European Journal of Obstetrics & Gynecology and Reproductive Biology, 41(1992) 17-23
17
0 1992 Elsevier Science Publishers B.V. All rights reserved 002&X-2243/92/$05.00 EUROBS 01410
Predictive value of pelvic scores for induction of labor by local PGE, Jens Lyndrup a,b,Jesper Legarth a,b,Tom Weber b, Carsten Nickelsen b and Else Guldbaek b Departments of Gynecology and Obstetrics, ’ Herlev University Hospital and ’ Hvidovre University Hospital, Copenhagen, Denmark
Accepted for publication 15 July 1992
Summary
The predictive value of pelvic scores, parity, age and gestational age for induction of labor by local prostaglandin-E, (PGE,) was examined in 336 women attempting induction of labor by intracervical or vaginal PGE,. The patient characteristics were correlated to: (1) vaginal delivery within 48 h, (2) the period from induction to onset of labor (latency period), and (3) the duration of labor. The Bishop score (P < 0.01) and even more the Lange score (P < 0.0001) were significantly inversely correlated to both latency period and induction-delivery period. This was caused by cervical dilatation (P < O.OOl), fetal station (P < 0.05) and cervical length (P < 0.051, whereas position and consistency of the cervix were of no importance. All three periods studied were significantly (P < 0.0001) shorter in parous women. In primiparous women, gestational age was of no importance for the latency period; however, higher gestational age was associated with longer labor (P < 0.001). We conclude that the predictive value of pelvic scores on induction hardly differs using local PGE, compared to conventional methods; furthermore, the Bishop score should be substituted, disregarding position and consistency of the cervix, but putting more weight to cervical dilatation. A new pelvic score is proposed. Induction of labor; PGE,; Pelvic scores; Parity; Gestational age
Introduction Before attempting prospects for readily should be evaluated. For a long time it duration of labor in
induction of labor, the inducing vaginal delivery has been known that the spontaneously delivering
Correspondence to: J. Lyndrup, Gammel Hareskowej 3500 Vaerloese, Denmark.
329,
women is correlated to the pre-labor station of the fetus, as well as to the dilatation and effacement of the cervix [1,2,]. Furthermore, similar correlations were observed when attempting induction of labor by oxytocin [3,4]. In 1964 Bishop [5] presented a pelvic score, compiling subscores for five pelvic findings: station of the fetus and dilatation, effacement, consistency and position of the cervix (Table 1). The Bishop score was elaborated for induction of multiparous women only and shown to determine the proximity of spontaneous onset of labor [5]. As time has passed
18 TABLE I The modified Bishop score as used in the present study Score
Station of head * Cervical dilatation (cm) length ** (cm) consistency position
0
1
2
3
> pi
= pi
= ps
< Ps
0 3 firm post.
l-2 2 medium mid
3-4 1 soft ant.
>4 0
* In relation to ( >, above; =, at; 4
0
The consistency and position of the cervix is ignored Otherwise, the score differs from the Bishop score doubling the weight of dilatation. * In relation to (beneath) the spines. In the present expressed in relation to the pelvic inlet and the pelvic as in the modified Bishop score (Table I).
( = 0). [5] by study spines,
of labor, especially when the cervix is unfavorable. PGE, ripens the pregnant cervix [lo], possibly by a direct action on the cervical connective tissue, mimicking the spontaneous cervical priming [ll]. Accordingly, when induction is attempted by applying PGE, locally to the cervix, the predictive value of cervical scorings may differ from that of systemic oxytocics. The present study examines the effect of induction in relation to the pre-induction information given by pelvic scorings and other patient characteristics when using locally applied PGE, for induction of labor. Materials The material was collected from four prospective, randomized studies dealing with induction methodology [12-l% In the present study, however, only the 336 pregnant women induced by local PGE, (in intracervical gel [12,15] or in vaginal pessaries 112-151) participated. The induction protocols, all running for two consecutive days (48 h), were conducted by the present authors at two university hospitals: Herlev Hospital [ 13,141 and Hvidovre Hospital [12,15]. All the women had singleton pregnancies, fetus in vertex presentation, intact membranes, a normal non-stress test and no previous operation on the uterus. Immediately before induction, all the patients were scored by the individual Bishop subscores (Table I), from which the total Bishop score [5] as well as the Lange score (Table II) were calculated. In one study [14], half of the patients temporarily had a Lamicel installed in the cervical canal as well; however, this was found to be of no importance [14,16]. The material is shown in Table III and Fig. 1 (for further details see previous publications). The induction period was analyzed in two intervals divided by the onset of labor (rhythmic regularity of uterine contractions followed by progressive cervical dilatation - Friedman et al. [4]), as judged by the attending midwife. Three periods were thus analyzed: (1) induction to onset of labor (“latency”); (2) onset of labor to vaginal delivery (“labor”); (3) induction to vaginal delivery (“induction”). For all three periods, the predictive value of
19 TABLE III
impact of maternal age, fetal weight (at delivery) and gestational age (confirmed by midtrimester ultrasound) were analyzed as well.
