Volume 99, Issue 12 p. 1700-1709
ORIGINAL RESEARCH ARTICLE
Free Access

Induction of labor and nulliparity: A nationwide clinical practice pilot evaluation

Ingvil K. Sørbye

Corresponding Author

Ingvil K. Sørbye

Department of Obstetrics and Gynecology, Oslo University Hospital, Oslo, Norway

Correspondence

Ingvil K. Sørbye, Department of Obstetrics and Gynecology, Oslo University Hospital, Sognsvannsveien 20, 0424 Oslo, Norway.

Email: [email protected]

Search for more papers by this author
Kevin S. Oppegaard

Kevin S. Oppegaard

Department of Obstetrics and Gynecology, Finnmark Hospital Trust, Hammerfest, Norway

Search for more papers by this author
Andrew Weeks

Andrew Weeks

Liverpool Women’s Hospital and University of Liverpool for Liverpool Health Partners, Liverpool, UK

Search for more papers by this author
Kjersti Marsdal

Kjersti Marsdal

Department of Obstetrics and Gynecology, Oslo University Hospital, Oslo, Norway

Oslo Metropolitan University, Oslo, Norway

Search for more papers by this author
Anne F. Jacobsen

Anne F. Jacobsen

Department of Obstetrics and Gynecology, Oslo University Hospital, Oslo, Norway

Faculty of Medicine, University of Oslo, Oslo, Norway

Search for more papers by this author
First published: 01 July 2020
Citations: 11

Abstract

Introduction

Induction of labor has become an increasingly common obstetric procedure, but in nulliparous women or women with a previous cesarean section, it can pose a clinical challenge. Despite an overall expansion of medical indications for labor induction, there is little international consensus regarding the criteria for induction, or for the recommended methods among nulliparous women. In this light, we assessed variations in the practice of induction of labor among 21 birth units in a nationwide cohort of women with no prior vaginal birth.

Material and methods

We carried out a prospective observational pilot study of women with induced labor and no prior vaginal birth, across 21 Norwegian birth units. We registered induction indications, methods and outcomes from 1 September to 31 December 2018 using a web-based case record form. Women were grouped into “Nulliparous term cephalic”, “Previous cesarean section” and “Other Robson” (Robson groups 6, 7, 8 or 10).

Results

More than 98% of eligible women (n=1818) were included and a wide variety of methods was used for induction of labor. In nulliparous term cephalic pregnancies, cesarean section rates ranged from 11.1% to 40.6% between birth units, whereas in the previous cesarean section group, rates ranged from 22.7% to 67.5%. The indications “large fetus” and “other fetal” indications were associated with the highest cesarean rates. Failed inductions and failure to progress in labor contributed most to the cesarean rates. Uterine rupture occurred in two women (0.11%), both in the previous cesarean section group. In neonates, 1.6% had Apgar <7 at 5 minutes, and 0.4% had an umbilical artery pH <7.00.

Conclusions

Cesarean rates and applied methods for induction of labor varied widely in this nationwide cohort of women without a prior vaginal birth. Neonatal outcomes were similar to those of normal birth populations. Results could indicate the need to move towards more standardized induction protocols associated with optimal outcomes for mother and baby.

Abbreviations

  • aOR,
  • adjusted odds ratio
  • CI,
  • confidence interval
  • CS
  • cesarean section
  • IUGR
  • intrauterine growth restriction
  • GDM
  • gestational diabetes mellitus
  • Key message

    In induction of labor among women without a prior vaginal birth, large variations in methods used and cesarean rates were observed in this nationwide clinical practice evaluation.

