in the risk of stillbirth, perinatal death, shoulder dystocia or NICU admission in the babies of women with gestational diabetes who had an elective delivery compared with those who were managed expectantly (table 80):

• stillbirth (one cohort study: RR 0.35, 95% CI 0.01 to 8.48, n=99);
• perinatal death (one RCT: RR not calculable [NC], 95% CI NC, n=200; and one cohort study: RR NC, 95% CI NC, n=260);
• shoulder dystocia (one RCT: RR 0.14, 95% CI 0.01 to 2.73, n=200; and two cohort studies: RR 0.26, 95% CI 0.03 to 2.05, n=207; RR NC, 95% CI NC, n=99);
• NICU admission (one cohort study: RR 0.18, 95% CI 0.02 to 1.42, n=99).

The evidence for these findings was of low and very low quality.

With regard to fetal size (table 80), one RCT found that the risk of macrosomia (defined as birthweight more than 4000 g) was lower in babies of women with gestational diabetes who had an elective delivery compared with those who were managed expectantly (RR 0.56, 95% CI 0.32 to 0.98, n=200) although two cohort studies reported no difference in risk of macrosomia (RR 0.51, 95% CI 0.25 to 1.03, n=260; RR 0.58, 95% CI 0.23 to 1.44, n=99). When macrosomia was defined as birthweight more than 4500 g, there was no difference in risk for babies of women who had elective delivery compared with those who were managed expectantly. There were no incidences of respiratory disease (one cohort study, n=260) or neonatal hypoglycaemia (one RCT, n=200) reported in the studies that reported these outcomes. The evidence for these findings was of low and very low quality.

Om de uitgangsvraag te beantwoorden is een systematische literatuuranalyse verricht naar de volgende wetenschappelijke vraagstelling(en):

1. Is er verschil in risico op zwangerschaps-, partus- en neonatale complicaties tussen afwachten van het spontane begin van de bevalling en inleiden van de bevalling bij een zwangere vrouw met diabetes mellitus of GDM met insuline en/of orale glucoseverlagende medicatie zonder complicaties? En is er verschil in frequentie van ongeplande sectio’s caesarea?

2. Als inleiden tot minder complicaties leidt, vanaf welke zwangerschapsduur is inleiden van de baring geïndiceerd?

Relevante uitkomstmaten

De werkgroep achtte de items sectio caesarea, schouderdystocie, neonataal trauma, perinatale sterfte en asfyxie voor de besluitvorming kritieke uitkomstmaten en de items macrosomie, vaginale kunstverlossing en fluxus postpartum voor de besluitvorming belangrijke uitkomstmaten.

De werkgroep definieerde niet a priori de genoemde uitkomstmaten, maar hanteerde de in de studies gebruikte definities.

Zoeken en selecteren (Methode)

In de databases Medline (OVID), Embase and Cochrane is met relevante zoektermen gezocht naar systematische reviews, RCT’s en observationele studies. De zoekverantwoording is weergegeven onder het tabblad Verantwoording. De literatuurzoekactie leverde 256 treffers op. Studies werden geselecteerd op grond van de volgende selectiecriteria:

1. relevant voor beantwoording van de uitgangsvraag;
2. betreft primair (origineel) onderzoek of systematische review;
3. studiepopulatie komt overeen met patiëntengroep waarvoor richtlijn bedoeld is;
4. uitkomstmaten komen overeen met gekozen uitkomstmaten.

Direct na de selectieperiode bleek een NICE-richtlijn (februari 2015) met onder meer dezelfde uitgangsvragen te zijn verschenen, waarin ook grotendeels dezelfde uitkomstmaten worden meegenomen. De werkgroep heeft deze richtlijn vergeleken met de eigen search en uitkomsten en heeft besloten dat deel van de NICE-richtlijn wat betreft literatuursamenvatting integraal over te nemen (NICE guideline Diabetes in Pregnancy, February 2015; 6.1.2 Timing of birth).

De werkgroep heeft ervoor gekozen de samenvatting en beoordeling van de literatuur over te nemen en niet de paragrafen waarin de mening of de overwegingen van de Britse werkgroep zijn opgenomen. De tekst, evidence tabellen en beoordeling van individuele studiekwaliteit zijn derhalve integraal overgenomen en kunt u onder het tabblad Onderbouwing vinden.

Uit de NICE-richtlijn zijn overgenomen uit hoofdstuk 6.1 Timing and mode of birth:

6.1.2 Timing of birth

6.1.2.1 Review question

6.1.2.3 Description of included studies

6.1.2.4 Evidence profile

6.1.2.5 Evidence statements

Voor de verantwoording van literatuursearch en -selectie bij dit hoofdstuk wordt verwezen naar de appendices bij de NICE-richtlijn (appendix D.13 review protocol, appendix E.9 Zoekstrategie, appendix F aantal gevonden en geïncludeerde studies, appendix G.16 verantwoording geëxcludeerde studies, appendix H Evidence tabellen).