SEARCH

SEARCH BY CITATION

Keywords:

  • Labor induction;
  • expectant management;
  • cesarean delivery;
  • gestational age;
  • pregnancy

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

Objective. To compare induction of labor and expectant management by gestational week with regard to the need for cesarean section (CS) in labor. Design. Cohort study. Setting. National study based on the Danish Birth Registry. Population. Aggregated data from 230 528 deliveries from 2004 until mid-year 2009. Women with cesarean section before labor, previous cesarean, preterm birth, breech presentation, multiple pregnancy and specified medical illnesses in pregnancy were excluded. Methods. We compared by gestational week nulliparous and parous women who were either induced or had expectant management until later spontaneous or induced labor. For each of five comparisons, we conducted multivariable logistic regression analysis, controlling for body mass index, age, smoking in pregnancy and use of epidural analgesia during labor. Main Outcome Measure. Rate of CS in labor. Results. From gestational week 39 and thereafter, there was no difference with regard to CS rates in labor among nulliparous and parous women when comparing women with induced labor and those women who waited for a later labor, either induced or spontaneous. The odds ratios (with confidence intervals) for CS in labor in nulliparous women were 0.99 (0.84–1.17), 1.16 (1.04–1.30), 1.04 (0.94–1.15), 1.00 (0.92–1.10) and 0.97 (0.88–1.07) for weeks 37–41, respectively. For parous women the corresponding figures were 1.72 (1.35–2.20), 1.27 (1.04–1.55), 1.15 (0.95–1.39), 1.18 (0.99–1.40) and 1.07 (0.87–1.32), respectively. Conclusions. Induction of labor provides a sound tool when counseling a woman either going past term or presenting with a problem, even if it is not a severe medical illness.


Abbreviations: 
BMI

body mass index

CI

confidence interval

CS

cesarean section

Exp

expectant management

IOL

induction of labor

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

Traditionally in observational studies, outcomes of induced labor are compared with spontaneous labor (1,2). This type of comparison typically shows a factor of 1.5–2.0 for a raised risk of cesarean section (CS) after induced labor when compared with spontaneous labor, both in nulliparous and in parous women. However, this type of comparison is not relevant because the choice in the clinical situation is between induction of labor and expectant management. It is obviously not an option to choose immediate, spontaneous labor. In order to overcome this problem when using registry-based data, in 2006 Caughey et al. proposed a new method, creating a comparison between a group of women with induction of labor (IOL group) in a specific gestational week and a group of women with expectant management who waited for a later either induced or spontaneous labor (Exp group) (3). In this way, the method took into consideration both the aspect of waiting in the Exp group and the aspect of comprising all women at risk, i.e. all women still pregnant at the time of IOL (4,5).

Caughey et al. used observational data, and the adjusted odds ratios (ORs) for the risk of emergency CS in nulliparous women at term were all significantly in favor of induction of labor in a specific gestational week when comparing this group with women who waited for a later gestational week and ended up with either induced or spontaneous labor. Among parous women, there was no difference in the CS rates after adjustment for potential bias (3).

In a Cochrane Database review, the protective effect from induction on the risk of CS was significant in gestational weeks 37–40, but not in week 41 and beyond (6), while in two meta-analyses (7,8) the risk for CS was significantly lower in the induced group from week 41. In yet another recent systematic review with meta-analysis, comparing elective induction of labor with expectant management, a significant odds ratio of 1.2 for CS risk was found for expectant management at or beyond 41 completed weeks (9). Of special interest is that a difference was found between studies conducted in the USA compared with studies conducted outside that country. In the non-USA studies, a significant protective effect of inducing labor was found, while in the USA studies there was no difference. The most recent of the included randomized controlled trials, a Norwegian study, reported no difference between CS rates in the induced group compared with the expectant group after controlling for potential confounders (10).

We wanted to add to international comparisons between women with induced labor and women with a later labor, by using a large national database on observational data to assess the risk of CS in labor and applying the new way of comparing the two management strategies proposed by Caughey et al. (3).