    1 INTRODUCTION

    The worldwide rate of induction of labor has been rising steadily since the turn of the millenium . Currently, approximately 25% of births in high-income countries are induced.1, 2 When faced with unfavorable factors for the mother or the baby if pregnancy continues, induction of labor can be indicated.3 In pregnancies complicated by maternal diabetes or preeclampsia, postterm pregnancies and prolonged prelabor rupture of membranes, induction of labor compared with expectant management reduces the risk of perinatal death and maternal complications.3-7 Over the last decades, an expansion of medical indications for labor induction has occurred, including such conditions as hypertensive disorders,5 advanced maternal age,8 gestational diabetes mellitus (GDM)9 and suspected large fetus for gestation.10 Newer studies have demonstrated the safety of induction of labor without a medical indication, with fetal outcomes and cesarean section (CS) rates comparable to rates among women awaiting spontaneous labor.11

    However, there are some concerns as to the generalizations of these findings into routine practice. First, results produced in a setting with relatively high overall CS rates cannot necessarily be extrapolated to settings with average low CS rates. A clinical challenge is also posed by the considerable number of nulliparous women and women with a previous uterine scar12 giving birth today. Furthermore, induction of labor is not risk-free, as more interventions are performed in induced than in spontaneous labors.13, 14 Finally, in recent studies of induction of labor, few have used standardized and consistent protocols in terms of the methods used. There is currently no international agreement as to what is the best induction method in women without a prior vaginal birth1, 15 and there is large diversity in clinical practice.1, 2

    The authors of this study considered that assessing variation in induction practices in a national sample from a setting with free universal public delivery care and low average CS rates,16 such as Norway, might be a good start to evaluate current practices and results. The aim of this pilot study was to examine variation in indications for induction of labor, methods and associated CS prevalence among women with no previous vaginal birth across 21 birth units nationwide. We used the Robson classification framework to distinguish women with nulliparous term cephalic pregnancies vs those with a previous uterine scar attempting a vaginal birth after cesarean section.17 Ultimately, we aimed to identify practices associated with the best outcomes in terms of maternal and neonatal safety to inform obstetric providers.

    2 MATERIAL AND METHODS

    Between 1 September and 31 December 2018, we carried out a prospective pilot registration of women undergoing induction of labor with a live fetus beyond 23 completed gestational weeks and with no prior vaginal birth. We invited Norwegian obstetric departments with >1000 annual births to participate in the study. Of 22 eligible units, 21 units were included (Supporting Information Figure S1). Participating units selected women whose labor was to be induced and decided upon the method(s) according to local practices, guidelines and definitions. Outpatient induction of labor was not practiced. Anonymous individual patient data were prospectively registered by clinicians in each department on a web-based electronic case record form. Only women with induction of labor were included. The number of nulliparous women without a previous birth and the induction rate during the period were also reported. The paper is reported using the STROBE guidelines for cohort studies.18 Data were stored in Services for Sensitive Data, University of Oslo, Norway. The project is registered in ClinicalTrials.gov, no. NCT03730220.19

    The primary outcome was the occurrence of CS according to indication for induction and method of induction, stratified by obstetric group. Indications for CS were defined according to national/regional guidelines. We also assessed CS rates according to level of birth unit (university hospital or not). Secondary outcomes included uterine rupture, estimated maternal blood loss, adverse neonatal outcomes and the time interval from drug administration to birth. Estimated postpartum blood loss in milliliters was reported in categories. Adverse neonatal outcomes were defined as a composite outcome of Apgar score <7 at 5 minutes and/or transfer to neonatal intensive care unit and/or pH in umbilical artery <7.10 within 1 hour of birth.

    We categorized cases into three groups, as follows: “Nulliparous term cephalic” (Robson 2), "Previous CS" (classified as Robson 5: multiparous women with a previous uterine scar, with a single cephalic term pregnancy; however, with no previous vaginal birth) and “Other Robson” (including Robson groups 6 and 7: women with a single breech pregnancy; Robson group 8: women with multiple pregnancies, and Robson group 10: women with a single cephalic pregnancy <37 weeks’ gestation).

    The indication for induction was categorized into 12 groups: Postdates (as defined locally; latest 42+0 weeks), prelabor rupture of membranes, preeclampsia/hypertension, intrauterine growth restriction (IUGR)/oligohydramnios, insulin-treated diabetes in pregnancy including insulin-treated GDM, non-insulin–treated GDM, suspected large fetus, reduced fetal movements, intrahepatic cholestasis of pregnancy, maternal request, “other maternal” and “other fetal.” The starting method for induction was categorized as Foley balloon catheter, misoprostol (oral, vaginal insert or vaginal tablet) or dinoprostone. As per the protocol, we performed three comparisons: induction regime with Foley balloon catheter vs no catheter; induction regime including misoprostol vaginal insert vs other misoprostol administration forms, and induction regime including dinoprostone vs misoprostol.