The materials Material
Patients Vaginal delivery ’
Lost 2
Induction failure ’
Total
Vaginal pessaries at Herlev H.
48 29
[I31 1141
2 8
Intracervical gel at Hvidovre H. 35 11 [121 34 10 [ISI Vaginal pessaries at Hvidovre H. 41 12 42 14 229 57
[121 1151
0 6
50 43
13
20
59 64
6 5
59 61
50
336
The present material as recalculated and compiled from four prospective studies [12-151. ’ The “event” of vaginal delivery. 2 “Lost to follow-up”, i.e. abdominal delivery or protocol deviation. 3 “Still at risk” for vaginal delivery at 48 h.
the individual Bishop subscores, the Bishop score, the Lange score and parity were analyzed. Furthermore, in primiparous women (only this group was large enough to enable such an analysis) the
Patients 50
1
0
m m
Vaginal pessaries at HerIev Hospital. Intracervical gel at Hvidovre Hospital. Vaginal pessaries at Hvidovre Hospital.
Bishop score Fig. 1. Distribution of patients according to Bishop score (Table I) and site of PGE, application (general strata).
Statistical procedures
Statistical analysis was conducted by the Statistical Research Unit of Copenhagen University employing life table analysis with compiled stratification (log rank test and Cox-Mantel test, BMDP pack 1988). Materials, not previously presented using life table analysis [12,13], have been recalculated to suit this ’ ([15], Table III). Patients delivered abdominally (42 patients) and protocol deviators (15 patients) were included until the time of caesarean section/ protocol deviation and then censored (as “lost to follow-up”). If onset of labor/vaginal delivery did not occur within the 48-h study period, the patients were censored after 48 h (as “still at risk”). To adjust for influences caused by differences in site of PGE, application, hospital procedures, parity and cervical ripeness, the patient material has been analyzed within a number of more uniform subgroups, as produced by dividing the material into strata on (two or more) decreasing levels. The subgroup results have then been computerized into compiled results. For all analysis, the patients were primarily arranged into three general strata according to: site of prostaglandin application and hospital (the general strata) 2: (1) vaginal pessaries at Her-
’ When considering the interval between induction and onset of labor, the onset of labor has been registered as the “event” and when considering that of induction to (vaginal) delivery or that of labor to (vaginal) delivery, vaginal delivery has been registered as the “event”. ’ Within each of the general strata two different materials are included. This is permissible as: (1) the two materials at Herlev Hospital followed identical protocols, except by including patients with ripe cervices [13]/unripe cervices [14]; (2) the two gel materials at Hvidovre Hospital [12,15] did not differ in regard to the periods analyzed; (3) although the results differed regarding “labor” (P < 0.01) between the two vaginal pessary materials at Hvidovre Hospital [12,15], no difference was apparent in respect to “latency” and “induction”; further recalculation with four general strata (the two vaginal pessary materials at Hvidovre in separate strata) did not change any conclusions.
20 lev Hospital 113,141; (2) intracervical gel at Hvidovre Hospital [12,15]; (3) vaginal pessaries at Hvidovre Hospital [12,15]. For trend analysis of cervical scores the three general strata were further divided into: strata for purity: (1) primiparous women; (2) parous women. For analysis of the influence of parity, the three general strata were divided into: strata for cervical ripeness: (1) Bishop score < 5 and; (2) Bishop score > 5. For trend analysis of patients’ age, gestational age and fetal weight, the three general strata were further stratified adjusting for both cervical ripeness (as above) and parity. However, as only the larger primiparous parity group (237 of 336 women) permitted such detailed stratification, these calculations were conducted for primiparous women only. P < 0.05 was considered statistically significant.
