- Email: [email protected]
Factors influencing the duration of late first and second-trimester termination of pregnancy with prostaglandin derivates
European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
Contents lists available at ScienceDirect
European Journal of Obstetrics & Gynecology and Reproductive Biology journal homepage: www.elsevier.com/locate/ejogrb
Factors influencing the duration of late first and second-trimester termination of pregnancy with prostaglandin derivates Norbert Wagner a,1, Harald Abele a,1, Markus Hoopmann a, Eva-Maria Grischke a, Gunnar Blumenstock b, Diethelm Wallwiener a, Karl Oliver Kagan a,* a b
Department of Obstetrics and Gynaecology, University of Tuebingen, Calwer Strasse 7, 72076 Tuebingen, Germany Department of Medical Biometry, University of Tuebingen, Westbahnhofstrasse 55, 72070 Tuebingen, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 June 2010 Received in revised form 14 September 2010 Accepted 26 October 2010
Objective: To examine the time interval between first induction and fetal expulsion in fetal defect-related termination of pregnancy and to determine the relevant factors that influence this time interval. Study design: Retrospective study involving singleton pregnancies that were terminated due to fetal abnormalities between 2005 and 2009. Induction was done by either 200 mg misoprostol, 1 mg gemeprost or in case of a previous caesarean section by 0.5 mg dinoprostone gel or 1 mg gemeprost. The length of the induction interval between first induction and fetal expulsion was examined according to different maternal and fetal characteristics. Univariate and multivariate logistic regression analyses were used to determine the significant contributors for a delivery within 24 h. Results: One hundred and eighty-four singleton pregnancies fulfilled the inclusion criteria. The median gestational age at first induction was 19.2 weeks of gestation. The median time interval between first induction and expulsion was 18 h (25–75th centile, 11–30 h) and in 59 (32.1%) cases, the induction interval was longer than 24 h. Multivariate logistic regression analysis in the prediction of fetal expulsion within 24 h after first induction indicated significant contribution from gestational age and history of spontaneous delivery without previous caesarean section. Conclusions: The only relevant contributors to an estimation of the time interval were gestational age and previous history of spontaneous delivery. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases. ß 2010 Elsevier Ireland Ltd. All rights reserved.
Keywords: Termination of pregnancy Induction Prostaglandin Duration
1. Introduction Termination of pregnancy (TOP) remains one of the most common procedures in obstetrics and gynaecology, with an estimated 50 million induced abortions each year worldwide [1]. Furthermore, as a result of the introduction of successful largescale antenatal screening programs for the detection of chromosomal abnormalities and major structural fetal malformations, first and second-trimester abortions are on the rise [2,3]. TOP is performed either by surgical evacuation or medically by prostaglandins. In general, these are misoprostol, gemeprost and dinoprostone, which are widely used and which have been proved to be safe and efficient for cervical ripening, induction of labour and termination of pregnancy [4]. In contrast to surgical evacuation, medically induced termination offers the opportunity to perform a post-mortem autopsy
* Corresponding author. Tel.: +49 7071 29 84807; fax: +49 7071 29 5619. E-mail address: [email protected] (K.O. Kagan). 1 Shared authorship. 0301-2115/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejogrb.2010.10.019
when further information is required. In addition, it was shown that maternal morbidity rates are lower after medically induced terminations. In a randomised prospective study by Niinima¨ki et al., surgical and medical terminations were compared in 98 miscarriages at 11 weeks of gestation. There were significantly more infections in the surgical treatment group [5]. Bartlett et al. studied the maternal mortality rate after legally induced abortions from 1988 to 1997 in the United States. The overall mortality rate was 0.7 per 100,000. The strongest risk factor was gestational age at the time of termination with an increase in the mortality rate of 38% by additional week of gestation from eight weeks. There were no deaths associated with medical abortion procedures [6]. Surgical termination, however, is faster and scheduling is easier. Therefore, most patients would request surgical rather than medically induced termination after a previous termination of pregnancy [5]. In particular, the time from first induction until expulsion is difficult to determine. In a study by Dickinson et al. gestational age, maternal age and parity were identified as significant covariates influencing the time between first induction and expulsion when using vaginal misoprostol [7].