    Other covariates included maternal age in categories, prepregnancy body mass index (<30 or ≥30), gestational age at induction, Bishop score at induction (≤5, >5 or missing), epidural, infant birthweight and tachysystole (>5 contractions per 10 seconds with abnormal fetal tracing).

    2.1 Statistical analyses

    A statistical analysis plan included a power analysis. Assuming two groups of birth units with different induction methods resulting in a difference in CS rate between 20% and 25%, a significance level (α) of 0.05 and 80% power (β), the study would need 2182 participants. Applying the inclusion criteria, we estimated 2250 births during the period.20 Baseline characteristics and outcomes were summarized according to the obstetric group. Small cell numbers (n < 10) were censored when calculating CS rates. For categorical outcomes we compared proportions with 95% confidence interval (CI) with the Chi-square test and/or Fisher’s exact test. We estimated the risk of CS by logistic regression analysis in generalized linear models adjusting for confounders as identified in the literature and according to biological plausibility, estimating crude and adjusted effect estimates as odds ratios (OR) with 95% CI with corresponding P values. We restricted analyses to nulliparous term cephalic and previous CS only, due to small cell numbers. In subanalyses of CS deliveries only, we determined indications for the procedure and the subtype (type 1—immediate delivery; type 2—within 20-30 minutes or type 3—within a given timeframe >30 minutes and <8 hours). Calculated P values were two-sided and compared with a 5% significance level. Statistical analyses were performed in IBM SPSS Statistics for Windows, Version 25.0. (IBM Corp., Armonk, NY, USA).

    3 RESULTS

    During the study period, a total of 7160 women without a prior vaginal birth gave birth in the 21 participating departments. Of these, 1874 were induced. (26.2%; range 11.7%-34.3% between hospitals). Of all eligible women for inclusion in the study, 1818 (98.5%) were included (Figure 1). Most birth units had a written induction protocol; however, most were not differentiated according to their Robson group.

    Details are in the caption following the image
    Flowchart of study participants

    Nulliparous term cephalic pregnancies constituted 80.4% of births, followed by previous CS pregnancies (12.2%) (Table 1). The “Other Robson” group included 52 twin pregnancies (2.9%), 59 preterm births (3.2%) and 25 planned breech births (1.4%). Overall, 20.3% were aged 35 years or more and 18.4% had a prepregnancy body mass index of 30 or more. Of all women, 16.6% had a registered comorbidity, preeclampsia/hypertension being the most prevalent. Prelabor rupture of membranes, postdate pregnancy and preeclampsia/hypertension were the most common indications for induction. Maternal request was an indication for induction in only 3.5% of women.