Results The inducibility of labor (as inversely expressed by the “latency” period, Table IV) was significantly related to the Bishop score (P < 0.01) and the individual subscores: cervical dilatation (P < 0.000, station of the fetal head (P < 0.05) and cervical length (P < 0.05). Consistency and
position of the cervix did not predict inducibility of labor. Accordingly, Lange score, ignoring consistency and position of the cervix and doubling the weight of cervical dilatation, had a highly significant (P < 0.0001) predictive value on labor inducibility. The Bishop score had an even more significant (P < 0.001) predictive value on inducibility of (vaginal) delivery (inversely expressed as “induction” period, Table IV) than on the latency period. Furthermore, as shown in Table IV, the same Bishop subscorings: fetal station, dilatation and cervical length were significant with slightly more extreme P values than for the latency period. Accordingly, the trend of Lange score again was extremely significant (P < 0.0001). No relationship was found between the “labor” period (Table IV) and any of the individual Bishop subscores (obtained before induction). The total Bishop score, however, reached the level of significance (P = 0.0346), whereas the Lange score was not related to the duration of labor. Parity, crudely corrected for cervical ripeness, was important for inducibility of labor, inducibility of vaginal delivery and decreased duration of labor to vaginal delivery (P < 0.0001, for all three periods). In primiparous women, the material was analyzed for trend of maternal age, gestational age and fetal weight at delivery taking site of
TABLE IV Analysis for trend of pelvic scores Score examined
Total Bishop score (< 3,4,5, > 6) Bishop subscores station of head (0, > 1) cervical dilatation (0, 2 1) cervical length (d 1, 2 2) cervical consistency (< 1, > 2) cervical position (0, > 1) Lange score (< 2,3,4, > 51
Periods examined “Latency”
“Induction”
“Labor”
P
P
P
0.0093 * 0.0463 0.0004 0.0457 0.2085 0.3755 < 0.0001
* * *
*
0.0007 * 0.0083 * 0.0901* 0.0176 * 0.4558 0.3015 < 0.0001 *
0.0346 * 0.0780 0.3228 0.5164 0.3522 0.3207 0.1090
Analysis for trend of pelvic scores (according to Tables I and II). P is the probability of “no influence” of the score on the periods. The material has been stratified for both site of PGE, application and parity. * P < 0.05. “Latency”: induction to onset of labor. “Induction”: induction to (vaginal) delivery. “Labor”: onset of labor to (vaginal) delivery.
21
prostaglandin application as well as cervical ripeness into account (Table V). Maternal age, was not related to either of the studied periods. Gestational age had no influence on the “latency” period or on the “induction” period; however, there was a highly significant correlation between (higher) gestational age and the duration of labor (2’ < 0.001). Fetal weight had no impact on either “latency” period or “induction” period though larger babies required longer labor (P < 0.01). Discussion Most investigators studying inducibility profiles have focused solely on patients delivering vaginally within 24 h, either ignoring caesarean sections or considering them as failures. In this study, the use of life table statistics has permitted inclusion of patients delivered abdominally or deviating from the protocol. Life table statistics also has the advantage of permitting compiled stratification, allowing - as with multiple regression analysis - the exclusion of competitive factors such as methods, hospital, parity and cervical ripeness. The value of pelvic scores for induction of labor using local PGE, has been studied before [ 17- 191. A significant inverse relationship was noted between Bishop score and induction-delivery time after vaginal PGE, in 109 patients [17]; however, the schedule for PGE, application and amniotomy differed whether or not the Bishop score was more than 6. In a pilot study,
primarily focusing on transvaginal ultrasound cervical assessment before induction of labor and only employing vaginal prostaglandin in patients with Bishop scores Q 6, the Bishop score was positively related to the proportion of women delivered vaginally [ 181. Furthermore, a system of reduced doses of vaginal PGE, with increasing Bishop score has been recommended [19]. The present study however, is the first large study evaluating in detail the relationship between preinduction pelvic findings and inducibility by locally administered PGE,. Applying local prostaglandins, our results resemble those of Lange et al. [9] who employed systemic oxytocics. Both the total Bishop score and even more the Lange score showed a significant inverse correlation to both the latency period (P < 0.01 and P < 0.0001, respectively) and the entire induction period (P < 0.001 and P < 0.0001, respectively). Further, the same Bishop subscores: fetal station, cervical dilatation and length of the cervix were significantly related to both the latency period and the induction period, whereas little could be learned from position and consistency of the cervix. In our study we found the impact of pelvic scores obtained at initiation of induction to differ little whether analyzing the latency period or the induction period (Table IV). As a matter of fact, in all respects we reached the same conclusions for both periods with lowest P values for the induction period, possibly reflecting delivery to be more accurately determined than the onset of
TABLE V The impact of maternal age, gestational age and fetal weight in primiparous women Item examined
Period examined “Latency”
Maternalage (d 23,24 - 27, 2 28 yr) Gestational age (< 38, 39 - 40, B 41 weeks) Fetal weight (Q 3ooO,3001 - 3799, > 3800 g)
“Induction”
“Labor”
P
P
P
0.1600 0.7864 0.9372
0.0579 0.3402 0.3292
0.5833 0.0002 * 0.0011 * *
P is the likelihood of “no trend” of the items on the periods. * Higher gestational age and * * larger babies were associated with longer labor. The material only comprises primiparous women; it is further stratified for site of prostaglandin medication and cervical ripeness.