76
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
In this study we used multiple regression analysis to determine the relevant factors that influence the time interval between first induction and fetal expulsion in medically induced termination of pregnancy. 2. Materials and methods This retrospective study consists of all singleton pregnancies terminated with prostaglandins between 2005 and 2009 in our hospital. The pregnancies were terminated between 11 and 24 weeks of gestation on the basis of present or expected severe maternal distress due to fetal abnormalities (German criminal law §218a, second paragraph). Social terminations were excluded from this study. In each case maternal and pregnancy characteristics, the induction interval and the applied prostaglandin were recorded into a database. The induction interval as outcome parameter was defined as the time between first application of the prostaglandin and the end of the fetal expulsion. Three different prostaglandin protocols were used for termination. For those patients without previous caesarean section either misoprostol (Cytotec1) or gemeprost (Cergem1) were used. Gemeprost (1 mg) was placed into the vagina whereas misoprostol was given vaginally (200 mg) and orally (200 mg) at the same time. Induction was repeated every 4–6 h until relevant contractions were recorded. Patients with a previous history of a spontaneous delivery without previous caesarean section received either misoprostol or gemeprost. For patients with a previous caesarean section 1 mg gemeprost or 0.5 mg dinoprostone gel (Prepidil1) was used. Both prostaglandins were inserted vaginally every 4–6 h until relevant contractions were recorded. Throughout the induction the patients were in the hospital. The decision between the misoprostol and the gemeprost protocols for patients without caesarean section and between the gemeprost and the dinoprostone protocols for patients with a previous caesarean section was made by the gynaecologist in charge. For this study, only those cases were considered where the type of prostaglandin was not changed between the first induction and delivery. Those few cases where different prostaglandins were given during the induction or where the time interval between two applications was different from 4 to 6 h were excluded. All patients were consented about the off-label-use of misoprostol for induction of labour and agreed by signature. 2.1. Statistical analysis In each case, maternal age (years), BMI (kg/m2), gravidity (n), parity (n), history of caesarean section (yes/no), history of [()TD$FIG]
Table 1 Maternal and pregnancy characteristics of study population. Maternal and pregnancy characteristics
Median
25th–75th centile
Maternal age (years) BMI (kg/m2) Gravida (n) Para (n) Gestational age at first IOL (weeks)
32 24.5 2 1 19.2
28–37 22.0–27.1 1–3 0–1 16.2–21.7
Table 2 Time interval between first induction and expulsion according to maternal and pregnancy characteristics. Maternal and pregnancy characteristics. Mode of previous deliveries No. previous delivery At least one caesarean section At least one spontaneous delivery and no caesarean section Kruskal–Wallis test Prostaglandin protocol Gemeprost Misoprostol Dinoprostone Kruskal–Wallis test
87 (47.3) 26 (14.1) 71 (38.6)
Time interval (h) Median and IQR 22 (18–31) 32 (17–67) 11 (9–16) p < 0.0001*
48 (26.1) 125 (67.9) 11 (6.0)
20 (13–31) 15 (10–25) 64 (41–76) p < 0.0001**
Maternal age Maternal age below 30 Maternal age between 30 and 34 Maternal age 35 or older Kruskal–Wallis test
65 (35.3) 47 (25.5) 72 (39.1)
18 (11–30) 20 (10–33) 15 (11–27) p = 0.617
Maternal body mass index BMI less than 20.0 kg/m2 BMI between 20.0 and 23.9 kg/m2 BMI between 24.0 and 30.9 kg/m2 BMI 31 kg/m2 or more Kruskal–Wallis test
13 66 86 19
(7.1) (35.9) (46.7) (10.3)
17 (14–22) 15 (10–25) 18 (12–32) 28 (18–33) p = 0.086
22 (12) 78 (42.4)
14 (10–20) 17 (11–26)
77 (41.8)
21 (12–34)
7 (3.8)
31 (12–39) p = 0.074
Gestational age at first IOL Gestational age less than 14 + 0 Gestational age between 14 + 0 and 19 + 6 Gestational age between 20 + 0 and 23 + 6 Gestational age 24 + 0 or higher Kruskal–Wallis test Amniotic fluid Normal Polyhydramnios Oligo-/anhydramnios Kruskal–Wallis test Fetal defect leading to TOP Aneuploidy Major cardiac defect Non-immune fetal hydrops Neural tube defect Neuromuscular disease or skeletal dysplasia Renal malformation Malformation of the central nervous system Others Kruskal–Wallis test
Fig. 1. Proportion of pregnancies where TOP is not completed according to the time interval between first induction and fetal expulsion.
n (%)
131 (71.2) 21 (11.4) 32 (17.4)
66 8 14 19 27
(35.9) (4.3) (7.6) (10.3) (14.7)
18 (11–30) 15 (11–30) 20 (11–29.5) p = 0.949
18 (11–29) 16.5 (13–28) 26.5 (17–41) 18 (13–30) 14 (11–30)
16 (8.7) 18 (9.8)
21.5 (14–31.5) 23 (10–37)
16 (8.7)
12 (9.5–18) p = 0.202
IQR = interquartile range. * Mode of previous deliveries: Mann–Whitney U test: Nullipara vs. at least one delivery: p < 0.0001, Nullipara vs. at least one spont. delivery and no caesarean section: p = 0.065, at least one delivery vs. at least one spont. delivery and no caesarean section: p < 0.0001. ** Prostaglandin protocol: Mann–Whitney U test: gemeprost vs. misoprostol: p = 0.03, gemeprost vs. dinoprostone: p < 0.0001, misoprostol vs. dinoprostone: p < 0.0001.