    TABLE 1. Maternal characteristics in 1818 women with no prior vaginal birth undergoing induction of labor
    All Nulliparous term cephalic Previous CS Other Robson
    n % n = 1461 % n = 221 % n = 136 %
    Maternal age (years)
    16-24 246 13.5 222 15.2 15 6.8 9 6.6
    25-34 1203 66.2 956 65.4 149 67.4 98 72.1
    35-54 369 20.3 283 19.4 57 25.8 29 21.3
    BMI prepregnancya, a
    <30 1458 81.6 1185 82.3 160 74.8 113 85.0
    ≥30 328 18.4 254 17.7 54 25.2 20 15.0
    Bishops scoreb
    0-5 1366 82.8 1077 81.8 185 88.9 104 83.2
    6-10 284 17.2 240 18.2 23 11.1 21 16.8
    Birth at University hospital
    Yes 1173 64.5 957 65.5 124 56.1 93 68.4
    Gestational age
    Median (IQR) 40 + 1 (21) 40 + 3 (18) 40 + 0 (18) 36 + 5 (16)
    Maternal comorbidityc
    IDDM/GDM insulin 137 7.5 105 7.2 22 10.0 10 7.4
    GDM, non-insulin 91 5.0 74 6.4 7 3.2 10 7.4
    Preeclampsia/hypertension 238 13.1 189 12.9 13 5.9 37 27.2
    Intrahepatic cholestasis 34 1.9 25 1.7 6 2.7 3 2.2
    Another comorbidity 272 15.0 210 14.4 39 17.6 23 5.9
    Decision induction
    Consultant 1181 65.0 897 61.4 162 72.4 123 90.4
    Resident 534 29.4 472 32.3 49 22.2 13 9.6
    Midwife 103 5.7 93 6.4 10 4.5 0 0
    Main indication for induction
    PROM 357 19.3 286 19.6 47 21.3 24 17.6
    Postdates 336 18.5 299 20.5 33 14.9 4 2.9
    Preeclampsia/hypertension 279 15.3 228 15.6 17 7.7 34 25.0
    IUGR/oligohydramnios 280 15.4 231 15.8 24 10.9 25 18.4
    IDDM/GDM – insulin 97 5.3 81 5.5 13 5.9 3 2.2
    Large fetus 67 3.7 45 3.1 16 7.2 6 4.4
    Maternal request 61 3.5 36 2.5 22 10.0 3 2.2
    GDM, non-insulin 35 1.9 32 2.2 3 1.4 0 0
    Intrahepatic cholestasis 43 2.4 35 2.4 5 2.3 3 2.2
    Reduced fetal movements 40 2.2 36 2.5 3 1.4 1 0.7
    Other maternald 164 9.0 101 6.9 33 14.9 30 22.1
    Other fetale 59 3.2 51 3.5 5 2.3 3 2.2

    Note

    • Other Robson includes Robson groups 6, 7, 8 and 10.
    • Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus; IDDM, insulin-dependent diabetes mellitus; IQR, interquartile range; IUGR, intrauterine growth restriction; PROM, prelabor rupture of membranes.
    • a Missing 1.8%.
    • b Not assessed in 9.2%.
    • c More than one condition might be registered.
    • d Including twin pregnancy, previous obstetric history, chronic disease, prolonged latency phase, vaginal bleeding.
    • e Including polyhydramnios, non-reassuring fetal tracing, known malformations, unknown gestational length.

    In the nulliparous term cephalic group, one in five women gave birth by CS, whereas in the previous CS group and in the Other Robson group, rates were doubled (Table 2). Hospital CS rates varied between 9.4% and 45.5% in the nulliparous term cephalic group and between 31.3% and 54.5% in the previous CS group (Figure 2). In the whole cohort, university hospital CS rates did not differ significantly from non-university hospitals rates (24.2% vs 26.8%, P = .22). In the nulliparous term cephalic group, CS rates by indication for the induction of labor ranged from 11.1% to 40.6%, whereas in the previous CS group, rates were overall higher and ranged from 22.7% to 67.5% (Figure 3; Supporting Information Table S1.). In the nulliparous term cephalic group, “non-insulin GDM”, “other fetal” and “large fetus” were the indications associated with the highest CS rates (40.6%-33.3%). In the previous CS group “large fetus”, “insulin-treated diabetes” and “other fetal” were associated with the highest CS rates (60.0%-62.5%).

    TABLE 2. Delivery mode after induction of labor in 1818 women with no prior vaginal birth according to obstetric group
    N Cesarean sectiona, a Operative vaginal Spontaneous vaginal
    n % 95% CI n % 95% CI n % 95% CI
    Nulliparous term cephalic 1461 320 21.9 19.8-24.1 314 21.5 19.4-23.7 827 56.6 54.0-59.2
    Previous CS 221 89 40.3 33.7-47.1 40 18.1 13.3-23.8 92 41.6 35.1-48.4
    Other Robson 136 50 36.8 28.9-45.8 28 20.6 14.2-28.6 60 44.1 35.9-53.2
    All 1818 459 25.2 23.3-27.3 382 21.0 19.2-23.0 979 53.9 51.5-56.2
    • Other Robson includes Robson groups 6, 7, 8 and 10.
    • a Including cesarean section second twin.
    Details are in the caption following the image
    Proportions of cesarean section (CS) after induction of labor by delivery unit in nulliparous term cephalic (a) and previous CS pregnancies (b). Other Robson includes Robson groups 6, 7, 8 and 10. Results from delivery units with <10 deliveries per cell are censored
    Details are in the caption following the image
    Cesarean section (CS) rates according to indication for induction of labor in nulliparous term cephalic and previous CS pregnancies. GDM, gestational diabetes mellitus; IDDM, insulin-dependent diabetes mellitus; IUGR, intrauterine growth restriction; PROM, prelabor rupture of membranes; ICP, intrahepatic cholestasis of pregnancy