22
labor. Accordingly, only little information was obtained from pre-induction pelvic findings regarding the labor period (onset of labor to vaginal delivery), the total Bishop score alone being significant 0’ = 0.0346). We have previously shown local PGE, (in vaginal pessaries) to be more efficient for induction of labor than i.v. oxytocin [13,14], especially within the first 24 h of induction [14]. However, disregarding such differences in the efficiency of the induction methods, our results considering the predictive value of pelvic scores on latency period/induction period are much like those of Friedman [6], Harrison et al. [7] and Lange et al. [91 using systemic oxytocics and/or amniotomy. This may’lead to two conclusions: (1) the impact of pelvic scores on induction hardly differs substantially whether local prostaglandin or systemic oxytocics and/or amniotomy are employed for induction, (2) the Bishop score as “the standard pelvic score” should be substituted by a scoring system, ignoring cervical consistency and cervical position as further discussed below. In this study, parity - crudely .adjusted for cervical ripeness - was highly significantly related to shorter latency period, shorter labor and thus sooner delivery (P < 0.0001, for al1 three periods). The influence of parity on labor is in .agreement with other studies using systemic oxytocics [5,9]. However, Lange et al. [9] in their study using multiple regression analysis found only a little impact of parity on the latency period. Our analysis of other inducibility factors in primiparae, showed, as reported by others [91, maternal age, gestational age and fetal weight to have little influence on onset of labor and induction of vaginal delivery. However, higher gestational age (P < 0.001) and larger babies (P < 0.01) in our study required longer labor. These observations are in contrast to those of Lange et al: [9], and we are not aware of other studies having examined these items adjusted for parity and “cervical ripeness”. As longer gestation was significantly correlated to larger babies (P < 0.00001, Spearman test), it is not possible to tell, if the observations simply reflect more labor being required to deliver larger babies. On the other hand, near term, our data do not support any
assumption of further improvement in uterine contractive potential with increasing gestational age. For almost 30 years, the Bishop score has been the world-wide approved pelvic score, irrespective of many other proposed pelvic scoring systems. However, including the present material, several large studies have shown cervical position [6,7,9] and cervical consistency [7,9] to contribute little to inducibility. Furthermore, all the studies [6,7,9] and the present one have shown cervical dilatation to be more important than the other scorings. Also, in the original Bishop score [S], effacement was expressed as a percentage of the (estimated) uneffaced cervix (see legend to Table I). However, as substantial variation exists in belief regarding the length of the uneffaced cervix near term [20], cervical length expressed in centimeters seems more appropriate [20]. A revision of the Bishop score as “the standard pelvic score” is thus relevant. We propose a pelvic score, essentially the same as proposed by Lange et al. ([93, Table II) but describing fetal position as in TabIe VI, since it seems more appropriate taking variations in size and shape of the pelvis into account. Furthermore, we suggest that the obvious consequences of doubling the weight of dilatation is taken and one point given for each centimeter of cervical dilatation (max. 4 points), rather than stepwise increasing the points by 2 point for every 2 cm. The proposed score (O-10 points) is shown in Table VI.