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

An observational cohort was created with aggregated data from the national Danish Birth Registry and examined for the risk of CS in labor, comparing induction of labor with a later induced or spontaneous labor. The Birth Registry comprises data from all pregnancies and deliveries in Denmark. Data on the woman and outcome of delivery are reported by the attending midwife to the Birth Registry shortly after birth. From 2004, prepregnancy body mass index (BMI) was added as a routine for all registrations. The registry has previously been validated (11).

We first considered data from a total of 352 844 women delivering from 1 January 2004 until 30 June 2009. We excluded women from the study cohort where there was missing information about parity, age, year of delivery, BMI or presentation. Thus, we excluded almost one-third of the group and ended up with 230 528 deliveries which were used in the analysis. Table 1 shows information on all exclusion criteria. By this, we attempted to exclude conditions that might possibly influence the analysis. The criteria for exclusion were adopted from the definitions by the Danish Society for Obstetricians and Gynecologists (DSOG) and the National Board of Health in Denmark (12). Women with the following indications for inducing labor remained in the cohort used for analysis: essential hypertension, varicose veins, unstable lie, intrauterine fetal demise, intrauterine growth retardation, polyhydramnios, prelabor rupture of membranes, prolonged pregnancy, malnutrition and urinary tract infection.

Table 1.  Exclusion criteria used for the selected cohort 2004–2009.
 nPercentage
  1. Note: The groups are not mutually exclusive.

  2. * Pre-eclampsia, eclampsia, hemolysis–elevated liver enzymes–low platelets syndrome, deep vein thrombosis, gestational diabetes, intrahepatic cholestasis of pregnancy, congenital malformation of the fetus, erythrocyte immunization and other medical illnesses.

  3. † Parity, age, year of delivery, body mass index, presentation, use of epidural analgesia during labor, gestational week or onset of labor.

Total cohort352 844100.0
Cesarean section before labor40 60711.5
Previous cesarean section39 55811.2
Preterm birth (<37+0weeks)28 2698.0
Breech presentation and transverse lie16 1544.6
Spontaneous delivery in gestational week 37+0–37+611 0843.1
Multiple pregnancy8 0682.3
Specified medical illnesses in pregnancy*4 6251.3
Cesarean section because of complication in pregnancy3 3430.9
Missing information†47 65913.5
Cohort after exclusion230 52865.3

In subgroup analyses, we excluded women with missing information about smoking during pregnancy. The eligible women were categorized into different subgroups by BMI and age, while use of epidural anesthesia and smoking during pregnancy were dichotomized (Table 2).

Table 2.  Distribution of patient characteristics in the group with induced labor compared with the group with spontaneous labor.
Column percentages in different categoriesInduction of labor (n=34 499)Spontaneous labor (n=196 029)p-Value (χ2) (two-tailed)
nPercentagenPercentage
  1. *Women with missing information have been excluded (n induced=33 801 and n spontaneous=192 637).

Nulliparous17 87851.892 57347.2<0.001
Gestational age (weeks)    <0.001
 372 5247.3Not included 
 385 20015.121 35410.9 
 395 76316.750 23825.6 
 406 78619.771 22236.3 
 415 20615.145 55723.2 
 428 93625.97 5373.8 
 43+840.21210.1 
Smoking in pregnancy*4 36712.924 23112.6 0.082
Epidural use12 02834.934 81017.8<0.001
Age (years)    <0.001
 ≤19.95101.52 9721.5 
 20–24.93 66510.621 53911.0 
 25–29.910 89631.667 48734.4 
 30–34.912 51236.372 77237.1 
 ≥356 91620.031 25915.9 
Body mass index (kg/m2)    <0.001
 ≤18.41 1403.38 7684.5 
 18.5–24.919 00555.1130 24066.4 
 25–29.98 16623.738 77519.8 
 ≥306 18817.918 2469.3 

There are 30 delivery departments in Denmark in all. The largest delivery departments are tertiary centers with about 5 000 deliveries per year and the smallest have around 1 000 deliveries per year. Since 1997, obstetricians from the whole country have joined annually in a consensus meeting to agree on national guidelines for pregnancy and delivery (http://www.dsog.dk).