    The most common CS indication was suspected fetal hypoxia in the nulliparous term cephalic group and failed induction in the previous CS group (Table 3). Of all cesarean procedures, 9.2% were reported as grade 1 (immediate) (Table 3). Overall, 2.1% of women in the nulliparous term cephalic group and 3.6% of the previous CS group experienced an immediate CS. Suspected uterine rupture or abruptio placentae were indications for seven (0.4%) cesarean procedures.

    TABLE 3. Main indication and subtype of 459 cesarean sectionsa after induction of labor according to obstetric group
    All Nulliparous term cephalic Previous CS Other Robson
    n = 459 % n = 320 % n = 89 % n = 50 %
    Main cesarean indication
    Prolonged 1st stage 117 25.5 85 26.6 23 25.8 9 18.8
    Prolonged 2nd stage 26 5.7 19 5.9 4 4.5 3 6.5
    Suspected fetal hypoxia 143 31.2 112 35.0 19 21.3 12 25.0
    Failed induction 109 23.7 69 21.6 31 34.8 9 18.8
    Uterine rupture 2 0.4 0 0 2 2.2 0 0
    Abruptio placentae 5 1.1 3 0.9 1 1.1 1 2.1
    Other 55 12.0 31 9.7 10 11.2 14 29.2
    Subtype
    Type 1 (immediate) 42 9.2 30 9.4 8 9.0 6 12.0
    Type 2 (<20 minutes) 234 51.0 172 53.8 36 40.4 26 52.0
    Type 3 (>20 minutes) 181 39.4 118 36.9 45 50.6 18 36.0
    • a Including cesarean section of second twin only (n = 2). Other Robson includes Robson groups 6, 7, 8 and 10.

    The most common methods of induction are presented in Table 4. In the nulliparous term cephalic group, a combination of Foley + misoprostol was the most common initiation method (37.3%), followed by Foley + amniotomy/oxytocin (11.9%) (Table 4). In the previous CS group, most women received Foley + dinoprostone (34.4%), followed by Foley + amniotomy/oxytocin. However, more than 40 different method combinations and sequences were registered altogether.

    TABLE 4. Induction methods in 1818 women with no prior vaginal birth according to obstetric group
    Induction method All Nulliparous term cephalic Previous CS Other
    n = 1818 % n = 1461 % n = 221 % n = 136 %
    Foley start combinations
    Foley alone 135 7.4 102 7.0 19 8.6 14 10.3
    Foley + oral misoprostol ± AT/oxytocin 191 10.5 178 12.2 4 1.8 9 6.6
    Foley + insert misoprostol ± AT/oxytocin 198 10.9 190 13.0 1 0.5 7 5.1
    Foley + vaginal misoprostol ± AT/oxytocin 213 11.7 177 12.1 11 5.0 25 18.4
    Foley + dinoprostone ±AT/oxytocin 108 5.9 28 1.9 76 34.4 4 2.9
    Foley ± AT/oxytocin 241 13.3 174 11.9 49 22.2 18 13.2
    Misoprostol start combinations
    Oral misoprostol alone 118 6.5 107 7.3 1 0.5 10 7.4
    Oral misoprostol ± AT/oxytocin 45 2.5 41 2.8 0 0 4 2.9
    Insert misoprostol alone 67 3.7 66 4.5 0 0 1 0.7
    Insert misoprostol ± AT/oxytocin 29 1.6 28 1.9 0 0 1 0.7
    Vaginal misoprostol alone 165 9.0 148 10.1 3 1.4 14 10.3
    Vaginal misoprostol ± AT/oxytocin 88 4.8 79 5.4 1 0.5 8 5.9
    Other combinations
    Dinoprostone alone 39 2.1 9 0.6 26 11.8 4 2.9
    Dinoprostone ± AT/oxytocin 11 0.6 2 0.1 8 3.6 1 0.7
    Amniotomy ± oxytocin 130 7.2 103 7.0 19 8.6 8 5.9
    Any misoprostol/dinoprostone + successive Foley ± AT/oxytocin 24 1.3 17 1.2 3 1.4 4 2.9
    Other 16 0.9 12 0.8 0 0 4 2.9

    Note

    • Other Robson includes Robson groups 6, 7, 8 and 10.
    • Abbreviation: AT, amniotomy.