TABLE VI The proposed pelvic score Points
Fetal station * Cervical dilatation (cm) length (cm)
0
1
2
3
> pi
= pi
= ps
4
* Fetal position in relation to ( > , above; =, at; < , below) the pelvic inlet (pi) and the pelvic spines (PSI. A maximum of 10 points is obtainable.
23
In conclusion we found: (1) The predictive value of pelvic scores differed little whether it is related to the latency period or to the induction-delivery period. (2) The predictive value of pelvic scores on inducibility when using locally administered prostaglandin-E does not differ substantially from what has been described using systemic oxytocics (oxytocin or oral PGE,) and/or amniotomy; (however, the period of time needed for induction may vary with the different induction methods 1141). (3) Among the individual subscores of the Bishop score, we found cervical dilatation to be the most significant followed by fetal station and cervical length, whereas cervical consistency and cervical position were of no importance. (4) The predictive value of the Lange score on inducibility was more significant than that of the Bishop score. Accordingly, we suggest a score (Table VI) based on the score proposed by Lange et al. [9]. (5) Parous women have shorter latency periods than primiparous women. (6) Larger fetal weight and longer gestation is related to longer duration of labor. Acknowledgements
We sincerely thank our co-authors from previous publications: Niels Jorgen Secher, Carsten Dahl, Torben Philipsen and Poul Sindberg Eriksen for letting us use original material, which they share part of. This study has been supported by grants from The Danish Medical Research Council. References Calkins LA. On predicting the length of labor. Am J Obstet Gynecol 1941;42:802-813. Friedman EA. Primigravid labor: a graphicostatistical analysis. Obstet Gynecol 1955;6:567-589. Bishop EH. Elective induction of labor. Obstet Gynecol 1955;5:519-527. Friedman EA, Sachtleben MR. Determinant role of initial cervical dilatation on the course of labor. Am‘J Obstet Gynecol 1962;84:930-935.
5 Bishop EH. Pelvic scoring for elective induction. Obstet Gynecol 1964;24:266-268. 6 Friedman EA, Niswander KR, Bayonet-Rivera NP, Sachtleben MR. Relation of prelabor evaluation to inducibility and the course of labor. Obstet Gynecol 1966;28:495-501. 7 Harrison RF, Flynn M, Craft I. Assessment of factors constituting an “inducibility profile”. Obstet Gynecol 1977;49:270-274. 8 Hughey MJ, McElin TW, Bird CC. An evaluation of preinduction scoring systems. Obstet Gynecol 1976;48: 635-641. 9 Lange AP, Secher NJ, Westergaard JG, Skovgird I. Prelabor evaluation of inducibility. Obstet Gynecol 1982; 60:137-147. 10 Keirse MJNC, Van Oppen ACC. Preparing the cervix for induction of labor. In: Chalmers I, Enkin M, Keirse MJNC, eds. Effective care in pregnancy and childbirth, vol 2. Oxford: Oxford University Press, 1989. 11 Ekman S, MaImstrom A, Uldbjerg N, Ulmsten U. Cervical collagen: an important regulator of cervical function in term labor. Obstet Gynecol 1986;67:633-636. I2 Legarth J, Guldbek E, Secher NJ. The efficiency of prostaglandin E, vaginal suppositories versus intracervical prostaglandin gel for induction of labor in patients with unfavorable inducibility prospects. Eur J Obstet Gynecol Reprod Biol 1988;27:93-98. 13 Legarth J, Lyndrup J, Dahl C, Philipsen T, Eriksen PS. Prostaglandin E, vaginal suppository for induction of labor: an efficient, safe and popular method. Eur J Obstet Gynecol Reprod Biol 1987;26:233-238. 14 Lyndrup J, Legarth J, Dahl C, Philipsen T Eriksen PS, Weber T. Induction of labor: the effect of vaginal prostaglandin or i.v. oxytocin - a matter of time only? Eur J Obstet Gynecol Reprod Biol 1990;37:11l-l 19. 15 Lyndrup J, Nickelsen C, Guldbaek E, Weber T. Induction of labor by prostaglandin E,: intracervical gel or vaginal pessaries? Eur J Obstet Gynecol Reprod Biol 1991;42: 101-109. 16 Lyndrup J, Legarth J, Dahl C, Philipsen T, Eriksen PS. Lamicel (R) does not promote induction of labor. A randomized controlled study. Eur J Obstet Gynecol Reprod Biol 1989;30:205-208. 17 Houghton DJ. An evaluation of the Bishop scoring system in relation to a method of induction of labor by intravaginal prostaglandin. Postgrad Med J 1982;58:403-407. 18 Paterson-Brown S, Fisk NM, Edmonds DK, Rodek CH. Preinduction cervical assessment by Bishop’s score and transvaginal ultrasound. Eur J Obstet Gynecol Reprod Biol 1991;40:17-23. 19 Grunstein S, Jaschevatzky OE, Shali A, Noy A, Davidson A, Ellenbogen A. A scoring system for induction of labor using prostaglandin E, vaginal Tablets. Int J Gynecol Obstet 1990;31:131-134. 20 Holcomb WL, Smeltzer JS. Cervical effacement: variation in belief among clinicians. Obstet Gynecol 1991;78:43-45.