During the study period, all women were offered an ultrasound scan at 12 and 18weeks of gestation. Estimated date of delivery was determined by ultrasound scan in most departments at the 12week scan and at the 18week scan in others. Induction of labor was carried out using prostaglandin vaginally or in the case of a ripe cervix by artificial rupture of membranes and, if necessary, oxytocin intravenously. In the case of prelabor rupture of membranes, a varying number of hours up to 24hours were allowed until spontaneous onset of labor. If the time limit was exceeded, an oxytocin infusion was used.

We established one group comprising all nulliparous (110 451) and another group with all parous women (120 077). For each of these, we conducted a comparison between women who had their pregnancy induced in a specific gestational week with the group of women who waited for a later labor. We used the rate of CS in labor as the outcome variable, denoted in the NOMESCO Classification of Surgical Procedures as KMCA 10E: acute (=emergency) CS in labor because of a complication in labor (http://nomesco-eng.nom-nos.dk).

For each comparison, we conducted multivariable logistic regression analysis, controlling for BMI, age, smoking and use of epidural analgesia during labor. Open Epi version 2.2.1 (Open Source Epidemiologic Statistics for Public Health; http://www.OpenEpi.com) was used for statistical analysis of patient characteristics. The SAS system (SAS Institute Inc, Cary, North Carolina, USA) was used for logistic regression analysis. The study was approved by the Danish Data Protection Agency (2010-41-5040), and the extraction of aggregated data from the Medical Birth Registry was approved and performed by the National Board of Health (7-505-29-1581/1).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

Table 2 shows patient characteristics in the group of women with spontaneous labor and in the group with induced labor. The two groups of women with either induced or spontaneous labor differed significantly with regard to parity, gestational age, use of epidural analgesia during labor, age and prepregnancy BMI.

The total CS rate in labor in the whole cohort was 7.3% (12.8% in nulliparous and 2.3% in parous women). The overall induction rate was 15.0%, i.e. 16.2% for nulliparous and 13.8% for parous women. Women with induced labor had a 14.5% CS rate, compared with 6.1% in the group of women with spontaneous onset of labor (p<0.001). Among nulliparous women, the rate of epidural analgesia during labor was 32.6%, but 9.0% among the parous. Of the nulliparous women 11.9% smoked, while 13.3% of parous women smoked. The rates of CS in subgroups of the nulliparous and parous women are shown in Table 3, as well as the CS rates in spontaneous and induced labor by specified gestational weeks.

Table 3.  Rates of cesarean section in specified subgroups by parity.
AllNulliparousParousTotal
  1. Note: All percentages refer to the cesarean section rate in the specific group.

Cesarean section rate (%)12.82.37.3
Subgroups
Spontaneous labor10.71.96.1
 Week 386.11.8 
 Week 397.61.3 
 Week 4010.01.8 
 Week 4114.82.5 
 Week 4222.54.2 
 Week 4320.81.5 
Induced labor23.44.914.5
 Week 3716.15.8 
 Week 3818.94.3 
 Week 3919.24.1 
 Week 4022.25.0 
 Week 4126.95.6 
 Week 4228.15.3 
 Week 4344.410.3 
Epidural use during labor24.08.320.4
Smoking during pregnancy12.52.67.3
Body mass index subgroups (kg/m2)
 ≤18.47.71.2 
 18.5–24.911.11.9 
 25–29.915.72.7 
 ≥3020.83.9 
Age groups (years)
 ≤19.97.01.8 
 20–24.99.71.5 
 25–29.911.61.9 
 30–34.914.62.2 
 ≥3519.83.3 

The CS rate was significantly higher in the induced group when comparing nulliparous women who had their labor induced with the group waiting for a later labor until week 41 (unadjusted OR; Table 4). A similar result was seen among parous women (unadjusted OR; Table 5).