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
77
Table 3 Univariate and multiple regression analysis in the prediction of delivery within 24 h after induction of labour. Maternal and pregnancy characteristics
Univariate regression analysis
Multivariate regression analysis
OR
p
95% CI
OR
p
95% CI
Maternal age (years) Maternal body mass index (kg/m2) Gestational age (weeks)
1.044 0.019 0.868
0.125 0.019 0.002
0.988–1.103 0.854–0.986 0.793–0.951
0.93 0.89
0.076 0.03
0.868–1.008 0.808–0.989
Amniotic fluid Normal Polyhydramnios Oligohydramnios
1 1.222 1.075
0.699 0.865
0.443–3.369 0.468–2.470
Prostaglandin protocol Gemeprost Misoprostol Dinoprostone
1 0.627 0.034
0.201 0.0001
0.307–1.283 0.004–0.279
1.13 0.09
0.063 0.788
0.007–1.141 0.451–2.852
Previous history No previous delivery At least one caesarean section At least one spontaneous delivery and no caesarean section
1 0.344 5.128
0.021 0.0001
0.139–0.850 2.290–11.485
0.703 5.27
0.609 0.0001
0.183–2.706 2.273–12.222
Fetal defect leading to TOP Aneuploidy Major cardiac defect Non-immune fetal hydrops Neural tube defect Neuromuscular disease or skeletal dysplasia Renal malformation Malformation of the central nervous system Others
1 1.213 0.404 1.132 0.960 0.520 0.404 2.830
0.823 0.131 0.833 0.935 0.253 0.096 0.195
0.225–6.551 0.125–1.31 0.358–3.581 0.359–2.566 0.169–1.597 0.139–1.175 0.586–13.663
spontaneous delivery (yes/no), gestational age at first induction of labour (weeks), amniotic fluid volume (normal, polyhydramnios, oligohydramnios), induction courses (n), applied prostaglandin and induction interval (h) was recorded. Fetal defects were classified into eight groups (aneuploidy, major cardiac defect, nonimmune fetal hydrops, neural tube defects, neuromuscular disease or skeletal dysplasia, renal malformation, malformation of the central nervous system, other). For each factor the median time interval and the 25–75th centile are given. Differences were tested by Kruskal–Wallis and consecutive Mann–Whitney U tests. Univariate and multivariate logistic regression analyses were used to determine the significant contributors for a delivery within 24 h. A p-value of less than 0.05 was considered to be significant.
and expulsion was 22 h for nulliparous patients. For those with at least one caesarean section, it was 32 h and for those with at least one spontaneous delivery and no caesarean section, it was 11 h. The median induction length was 20, 15 and 64 h for gemeprost, misoprostol and dinoprostone, respectively. Table 3 shows the result of the univariate and multivariate logistic regression analysis in predicting of expulsion within 24 h after first induction. In the multiple regression analysis, only gestational age and history of spontaneous delivery without previous caesarean section were significant contributors for predicting an induction interval of 24 h or less.
[()TD$FIG]
3. Results One hundred and eighty-four singleton pregnancies fulfilled the inclusion criteria. Of these, 125 (67.9%) were induced according to the misoprostol protocol, 48 (26.1%) according to the gemeprost protocol and 11 (6.0%) according to the dinoprostone protocol, respectively. The maternal characteristics of the study population are shown in Table 1. In 22 (12.0%) cases, the gestational age was less than 14 + 0 weeks of gestation. The gestational age was between 14 + 0 and 19 + 6 weeks of gestation in 78 (42.4%) cases, between 20 + 0 and 23 + 6 weeks in 77 (41.8%) cases and above 24 + 0 weeks in 7 (3.8%) cases, respectively. The median gestational age at first application of the prostaglandin was 19.2 weeks of gestation. The median time interval between first induction and expulsion was 18 h (25–75th centile, 11–30 h). In 59 (32.1%) cases, the induction interval was longer than 24 h and in 18 (9.8%) pregnancies, the induction took 36 h or more (Fig. 1). The median induction length according to different maternal pregnancy and management characteristics is shown in Table 2. Significant differences were found for maternal history and applied prostaglandin. The median time interval between first induction
Fig. 2. Detection and false positive rates in the prediction of delivery within 24 h after first induction of labour.