17
0 1992 Elsevier Science Publishers B.V. All rights reserved 002&X-2243/92/$05.00 EUROBS 01410
Predictive value of pelvic scores for induction of labor by local PGE, Jens Lyndrup a,b,Jesper Legarth a,b,Tom Weber b, Carsten Nickelsen b and Else Guldbaek b Departments of Gynecology and Obstetrics, ’ Herlev University Hospital and ’ Hvidovre University Hospital, Copenhagen, Denmark
Accepted for publication 15 July 1992
Summary
The predictive value of pelvic scores, parity, age and gestational age for induction of labor by local prostaglandin-E, (PGE,) was examined in 336 women attempting induction of labor by intracervical or vaginal PGE,. The patient characteristics were correlated to: (1) vaginal delivery within 48 h, (2) the period from induction to onset of labor (latency period), and (3) the duration of labor. The Bishop score (P < 0.01) and even more the Lange score (P < 0.0001) were significantly inversely correlated to both latency period and induction-delivery period. This was caused by cervical dilatation (P < O.OOl), fetal station (P < 0.05) and cervical length (P < 0.051, whereas position and consistency of the cervix were of no importance. All three periods studied were significantly (P < 0.0001) shorter in parous women. In primiparous women, gestational age was of no importance for the latency period; however, higher gestational age was associated with longer labor (P < 0.001). We conclude that the predictive value of pelvic scores on induction hardly differs using local PGE, compared to conventional methods; furthermore, the Bishop score should be substituted, disregarding position and consistency of the cervix, but putting more weight to cervical dilatation. A new pelvic score is proposed. Induction of labor; PGE,; Pelvic scores; Parity; Gestational age
Introduction Before attempting prospects for readily should be evaluated. For a long time it duration of labor in
induction of labor, the inducing vaginal delivery has been known that the spontaneously delivering
Correspondence to: J. Lyndrup, Gammel Hareskowej 3500 Vaerloese, Denmark.
329,
women is correlated to the pre-labor station of the fetus, as well as to the dilatation and effacement of the cervix [1,2,]. Furthermore, similar correlations were observed when attempting induction of labor by oxytocin [3,4]. In 1964 Bishop [5] presented a pelvic score, compiling subscores for five pelvic findings: station of the fetus and dilatation, effacement, consistency and position of the cervix (Table 1). The Bishop score was elaborated for induction of multiparous women only and shown to determine the proximity of spontaneous onset of labor [5]. As time has passed
18 TABLE I The modified Bishop score as used in the present study Score
Station of head * Cervical dilatation (cm) length ** (cm) consistency position
0
1
2
3
> pi
= pi
= ps
< Ps
0 3 firm post.
l-2 2 medium mid
3-4 1 soft ant.
>4 0
* In relation to ( >, above; =, at; 4
0
The consistency and position of the cervix is ignored Otherwise, the score differs from the Bishop score doubling the weight of dilatation. * In relation to (beneath) the spines. In the present expressed in relation to the pelvic inlet and the pelvic as in the modified Bishop score (Table I).