Table 4.  Induction of labor compared with expectant management in nulliparous women.
Gestational week (IOL)Cesarean section in induced labor (%)Cesarean section in expectant management (%)IOL/Exp (unadjusted OR and 95% CI)IOL/Exp (adjusted OR and 95% CI)
  1. Note: The group with induced labor in a specific gestational week is compared with the group of women waiting for a later either induced or spontaneous labor. Abbreviations: CI, confidence interval; Exp, expectant management; IOL, induction of labor; OR, odds ratio.

  2. *p<0.05.

3716.112.71.32 (1.13–1.54)*0.99 (0.84–1.17) 
3818.913.31.52 (1.37–1.68)*1.16 (1.04–1.30)*
3919.214.91.35 (1.23–1.49)*1.04 (0.94–1.15) 
4022.218.81.23 (1.13–1.34)*1.00 (0.92–1.10) 
4126.925.71.07 (0.97–1.17) 0.97 (0.88–1.07) 
4228.1
Table 5.  Induction of labor compared with expectant management in multiparous women.
Gestational week (IOL)Cesarean section in induced labor (%)Cesarean section in expectant management (%)IOL/Exp (unadjusted OR and 95% CI)IOL/Exp (adjusted OR and 95% CI)
  1. Note: The group with induced labor in a specific gestational week is compared with the group of women waiting for a later either induced or spontaneous labor. Abbreviations: CI, confidence interval; Exp, expectant management; IOL, induction of labor; OR, odds ratio.

  2. *p<0.05.

375.82.32.66 (2.12–3.35)*1.72 (1.35–2.20)*
384.32.31.91 (1.58–2.30)*1.27 (1.04–1.55)*
394.12.51.64 (1.36–1.98)*1.15 (0.95–1.39) 
405.03.21.58 (1.34–1.87)*1.18 (0.99–1.40) 
415.64.71.20 (0.98–1.47) 1.07 (0.87–1.32) 
425.3

After adjusting for potential confounders, the increased risk for nulliparous women only persisted for inductions in gestational week 38 (Table 4), while for the parous women the increased risk was present for inductions in both weeks 37 and 38 (Table 5). From gestational week 39, there was no difference in CS rates between the two groups in both parity cohorts. We did not observe a decrease in the CS rates in induced labor at any gestational age.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

Our main finding was that induction of labor apparently does not convey an increased risk of acute (or emergency) CS when comparing gestational weeks 39, 40 or 41 with a later either induced or spontaneous labor. This is consistent with the findings of Caughey et al. (3), although they found a protective effect on risk of a CS in labor by induction for nulliparas in weeks 38–41.

In general, CS rates in our population were lower than seen in the study by Caughey et al. (3), although the induction rates were the same, 15.1% in their study vs. 15.0% in our data. Our overall CS rate among women with induced labor was 14.5% and among women with spontaneous labor it was 6.1%, compared with 20.6 and 10.9%, respectively, in the study by Caughey et al. The differences in CS rates reflect different guidelines and care practices during pregnancy and labor in the American and Danish labor wards. Our data represent, however, a complete and very large cohort from a well-defined geographical region with different types of labor wards from the smallest up to tertiary centers in university hospitals. Our data extend over a shorter time and thus have less chance of variation in induction policies and strategies for CS. The validity in our study is strengthened by more than 10 times as many deliveries as in the study by Caughey et al.

In a comparison with the same method as proposed by Caughey et al. and also used in the present article, Glantz found no significant differences between the induction group and the group laboring in a later week, among either nulliparous or parous women in any of the gestational weeks (13). This is in agreement with our results. Glantz, however, questioned this design and argued that it biases against expectant management if this group does not include women who will go into labor spontaneously in the same week as the induction week. Therefore, he introduced an at-or-above comparison, comparing IOL with all women laboring in the same week or in a later week. In this comparison, a significantly higher OR for a CS was found in all induced groups through the gestational weeks, except for the combined group and the nulliparous women in week 37 and the parous women in week 41.