78
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
Fig. 2 shows the respective ROC curve (area under the curve 0.761 (95% CI 0.687–0.834, p < 0.0001). For a detection rate of 70% the false positive rate was 23.7%. This means that the algorithm correctly identified 70% of those pregnancies where the induction length was 24 h or less. In 23.7% of the pregnancies where the induction length was more than 24 h an earlier delivery was anticipated. In practical terms, in our study delivery within 24 h was anticipated in 101 cases (70% of the 125 cases with an induction length of 24 h or less and 23.7% of 59 cases with induction length of more than 24 h). In 88 cases delivery was within 24 h indicating a positive predictive value of 87.1%. 4. Comments In this retrospective study we have used multiple regression analysis to determine the relevant covariates that influence the time interval between first induction and fetal expulsion in termination of pregnancy. Parameters with significant influence were gestational age at the time of induction, previous obstetric history and applied prostaglandin. The time interval was prolonged with increasing gestational age, previous caesarean section and nulliparity compared to previous spontaneous delivery and with dinoprostone and gemeprost compared to misoprostol. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases. Our findings are consistent with a previous study of Dickinson and Doherty [7]. In that large, prospective study involving 1066 terminations of pregnancy with 400 mg of vaginally applied misoprostol, maternal age, gestational age and previous obstetric history contributed significantly to the time interval between first induction and fetal expulsion [7]. Similarly to our study, they observed a prolonged induction time with increasing gestational age and attributed this to the need for greater cervical dilatation due to the increasing fetal size. Grigsby et al. demonstrated that with advancing gestation there is a reduction of some prostaglandin receptors in the maternal myometrium [8]. It could therefore be argued that the prolonged induction interval with gestational age could be due a reduction of the receptivity of the uterus to prostaglandins [8]. In contrast, Lo et al., who examined 280 terminations of pregnancy between 13 and 23 weeks by vaginal application of 400 mg misoprostol, could not find a relationship between gestational age and the duration of pregnancy termination [9]. Dependency on the previous obstetric history was also observed by Jannet et al. In their study, which consisted of 106 terminations of pregnancy with a combination of 600 mg mifepristone and 400 mg misoprostol, the time interval between first induction and fetal expulsion was significantly shorter in multiparous than in primiparous women [10]. This can be attributed to the faster ripening of the cervix after a previous spontaneous delivery. A longer time interval in women with a previous caesarean section in our study could be due to the absence of a previous spontaneous delivery and to the fact that in these cases dinoprostone or gemeprost was used. Compared to misoprostol, both prostaglandins increased the time interval significantly in the univariate regression, but in the multivariate regression analysis this no longer contributed significantly.
In our study the fetal anomaly leading to the termination of pregnancy did not affect the time interval between first induction and fetal expulsion. This is in contrast to Nesbitt et al., who reported a longer induction-delivery interval in cases of a fetal neural tube defect and/or hydrocephalus. Their finding was attributed to a defect-related decrease in fetal corticotrophinreleasing hormone [11]. Alternatively to induction with misopristol only, a combination of mifepristone and misoprostol can be used for termination of midtrimester pregnancies. Ashok et al. reported on a study of 1002 women between 13 and 21 weeks of gestation who underwent termination of pregnancy by administration of 600 mg mifepristone followed by 400 mg misoprostol [12]. The antiprogesterone mifepristone competes with progesterone at the progesterone receptor level and sensitizes the myometrial cells to the action of prostaglandins. In their study the median time interval between the vaginal application of misoprostol and fetal expulsion was 6.3 h, with 97.1% of the cases being within 15 h, which was substantially shorter than in our study [12]. In conclusion, in a multiple regression analysis the only relevant contributors to the time interval between first induction and fetal expulsion were gestational age and previous history of spontaneous delivery. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases.
References [1] Segal ST, LaGuardia KD. Termination of pregnancy—a global view. Baillie`re’s Clin Obstet Gynaecol 1990;4:235–47. [2] Liu S, Joseph KS, Kramer MS, et al. Relationship of prenatal diagnosis and pregnancy termination to overall infant mortality in Canada. JAMA 2002;287:1561–7. [3] Lalitkumar S, Bygdeman M, Gemzell-Danielsson K. Mid-trimester induced abortion: a review. Hum Reprod Update 2007;13:37–52. [4] Hertzen V, Piaggio H, Wojdyla GD, et al. Comparison of vaginal and sublingual misoprostol for second trimester abortion: randomized controlled equivalence trial. Hum Reprod 2009;24:106–12. [5] Niinima¨ki M, Jouppila P, Martikainen H, et al. A randomized study comparing efficacy and patient satisfaction in medical or surgical treatment of miscarriage. Fertil Steril 2006;86:367–72. [6] Bartlett LA, Berg CJ, Shulman HB, et al. Risk factors for legal induced abortionrelated mortality in the United States. Obstet Gynecol 2004;103:729–37. [7] Dickinson JE, Doherty DA. Factors influencing the duration of pregnancy termination with vaginal misoprostol for fetal abnormality. Prenat Diagn 2009;29:520–4. [8] Grigsby PL, Sooranna SR, Adu-Amankwa B, et al. Regional expression of prostaglandin E2 and F2alpha receptors in human myometrium, amnion, and choriodecidua with advancing gestation and labor. Biol Reprod 2006;75(2):297–305. [9] Lo TK, Lau WL, Lai FK, et al. The effect of gestational age on the outcome of second-trimester termination of pregnancies for foetal abnormalities. Prenat Diagn 2008;28:508–11. [10] Jannet D, Aflak N, Abankwa A, et al. Termination of 2nd 3rd trimester pregnancies with mifepristone and misoprostol. Eur J Obstet Gynecol Reprod Biol 1996;70:159–63. [11] Nesbitt D, Giles W. Prolonged induction to delivery time in termination of pregnancy using 16, 16-dimethyl-PGE1-methyl ester (gemeprost) for fetuses with a neural tube defect or hydrocephalus induction to delivery time in termination of pregnancy using 16, 16-dimethyl-PGE1-methyl ester (gemeprost) for fetuses with a neural tube defect or hydrocephalus. Aust N Z J Obstet Gynaecol 1996;36:300–3. [12] Ashok PW, Templeton A, Wagaarachchi PT, et al. Midtrimester medical termination of pregnancy: a review of 1002 consecutive cases. Contraception 2004;69:51–8.