    Use of Foley catheter was associated with birth by CS in the nulliparous term cephalic group (adjusted odds ratio [aOR] 1.71, 95% CI 1.14-2.55, P = .009) but not in the previous CS group (aOR 0.62, 95% CI 0.19-2.06, P = .44) (Table 5). Use of dinoprostone showed a borderline significant association with birth by CS compared with misoprostol. There was no association between route of administration of misoprostol and risk of CS (data not shown).

    TABLE 5. Method and risk of cesarean section in nulliparous term cephalic and previous CS pregnancies after induction of labor
    Proportion CS Risk of cesarean section
    CS % OR 95% CI p aORb 95% CIb P b
    1. Foley a (n = 1356)
    Nulliparous term cephalic
    Foley catheter 212 25.2 1.55 1.13-2.13 0.007 1.71 1.14-2.55 .009
    No Foley catheter 61 17.9 1 1
    Previous CS
    Foley catheter 60 40.3 1.08 0.46-2.54 0.86 0.62 0.19-2.06 .44
    No Foley catheter 10 38.5 1 1
    2. Dinoprostone vs misoprostol (n = 1195)
    Nulliparous term cephalic
    Dinoprostone 14 34.1 1.80 0.93-3.49 0.082 1.90 0.90-4.03 .09
    Misoprostol 230 22.4 1 1
    Previous CS
    Dinoprostone 49 47.1 1.29 0.51-3.27 0.60 1.02 0.32-3.23 .97
    Misoprostol 9 40.9 1 1
    • a Excluding women with prelabor rupture of membranes.
    • b Adjusted for maternal age groups 16-24, 25-34 and ≥35; prepregnancy BMI < 25, 25-29, 30-34, ≥35; Bishop score ≤5, >5 and missing; birthweight in grams and Foley catheter yes/no.

    Uterine rupture occurred in two women (0.11%), both in the previous CS group (Table 6). Maternal blood loss differed between groups (P = .049, Chi-square test); however, tachysystole did not. The composite adverse infant outcome occurred in 9.5% and 10.0% in the nulliparous term cephalic and the previous CS group, respectively. A higher proportion (30.9%) was found in the Other Robson groups due to more transfers to the neonatal ward because of prematurity. Overall, 29 infants (1.6%) had an Apgar score of <7 at 5 minutes. Seven infants (0.4%) had an umbilical artery pH <7.00, one of whom was transferred to the neonatal ward. The method of induction was not associated with an adverse maternal or neonatal outcome.

    TABLE 6. Maternal and fetal secondary outcomes after induction of labor according to obstetric groups
    All Nulliparous term cephalic Previous CS Other
    Maternal n % n % n % n %
    Uterine rupture 2 0.1 0 0 2 0.9 0 0
    Tachysystole 96 5.3 79 5.4 10 4.5 7 5.1
    Epidural 1355 74.5 1090 74.6 157 71.0 108 79.4
    Blood loss in Ml
    <500 1051 57.8 863 59.1 120 54.3 68 50.0
    500-999 552 30.4 439 30.0 69 31.2 44 32.4
    1000-1999 178 9.8 135 9.2 24 10.9 19 14.0
    2000-2999 33 1.8 22 1.5 6 2.7 5 3.7
    3000+ 4 0.2 2 0.1 2 0.9 0 0
    Fetala
    Mean birthweight in grams (SD) 3485 (597) 3513 (550) 3664 (522) 2887 (808)
    Adverse neonatal outcomeb 203 11.2 139 9.5 22 10.0 42 30.9
    Transfer NICU 132 7.4 85 5.9 10 4.6 37 27.6
    Apgar <7 at 5 min 29 1.6 20 1.4 7 3.2 2 1.5
    pH art umb <7.10c 72 4.0 59 4.0 9 4.0 4 2.9
    pH art umb < 7.00c 7 0.4 6 0.4 1 0.5 0 0