The method in this study most closely resembles that of Caughey et al., although we did not control for year of delivery. Unlike in the study by Glantz, we did control for use of epidural analgesia during labor. Glantz excluded women who presented for artificial rupture of membranes, which probably will bias against inductions because this group most probably represents parous women who are ‘easy’ to induce.

Glantz may have a point with his proposed at-or-above comparison, because in the clinical situation some women will go into spontaneous labor during the same gestational week as when the inductions take place. In contrast, a woman already in spontaneous labor should not be included in a comparison between IOL and Exp because there is no expectancy included in the management in that case. Therefore, we find that the at-or-above comparison obscures the idea of the new method and biases against induction. If it is feasible to create comparison groups comprising only half weeks or only a few days, then the criticism by Glantz would be taken into account and the method improved.

Women with elective CS were excluded from the analyses by Caughey et al., Glantz and ourselves. If the decision to carry out an elective CS was made during the included weeks, a CS should in fact have been counted with the group of women treated expectantly. Thus, excluding women with elective CS will bias against the induction group. In a retrospective analysis it is not possible to adjust for this. On the other hand, most elective CSs are decided upon before term, thus minimizing this type of bias in the term and post-term comparisons.

The Danish Birth Registry does not contain information about cervical status. A recent retrospective cohort study of nulliparous women with a favorable cervix found comparable results to those in the present study, showing no difference in risk for a CS in the women with induced labor beyond 39weeks of gestation compared with expectant management (14). In the Cochrane Database review, subgroup analysis by cervical status did not reveal any patterns when comparing induction with expectant management (6).

As with all observational data, there is a risk of introducing bias to the analysis because of the indication for the active management when comparing this with expectant management. We attempted to minimize this bias according to indication for induction, by excluding all women with conditions known to or thought to possibly influence the analysis, although more research is needed about specific indications and the associated risk for CS during an induction (15).

Our findings do not support the suggested differences between USA and non-USA studies (9). However, when examining those data more closely, it turns out that two-thirds of the included women in the group of ‘non-USA’ studies came from a Canadian trial (16). If, instead of country, the overall risk of emergency CS in a study is used for stratification, then this shows that there is no difference between active and expectant management when the overall CS risk is below 20% (p=0.44; OR 0.976; 95% CI 0.718–1.327), while the difference is significantly in favor of induction for the studies with an overall CS risk of 20% or more (p=0.001; OR 1.253; 95% CI 1.085–1.447). This supports the view that clinical factors, such as overall CS rates, may have an impact on the OR for CS in labor when comparing the induced and expectant groups.

The decisive factors that may have an influence on this comparison rely in part on the specific gestational week when the induction takes place. This was supported by the present study. The prevailing CS rate in labor also matters, as well as the induction rate, in combination or separately (17). The exact nature of these relations needs further investigation.

In conclusion, we found no added risk for CS in labor when comparing induction of labor in a specific gestational week from week 39 with a later labor. It is reassuring that several meta-analyses and also registry-based studies are in agreement that induction of labor from week 41 is not associated with an increased risk of CS in labor. Induction of labor provides a sound tool when counseling a woman either going past term or presenting with a problem, even if it is not a severe medical illness. In contrast, the clinical implications for gestational weeks 39 and 40 should be interpreted cautiously because several methodological questions remain unanswered.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References