Contents lists available at ScienceDirect
European Journal of Obstetrics & Gynecology and Reproductive Biology journal homepage: www.elsevier.com/locate/ejogrb
Factors influencing the duration of late first and second-trimester termination of pregnancy with prostaglandin derivates Norbert Wagner a,1, Harald Abele a,1, Markus Hoopmann a, Eva-Maria Grischke a, Gunnar Blumenstock b, Diethelm Wallwiener a, Karl Oliver Kagan a,* a b
Department of Obstetrics and Gynaecology, University of Tuebingen, Calwer Strasse 7, 72076 Tuebingen, Germany Department of Medical Biometry, University of Tuebingen, Westbahnhofstrasse 55, 72070 Tuebingen, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 June 2010 Received in revised form 14 September 2010 Accepted 26 October 2010
Objective: To examine the time interval between first induction and fetal expulsion in fetal defect-related termination of pregnancy and to determine the relevant factors that influence this time interval. Study design: Retrospective study involving singleton pregnancies that were terminated due to fetal abnormalities between 2005 and 2009. Induction was done by either 200 mg misoprostol, 1 mg gemeprost or in case of a previous caesarean section by 0.5 mg dinoprostone gel or 1 mg gemeprost. The length of the induction interval between first induction and fetal expulsion was examined according to different maternal and fetal characteristics. Univariate and multivariate logistic regression analyses were used to determine the significant contributors for a delivery within 24 h. Results: One hundred and eighty-four singleton pregnancies fulfilled the inclusion criteria. The median gestational age at first induction was 19.2 weeks of gestation. The median time interval between first induction and expulsion was 18 h (25–75th centile, 11–30 h) and in 59 (32.1%) cases, the induction interval was longer than 24 h. Multivariate logistic regression analysis in the prediction of fetal expulsion within 24 h after first induction indicated significant contribution from gestational age and history of spontaneous delivery without previous caesarean section. Conclusions: The only relevant contributors to an estimation of the time interval were gestational age and previous history of spontaneous delivery. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases. ß 2010 Elsevier Ireland Ltd. All rights reserved.
Keywords: Termination of pregnancy Induction Prostaglandin Duration
1. Introduction Termination of pregnancy (TOP) remains one of the most common procedures in obstetrics and gynaecology, with an estimated 50 million induced abortions each year worldwide [1]. Furthermore, as a result of the introduction of successful largescale antenatal screening programs for the detection of chromosomal abnormalities and major structural fetal malformations, first and second-trimester abortions are on the rise [2,3]. TOP is performed either by surgical evacuation or medically by prostaglandins. In general, these are misoprostol, gemeprost and dinoprostone, which are widely used and which have been proved to be safe and efficient for cervical ripening, induction of labour and termination of pregnancy [4]. In contrast to surgical evacuation, medically induced termination offers the opportunity to perform a post-mortem autopsy
* Corresponding author. Tel.: +49 7071 29 84807; fax: +49 7071 29 5619. E-mail address: [email protected] (K.O. Kagan). 1 Shared authorship. 0301-2115/$ – see front matter ß 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ejogrb.2010.10.019
when further information is required. In addition, it was shown that maternal morbidity rates are lower after medically induced terminations. In a randomised prospective study by Niinima¨ki et al., surgical and medical terminations were compared in 98 miscarriages at 11 weeks of gestation. There were significantly more infections in the surgical treatment group [5]. Bartlett et al. studied the maternal mortality rate after legally induced abortions from 1988 to 1997 in the United States. The overall mortality rate was 0.7 per 100,000. The strongest risk factor was gestational age at the time of termination with an increase in the mortality rate of 38% by additional week of gestation from eight weeks. There were no deaths associated with medical abortion procedures [6]. Surgical termination, however, is faster and scheduling is easier. Therefore, most patients would request surgical rather than medically induced termination after a previous termination of pregnancy [5]. In particular, the time from first induction until expulsion is difficult to determine. In a study by Dickinson et al. gestational age, maternal age and parity were identified as significant covariates influencing the time between first induction and expulsion when using vaginal misoprostol [7].