( = 0). [5] by study spines,
of labor, especially when the cervix is unfavorable. PGE, ripens the pregnant cervix [lo], possibly by a direct action on the cervical connective tissue, mimicking the spontaneous cervical priming [ll]. Accordingly, when induction is attempted by applying PGE, locally to the cervix, the predictive value of cervical scorings may differ from that of systemic oxytocics. The present study examines the effect of induction in relation to the pre-induction information given by pelvic scorings and other patient characteristics when using locally applied PGE, for induction of labor. Materials The material was collected from four prospective, randomized studies dealing with induction methodology [12-l% In the present study, however, only the 336 pregnant women induced by local PGE, (in intracervical gel [12,15] or in vaginal pessaries 112-151) participated. The induction protocols, all running for two consecutive days (48 h), were conducted by the present authors at two university hospitals: Herlev Hospital [ 13,141 and Hvidovre Hospital [12,15]. All the women had singleton pregnancies, fetus in vertex presentation, intact membranes, a normal non-stress test and no previous operation on the uterus. Immediately before induction, all the patients were scored by the individual Bishop subscores (Table I), from which the total Bishop score [5] as well as the Lange score (Table II) were calculated. In one study [14], half of the patients temporarily had a Lamicel installed in the cervical canal as well; however, this was found to be of no importance [14,16]. The material is shown in Table III and Fig. 1 (for further details see previous publications). The induction period was analyzed in two intervals divided by the onset of labor (rhythmic regularity of uterine contractions followed by progressive cervical dilatation - Friedman et al. [4]), as judged by the attending midwife. Three periods were thus analyzed: (1) induction to onset of labor (“latency”); (2) onset of labor to vaginal delivery (“labor”); (3) induction to vaginal delivery (“induction”). For all three periods, the predictive value of
19 TABLE III
impact of maternal age, fetal weight (at delivery) and gestational age (confirmed by midtrimester ultrasound) were analyzed as well.
The materials Material
Patients Vaginal delivery ’
Lost 2
Induction failure ’
Total
Vaginal pessaries at Herlev H.
48 29
[I31 1141
2 8
Intracervical gel at Hvidovre H. 35 11 [121 34 10 [ISI Vaginal pessaries at Hvidovre H. 41 12 42 14 229 57
[121 1151
0 6
50 43
13
20
59 64
6 5
59 61
50
336
The present material as recalculated and compiled from four prospective studies [12-151. ’ The “event” of vaginal delivery. 2 “Lost to follow-up”, i.e. abdominal delivery or protocol deviation. 3 “Still at risk” for vaginal delivery at 48 h.
the individual Bishop subscores, the Bishop score, the Lange score and parity were analyzed. Furthermore, in primiparous women (only this group was large enough to enable such an analysis) the
Patients 50
1
0
m m
Vaginal pessaries at HerIev Hospital. Intracervical gel at Hvidovre Hospital. Vaginal pessaries at Hvidovre Hospital.
Bishop score Fig. 1. Distribution of patients according to Bishop score (Table I) and site of PGE, application (general strata).
Statistical procedures
Statistical analysis was conducted by the Statistical Research Unit of Copenhagen University employing life table analysis with compiled stratification (log rank test and Cox-Mantel test, BMDP pack 1988). Materials, not previously presented using life table analysis [12,13], have been recalculated to suit this ’ ([15], Table III). Patients delivered abdominally (42 patients) and protocol deviators (15 patients) were included until the time of caesarean section/ protocol deviation and then censored (as “lost to follow-up”). If onset of labor/vaginal delivery did not occur within the 48-h study period, the patients were censored after 48 h (as “still at risk”). To adjust for influences caused by differences in site of PGE, application, hospital procedures, parity and cervical ripeness, the patient material has been analyzed within a number of more uniform subgroups, as produced by dividing the material into strata on (two or more) decreasing levels. The subgroup results have then been computerized into compiled results. For all analysis, the patients were primarily arranged into three general strata according to: site of prostaglandin application and hospital (the general strata) 2: (1) vaginal pessaries at Her-
’ When considering the interval between induction and onset of labor, the onset of labor has been registered as the “event” and when considering that of induction to (vaginal) delivery or that of labor to (vaginal) delivery, vaginal delivery has been registered as the “event”. ’ Within each of the general strata two different materials are included. This is permissible as: (1) the two materials at Herlev Hospital followed identical protocols, except by including patients with ripe cervices [13]/unripe cervices [14]; (2) the two gel materials at Hvidovre Hospital [12,15] did not differ in regard to the periods analyzed; (3) although the results differed regarding “labor” (P < 0.01) between the two vaginal pessary materials at Hvidovre Hospital [12,15], no difference was apparent in respect to “latency” and “induction”; further recalculation with four general strata (the two vaginal pessary materials at Hvidovre in separate strata) did not change any conclusions.