From the Institute of Public Health, Aarhus University, Aarhus, Denmark, we thank Michael Væth for advice on statistical methods and Ellen Aagaard Nøhr for valuable and constructive suggestions for writing this manuscript. Finn Lauszus from the Department of Obstetrics and Gynecology, Sygehusenheden Vest, Herning, Denmark, is thanked for valuable discussions and comments.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Funding
  8. Acknowledgements
  9. References
  • 1
    Heffner LJ, Elkin E, Fretts RC. Impact of labor induction, gestational age and maternal age on cesarean delivery rates. Obstet Gynecol. 2003;102:28793.
  • 2
    Ehrenthal DB, Jiang X, Strobino DM. Labor induction and the risk of a cesarean delivery among nulliparous women at term. Obstet Gynecol. 2010;116:3542.
  • 3
    Caughey AB, Nicholson JM, Cheng YW, Lyell DJ, Washington AE. Induction of labor and cesarean delivery by gestational age. Am J Obstet Gynecol. 2006;195:7005.
  • 4
    Caughey AB. Measuring perinatal complications: methodologic issues related to gestational age. BMC Pregnancy Childbirth. 2007;7:18.
  • 5
    Joseph KS. Theory of obstetrics: an epidemiologic framework for justifying medically indicated early delivery. BMC Pregnancy Childbirth. 2007;7:4.
  • 6
    Gülmezoglu AM, Crowther CA, Middleton P. Induction of labor for improving birth outcomes for women at or beyond term. Cochrane Database Syst Rev. 2006;(4):CD004945.
  • 7
    Sanchez-Ramos L, Olivier F, Delke I, Kaunitz AM. Labor induction versus expectant management for postterm pregnancies: a systematic review with meta-analysis. Obstet Gynecol. 2003;101:13128.
  • 8
    Wennerholm U-B, Hagberg H, Brorsson B, Bergh C. Induction of labor versus expectant management for post-date pregnancy: is there sufficient evidence for a change in clinical practice? Acta Obstet Gynecol Scand. 2009;88:617.
  • 9
    Caughey AB, Sundaram V, Kaimal AJ, Gienger A, Cheng YW, McDonald KM, et al. Systematic review: elective induction of labor versus expectant management of pregnancy. Ann Intern Med. 2009;151:25263.
  • 10
    Heimstad R, Skogvoll E, Mattsson LA, Johansen OJ, Eik-Nes SH, Salvesen KA. Induction of labor or serial antenatal fetal monitoring in postterm pregnancy: a randomized controlled trial. Obstet Gynecol. 2007;109:60917.
  • 11
    Validering af Landspatientregistret (LPR) mhp obstetrisk forskning og kvalitetssikring [Validation of the National Patient Registry (LPR) concerning obstetrical research and quality assurance]. National Board of Health. Copenhagen , 2003.
  • 12
    Nye tal fra Sundhedsstyrelsen. Årgang 11. Nr 1, januar 2007. Sygehusfordelte kvalitetsindikatorer for lavrisiko førstegangsfødende 2005–2006 (1.–3. kvartal). [National Board of Health. Hospital distributed quality indicators for low risk primipara 2005–2006 (1st–3rd quarter)]. National Board of Health. Copenhagen , 2007.
  • 13
    Glantz CJ. Term labor induction compared with expectant management. Obstet Gynecol. 2010;115:706.
  • 14
    Osmundson SS, Ou-Yang RJ, Grobman WA. Elective induction compared with expectant management in nulliparous women with a favorable cervix. Obstet Gynecol. 2010;116:6015.
  • 15
    Mozurkevich E, Chilimigras J, Koepke E, Keeton K, King V. Indications for induction of labour: a best-evidence review. Br J Obstet Gynaecol. 2009;116:62636.
  • 16
    Hannah ME, Hannah WJ, Hellman J, Hewson S, Milner R, Willan A. Induction of labour as compared with serial antenatal monitoring in post-term pregnancy. A randomised controlled trial. The Canadian Multicenter Post-term Pregnancy Trial Group. N Engl J Med. 1992;326:158792.
  • 17
    Brennan DJ, Robson MS, Murphy M, O’Herlihy C. Comparative analysis of international cesarean delivery rates using 10-group classification identifies significant variation in spontaneous labor. Am J Obstet Gynecol. 2009;201:308.e18.