76
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
In this study we used multiple regression analysis to determine the relevant factors that influence the time interval between first induction and fetal expulsion in medically induced termination of pregnancy. 2. Materials and methods This retrospective study consists of all singleton pregnancies terminated with prostaglandins between 2005 and 2009 in our hospital. The pregnancies were terminated between 11 and 24 weeks of gestation on the basis of present or expected severe maternal distress due to fetal abnormalities (German criminal law §218a, second paragraph). Social terminations were excluded from this study. In each case maternal and pregnancy characteristics, the induction interval and the applied prostaglandin were recorded into a database. The induction interval as outcome parameter was defined as the time between first application of the prostaglandin and the end of the fetal expulsion. Three different prostaglandin protocols were used for termination. For those patients without previous caesarean section either misoprostol (Cytotec1) or gemeprost (Cergem1) were used. Gemeprost (1 mg) was placed into the vagina whereas misoprostol was given vaginally (200 mg) and orally (200 mg) at the same time. Induction was repeated every 4–6 h until relevant contractions were recorded. Patients with a previous history of a spontaneous delivery without previous caesarean section received either misoprostol or gemeprost. For patients with a previous caesarean section 1 mg gemeprost or 0.5 mg dinoprostone gel (Prepidil1) was used. Both prostaglandins were inserted vaginally every 4–6 h until relevant contractions were recorded. Throughout the induction the patients were in the hospital. The decision between the misoprostol and the gemeprost protocols for patients without caesarean section and between the gemeprost and the dinoprostone protocols for patients with a previous caesarean section was made by the gynaecologist in charge. For this study, only those cases were considered where the type of prostaglandin was not changed between the first induction and delivery. Those few cases where different prostaglandins were given during the induction or where the time interval between two applications was different from 4 to 6 h were excluded. All patients were consented about the off-label-use of misoprostol for induction of labour and agreed by signature. 2.1. Statistical analysis In each case, maternal age (years), BMI (kg/m2), gravidity (n), parity (n), history of caesarean section (yes/no), history of [()TD$FIG]
Table 1 Maternal and pregnancy characteristics of study population. Maternal and pregnancy characteristics
Median
25th–75th centile
Maternal age (years) BMI (kg/m2) Gravida (n) Para (n) Gestational age at first IOL (weeks)
32 24.5 2 1 19.2
28–37 22.0–27.1 1–3 0–1 16.2–21.7
Table 2 Time interval between first induction and expulsion according to maternal and pregnancy characteristics. Maternal and pregnancy characteristics. Mode of previous deliveries No. previous delivery At least one caesarean section At least one spontaneous delivery and no caesarean section Kruskal–Wallis test Prostaglandin protocol Gemeprost Misoprostol Dinoprostone Kruskal–Wallis test
87 (47.3) 26 (14.1) 71 (38.6)
Time interval (h) Median and IQR 22 (18–31) 32 (17–67) 11 (9–16) p < 0.0001*
48 (26.1) 125 (67.9) 11 (6.0)
20 (13–31) 15 (10–25) 64 (41–76) p < 0.0001**
Maternal age Maternal age below 30 Maternal age between 30 and 34 Maternal age 35 or older Kruskal–Wallis test
65 (35.3) 47 (25.5) 72 (39.1)
18 (11–30) 20 (10–33) 15 (11–27) p = 0.617
Maternal body mass index BMI less than 20.0 kg/m2 BMI between 20.0 and 23.9 kg/m2 BMI between 24.0 and 30.9 kg/m2 BMI 31 kg/m2 or more Kruskal–Wallis test
13 66 86 19
(7.1) (35.9) (46.7) (10.3)
17 (14–22) 15 (10–25) 18 (12–32) 28 (18–33) p = 0.086
22 (12) 78 (42.4)
14 (10–20) 17 (11–26)
77 (41.8)
21 (12–34)
7 (3.8)
31 (12–39) p = 0.074
Gestational age at first IOL Gestational age less than 14 + 0 Gestational age between 14 + 0 and 19 + 6 Gestational age between 20 + 0 and 23 + 6 Gestational age 24 + 0 or higher Kruskal–Wallis test Amniotic fluid Normal Polyhydramnios Oligo-/anhydramnios Kruskal–Wallis test Fetal defect leading to TOP Aneuploidy Major cardiac defect Non-immune fetal hydrops Neural tube defect Neuromuscular disease or skeletal dysplasia Renal malformation Malformation of the central nervous system Others Kruskal–Wallis test
Fig. 1. Proportion of pregnancies where TOP is not completed according to the time interval between first induction and fetal expulsion.
n (%)
131 (71.2) 21 (11.4) 32 (17.4)
66 8 14 19 27
(35.9) (4.3) (7.6) (10.3) (14.7)
18 (11–30) 15 (11–30) 20 (11–29.5) p = 0.949
18 (11–29) 16.5 (13–28) 26.5 (17–41) 18 (13–30) 14 (11–30)
16 (8.7) 18 (9.8)
21.5 (14–31.5) 23 (10–37)
16 (8.7)
12 (9.5–18) p = 0.202
IQR = interquartile range. * Mode of previous deliveries: Mann–Whitney U test: Nullipara vs. at least one delivery: p < 0.0001, Nullipara vs. at least one spont. delivery and no caesarean section: p = 0.065, at least one delivery vs. at least one spont. delivery and no caesarean section: p < 0.0001. ** Prostaglandin protocol: Mann–Whitney U test: gemeprost vs. misoprostol: p = 0.03, gemeprost vs. dinoprostone: p < 0.0001, misoprostol vs. dinoprostone: p < 0.0001.