    Note

    • Other Robson includes Robson groups 6,7, 8 and 10.
    • Abbreviation: arthumb, arteria umbilicalis.
    • a Outcomes for first twin only. SD, standard deviation.
    • b Adverse neonatal outcome incl. pH arteria umbilicalis <7.10 and/or Apgar score at 5 min <7 and/or transfer neonatal intensive care unit. NICU, neonatal intensive care unit excluding planned transfers (n = 18).
    • c Missing 19.5%.

    Among nulliparous term cephalic births, 26.5% were still undelivered 48 hours after induction start, as were 31.6% in the previous CS group and 26.5% in the Other Robson group (data not shown). In the three groups, median duration from start of induction to birth were 32.6 hours (interquartile range 31.8), 34.1 hours (interquartile range 35.1) and 30.6 hours (interquartile range 32.6), respectively.

    4 DISCUSSION

    Our study showed large variations in the practice and results of induction of labor in this nationwide sample. The frequency of CS after induction of labor was highest in the previous CS group, where about two in five women gave birth by CS, and lowest in the nulliparous term cephalic group, where about one in five women gave birth by CS. CS rates after induction differed widely between units. CS performed due to failed induction of labor and prolonged first stage of labor accounted for nearly half of all CS. Our study found a wide variation of induction methods, with few units using standard induction protocols. Maternal and fetal safety outcomes were comparable to those in the existing literature.

    The strengths of this pilot study include the nationwide prospective design with more than 98% of eligible women included. We had access to detailed information regarding indications, the methods used, including the order and route of administration, as well as important safety and efficiency outcomes.

    One of the limitations of the study is that we lacked control data from induced multiparous women as well as on spontaneous labors. For this reason, we cannot comment on whether induction increases the rate of CS or adverse outcomes in comparison with spontaneous birth. Furthermore, we lacked detailed data on the local birth units, such as the number of referrals, socioeconomic spread etc. that might influence outcomes in terms of mode of birth. In the previous CS group we lacked information regarding the previous birth. However, CS rates were slightly lower in tertiary referral university hospitals than in non-university hospitals, where an accumulation of risks would be expected. Finally, our observational design does not warrant causal inference.

    Induction by “large fetus” indication revealed high rates of CS in our study. However, the CS rate at 33.3% is similar to other studies of induction in woman with “large babies”. In the comprehensive study by Boulvain et al,10 there was a CS rate of 28%, even though 53% were parous. These rates might be the result of a high gestational age in combination with maternal diabetic comorbidity. GDM non-insulin comorbidity had the highest CS rate, whereas insulin-treated pregestational or gestational diabetes comorbidity had a relatively low CS rate in nulliparous term cephalic pregnancies. In Norway, insulin users are induced between weeks 38 and 40, but women with non-insulin GDM are induced primarily on additional indications.21

    “Other fetal indication” for induction of labor had one of the highest CS rates in both nulliparous term cephalic and previous CS pregnancies. This is a mixed group including fetal malformations, polyhydramnios, non-reassuring antenatal fetal tracing and unknown gestational length. Polyhydramnios may cause insufficient contractions due to an over-distended uterus22 and non-reassuring fetal tracing must be handled with care; if it continues, delivery rather than expectant management is preferred. The group “maternal request” was surprisingly low with 3.5% of all inductions and we found a low CS rate both in the nulliparous term cephalic and the previous CS group. This is lower than previously reported.23 The distinction between “maternal request” or “medical problem” can be a fine one, especially when considering mental health and pregnancy complaints. However, this finding indicates a restrictive attitude among providers, in contrast to upcoming trends elsewhere.24

    The overall proportion of failed induction and prolonged first stage was unexpectedly high in our sample. However, as 27%-32% of women were undelivered 48 hours after the start of induction, this is not likely to reflect a use of rigid time limits. The 22%-35% rate of failed induction/poor progress in the first stage that we found in our sample might imply a practice emphasizing safety rather than effectiveness. This is also reflected in a relatively low uterine rupture rate, a low tachysystole rate of 5%, and few immediate CS procedures.