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
77
Table 3 Univariate and multiple regression analysis in the prediction of delivery within 24 h after induction of labour. Maternal and pregnancy characteristics
Univariate regression analysis
Multivariate regression analysis
OR
p
95% CI
OR
p
95% CI
Maternal age (years) Maternal body mass index (kg/m2) Gestational age (weeks)
1.044 0.019 0.868
0.125 0.019 0.002
0.988–1.103 0.854–0.986 0.793–0.951
0.93 0.89
0.076 0.03
0.868–1.008 0.808–0.989
Amniotic fluid Normal Polyhydramnios Oligohydramnios
1 1.222 1.075
0.699 0.865
0.443–3.369 0.468–2.470
Prostaglandin protocol Gemeprost Misoprostol Dinoprostone
1 0.627 0.034
0.201 0.0001
0.307–1.283 0.004–0.279
1.13 0.09
0.063 0.788
0.007–1.141 0.451–2.852
Previous history No previous delivery At least one caesarean section At least one spontaneous delivery and no caesarean section
1 0.344 5.128
0.021 0.0001
0.139–0.850 2.290–11.485
0.703 5.27
0.609 0.0001
0.183–2.706 2.273–12.222
Fetal defect leading to TOP Aneuploidy Major cardiac defect Non-immune fetal hydrops Neural tube defect Neuromuscular disease or skeletal dysplasia Renal malformation Malformation of the central nervous system Others
1 1.213 0.404 1.132 0.960 0.520 0.404 2.830
0.823 0.131 0.833 0.935 0.253 0.096 0.195
0.225–6.551 0.125–1.31 0.358–3.581 0.359–2.566 0.169–1.597 0.139–1.175 0.586–13.663
spontaneous delivery (yes/no), gestational age at first induction of labour (weeks), amniotic fluid volume (normal, polyhydramnios, oligohydramnios), induction courses (n), applied prostaglandin and induction interval (h) was recorded. Fetal defects were classified into eight groups (aneuploidy, major cardiac defect, nonimmune fetal hydrops, neural tube defects, neuromuscular disease or skeletal dysplasia, renal malformation, malformation of the central nervous system, other). For each factor the median time interval and the 25–75th centile are given. Differences were tested by Kruskal–Wallis and consecutive Mann–Whitney U tests. Univariate and multivariate logistic regression analyses were used to determine the significant contributors for a delivery within 24 h. A p-value of less than 0.05 was considered to be significant.
and expulsion was 22 h for nulliparous patients. For those with at least one caesarean section, it was 32 h and for those with at least one spontaneous delivery and no caesarean section, it was 11 h. The median induction length was 20, 15 and 64 h for gemeprost, misoprostol and dinoprostone, respectively. Table 3 shows the result of the univariate and multivariate logistic regression analysis in predicting of expulsion within 24 h after first induction. In the multiple regression analysis, only gestational age and history of spontaneous delivery without previous caesarean section were significant contributors for predicting an induction interval of 24 h or less.
[()TD$FIG]
3. Results One hundred and eighty-four singleton pregnancies fulfilled the inclusion criteria. Of these, 125 (67.9%) were induced according to the misoprostol protocol, 48 (26.1%) according to the gemeprost protocol and 11 (6.0%) according to the dinoprostone protocol, respectively. The maternal characteristics of the study population are shown in Table 1. In 22 (12.0%) cases, the gestational age was less than 14 + 0 weeks of gestation. The gestational age was between 14 + 0 and 19 + 6 weeks of gestation in 78 (42.4%) cases, between 20 + 0 and 23 + 6 weeks in 77 (41.8%) cases and above 24 + 0 weeks in 7 (3.8%) cases, respectively. The median gestational age at first application of the prostaglandin was 19.2 weeks of gestation. The median time interval between first induction and expulsion was 18 h (25–75th centile, 11–30 h). In 59 (32.1%) cases, the induction interval was longer than 24 h and in 18 (9.8%) pregnancies, the induction took 36 h or more (Fig. 1). The median induction length according to different maternal pregnancy and management characteristics is shown in Table 2. Significant differences were found for maternal history and applied prostaglandin. The median time interval between first induction
Fig. 2. Detection and false positive rates in the prediction of delivery within 24 h after first induction of labour.
78
N. Wagner et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 155 (2011) 75–78
Fig. 2 shows the respective ROC curve (area under the curve 0.761 (95% CI 0.687–0.834, p < 0.0001). For a detection rate of 70% the false positive rate was 23.7%. This means that the algorithm correctly identified 70% of those pregnancies where the induction length was 24 h or less. In 23.7% of the pregnancies where the induction length was more than 24 h an earlier delivery was anticipated. In practical terms, in our study delivery within 24 h was anticipated in 101 cases (70% of the 125 cases with an induction length of 24 h or less and 23.7% of 59 cases with induction length of more than 24 h). In 88 cases delivery was within 24 h indicating a positive predictive value of 87.1%. 4. Comments In this retrospective study we have used multiple regression analysis to determine the relevant covariates that influence the time interval between first induction and fetal expulsion in termination of pregnancy. Parameters with significant influence were gestational age at the time of induction, previous obstetric history and applied prostaglandin. The time interval was prolonged with increasing gestational age, previous caesarean section and nulliparity compared to previous spontaneous delivery and with dinoprostone and gemeprost compared to misoprostol. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases. Our findings are consistent with a previous study of Dickinson and Doherty [7]. In that large, prospective study involving 1066 terminations of pregnancy with 400 mg of vaginally applied misoprostol, maternal age, gestational age and previous obstetric history contributed significantly to the time interval between first induction and fetal expulsion [7]. Similarly to our study, they observed a prolonged induction time with increasing gestational age and attributed this to the need for greater cervical dilatation due to the increasing fetal size. Grigsby et al. demonstrated that with advancing gestation there is a reduction of some prostaglandin receptors in the maternal myometrium [8]. It could therefore be argued that the prolonged induction interval with gestational age could be due a reduction of the receptivity of the uterus to prostaglandins [8]. In contrast, Lo et al., who examined 280 terminations of pregnancy between 13 and 23 weeks by vaginal application of 400 mg misoprostol, could not find a relationship between gestational age and the duration of pregnancy termination [9]. Dependency on the previous obstetric history was also observed by Jannet et al. In their study, which consisted of 106 terminations of pregnancy with a combination of 600 mg mifepristone and 400 mg misoprostol, the time interval between first induction and fetal expulsion was significantly shorter in multiparous than in primiparous women [10]. This can be attributed to the faster ripening of the cervix after a previous spontaneous delivery. A longer time interval in women with a previous caesarean section in our study could be due to the absence of a previous spontaneous delivery and to the fact that in these cases dinoprostone or gemeprost was used. Compared to misoprostol, both prostaglandins increased the time interval significantly in the univariate regression, but in the multivariate regression analysis this no longer contributed significantly.
In our study the fetal anomaly leading to the termination of pregnancy did not affect the time interval between first induction and fetal expulsion. This is in contrast to Nesbitt et al., who reported a longer induction-delivery interval in cases of a fetal neural tube defect and/or hydrocephalus. Their finding was attributed to a defect-related decrease in fetal corticotrophinreleasing hormone [11]. Alternatively to induction with misopristol only, a combination of mifepristone and misoprostol can be used for termination of midtrimester pregnancies. Ashok et al. reported on a study of 1002 women between 13 and 21 weeks of gestation who underwent termination of pregnancy by administration of 600 mg mifepristone followed by 400 mg misoprostol [12]. The antiprogesterone mifepristone competes with progesterone at the progesterone receptor level and sensitizes the myometrial cells to the action of prostaglandins. In their study the median time interval between the vaginal application of misoprostol and fetal expulsion was 6.3 h, with 97.1% of the cases being within 15 h, which was substantially shorter than in our study [12]. In conclusion, in a multiple regression analysis the only relevant contributors to the time interval between first induction and fetal expulsion were gestational age and previous history of spontaneous delivery. By combining this information it was possible to correctly anticipate fetal expulsion within 24 h after first induction in 90% of the cases.
References [1] Segal ST, LaGuardia KD. Termination of pregnancy—a global view. Baillie`re’s Clin Obstet Gynaecol 1990;4:235–47. [2] Liu S, Joseph KS, Kramer MS, et al. Relationship of prenatal diagnosis and pregnancy termination to overall infant mortality in Canada. JAMA 2002;287:1561–7. [3] Lalitkumar S, Bygdeman M, Gemzell-Danielsson K. Mid-trimester induced abortion: a review. Hum Reprod Update 2007;13:37–52. [4] Hertzen V, Piaggio H, Wojdyla GD, et al. Comparison of vaginal and sublingual misoprostol for second trimester abortion: randomized controlled equivalence trial. Hum Reprod 2009;24:106–12. [5] Niinima¨ki M, Jouppila P, Martikainen H, et al. A randomized study comparing efficacy and patient satisfaction in medical or surgical treatment of miscarriage. Fertil Steril 2006;86:367–72. [6] Bartlett LA, Berg CJ, Shulman HB, et al. Risk factors for legal induced abortionrelated mortality in the United States. Obstet Gynecol 2004;103:729–37. [7] Dickinson JE, Doherty DA. Factors influencing the duration of pregnancy termination with vaginal misoprostol for fetal abnormality. Prenat Diagn 2009;29:520–4. [8] Grigsby PL, Sooranna SR, Adu-Amankwa B, et al. Regional expression of prostaglandin E2 and F2alpha receptors in human myometrium, amnion, and choriodecidua with advancing gestation and labor. Biol Reprod 2006;75(2):297–305. [9] Lo TK, Lau WL, Lai FK, et al. The effect of gestational age on the outcome of second-trimester termination of pregnancies for foetal abnormalities. Prenat Diagn 2008;28:508–11. [10] Jannet D, Aflak N, Abankwa A, et al. Termination of 2nd 3rd trimester pregnancies with mifepristone and misoprostol. Eur J Obstet Gynecol Reprod Biol 1996;70:159–63. [11] Nesbitt D, Giles W. Prolonged induction to delivery time in termination of pregnancy using 16, 16-dimethyl-PGE1-methyl ester (gemeprost) for fetuses with a neural tube defect or hydrocephalus induction to delivery time in termination of pregnancy using 16, 16-dimethyl-PGE1-methyl ester (gemeprost) for fetuses with a neural tube defect or hydrocephalus. Aust N Z J Obstet Gynaecol 1996;36:300–3. [12] Ashok PW, Templeton A, Wagaarachchi PT, et al. Midtrimester medical termination of pregnancy: a review of 1002 consecutive cases. Contraception 2004;69:51–8.