    At present, there are conflicting reports on how and when induction of labor should be offered to women. Trials have been conducted among women at term with no medical indications.11, 25 These randomized trials indicate no major safety concerns in terms of the CS risk or adverse infant or maternal outcomes. In addition, although the ARRIVE trial has been criticized as including many overweight and obese women,26 the 18.6% CS rate in the trial’s induced group (who were all low-risk nulliparous women) is similar to the 16.7% rate seen in “maternal request” in the nulliparous term cephalic group in our study. A Cochrane review looking at induction at 40 weeks vs expectant management found improved outcomes in the induction group, except for a higher operative vaginal delivery rate.3 However, a prerequisite in generalizing findings is that the induction process and labor are well managed with the necessary staff at hand. Like most high-resource countries, Norway has a rapidly increasing induction rate, reaching 23% of all births in 2018,20 but at the same time, overall CS rates–16.0% in 2017–are the second lowest across the Organization for Economic Co-operation and Development area.16 However, CS rates vary considerably between regions, beyond what can be expected due to case-mix.27 A national induction guideline lists medical indications and methods but leaves the choice of methods to individual departments and staff.21 In this clinical practice evaluation, we found that multiple induction protocols are used even within nulliparous term cephalic and previous CS groups. What this means is that women across the country do not receive similar treatment when undergoing induction of labor.

    Translating randomized controlled trial evidence into practical clinical protocols can be challenging in obstetric units facing logistical restraints such as delays in timely administration of uterotonics and performing rupture of membranes.28 Results from practice evaluations are therefore important to inform decisions in induction regimes tailored to specific groups. Women should be offered joint decision-making based on these facts. Careful selection of women for induction who have previously had a CS, as well as taking women’s preferences into account, are important factors in a pragmatic induction of labor protocol.

    5 CONCLUSION

    A wide variation of induction methods and CS rates after induction, as well as a high rate of failed inductions in women without a prior vaginal birth, point to a potential for improvement by moving towards more standardized protocols. The Robson groups provide a framework for the counseling of women about particular risks and benefits regarding induction of labor while working towards shared decision-making.

    ACKNOWLEDGMENTS

    We are indebted to our dedicated obstetrician and midwife collaborators in the National Induction Group at the participating birth units who made this study possible: Ines Panadero at Akershus University Hospital, Nina Marie Albretsen at Arendal Hospital, Kristin Hestvold at Drammen Hospital, Mette Kristine Hjertaas at Førde Hospital, Anja Holstad at Gjøvik Hospital, Line Olufsen-Melhus at Hammerfest Hospital, Kristin Urnes at Haugesund Hospital, Chen Sun at Haukeland University Hospital, Marte Eline Ween-Velken at Kristiansand Hospital, Dordi Bogfjellmo at Levanger Hospital, Jakob Nakling and Ida Olsen Hokland at Lillehammer Hospital, Kristin Skogøy at Nordland Hospital, Bodø, Anja Halleraker and Marianne Omland at Oslo University Hospital Rikshospitalet, Hilde Sellevoll and Marit Småvik Johansen at Oslo University Hospital Ullevål, Kjersti Skoe at Telemark Hospital, Ewa Margas at Tønsberg Hospital, Åse Torunn Pettersen at University Hospital of North Norway, Malin Dögl at St. Olavs University Hospital, Trondheim, Erik Andreas Torkildsen at Stavanger University Hospital, Katrine Sjøborg Dønvold and Lotte Martine Jacobsen at Østfold Hospital, and Åse Turid Rossevatn Svoren at Ålesund Hospital.

      CONFLICT OF INTEREST

      The authors state that there are no conflicts of interest in connection with this article.

      ETHICAL APPROVAL

      All women received oral and written information about the study. As routine data were gathered anonymously, informed consent was waived; however, women were able to opt out of the study upon request. The project was approved by the Norwegian Ethics Board, Region Health South East C, reference 2018/1087 on 27 June 2018 and by each hospital’s Data Protection Officer.