20 lev Hospital 113,141; (2) intracervical gel at Hvidovre Hospital [12,15]; (3) vaginal pessaries at Hvidovre Hospital [12,15]. For trend analysis of cervical scores the three general strata were further divided into: strata for purity: (1) primiparous women; (2) parous women. For analysis of the influence of parity, the three general strata were divided into: strata for cervical ripeness: (1) Bishop score < 5 and; (2) Bishop score > 5. For trend analysis of patients’ age, gestational age and fetal weight, the three general strata were further stratified adjusting for both cervical ripeness (as above) and parity. However, as only the larger primiparous parity group (237 of 336 women) permitted such detailed stratification, these calculations were conducted for primiparous women only. P < 0.05 was considered statistically significant.
Results The inducibility of labor (as inversely expressed by the “latency” period, Table IV) was significantly related to the Bishop score (P < 0.01) and the individual subscores: cervical dilatation (P < 0.000, station of the fetal head (P < 0.05) and cervical length (P < 0.05). Consistency and
position of the cervix did not predict inducibility of labor. Accordingly, Lange score, ignoring consistency and position of the cervix and doubling the weight of cervical dilatation, had a highly significant (P < 0.0001) predictive value on labor inducibility. The Bishop score had an even more significant (P < 0.001) predictive value on inducibility of (vaginal) delivery (inversely expressed as “induction” period, Table IV) than on the latency period. Furthermore, as shown in Table IV, the same Bishop subscorings: fetal station, dilatation and cervical length were significant with slightly more extreme P values than for the latency period. Accordingly, the trend of Lange score again was extremely significant (P < 0.0001). No relationship was found between the “labor” period (Table IV) and any of the individual Bishop subscores (obtained before induction). The total Bishop score, however, reached the level of significance (P = 0.0346), whereas the Lange score was not related to the duration of labor. Parity, crudely corrected for cervical ripeness, was important for inducibility of labor, inducibility of vaginal delivery and decreased duration of labor to vaginal delivery (P < 0.0001, for all three periods). In primiparous women, the material was analyzed for trend of maternal age, gestational age and fetal weight at delivery taking site of
TABLE IV Analysis for trend of pelvic scores Score examined
Total Bishop score (< 3,4,5, > 6) Bishop subscores station of head (0, > 1) cervical dilatation (0, 2 1) cervical length (d 1, 2 2) cervical consistency (< 1, > 2) cervical position (0, > 1) Lange score (< 2,3,4, > 51
Periods examined “Latency”
“Induction”
“Labor”
P
P
P
0.0093 * 0.0463 0.0004 0.0457 0.2085 0.3755 < 0.0001
* * *
*
0.0007 * 0.0083 * 0.0901* 0.0176 * 0.4558 0.3015 < 0.0001 *
0.0346 * 0.0780 0.3228 0.5164 0.3522 0.3207 0.1090
Analysis for trend of pelvic scores (according to Tables I and II). P is the probability of “no influence” of the score on the periods. The material has been stratified for both site of PGE, application and parity. * P < 0.05. “Latency”: induction to onset of labor. “Induction”: induction to (vaginal) delivery. “Labor”: onset of labor to (vaginal) delivery.
21
prostaglandin application as well as cervical ripeness into account (Table V). Maternal age, was not related to either of the studied periods. Gestational age had no influence on the “latency” period or on the “induction” period; however, there was a highly significant correlation between (higher) gestational age and the duration of labor (2’ < 0.001). Fetal weight had no impact on either “latency” period or “induction” period though larger babies required longer labor (P < 0.01). Discussion Most investigators studying inducibility profiles have focused solely on patients delivering vaginally within 24 h, either ignoring caesarean sections or considering them as failures. In this study, the use of life table statistics has permitted inclusion of patients delivered abdominally or deviating from the protocol. Life table statistics also has the advantage of permitting compiled stratification, allowing - as with multiple regression analysis - the exclusion of competitive factors such as methods, hospital, parity and cervical ripeness. The value of pelvic scores for induction of labor using local PGE, has been studied before [ 17- 191. A significant inverse relationship was noted between Bishop score and induction-delivery time after vaginal PGE, in 109 patients [17]; however, the schedule for PGE, application and amniotomy differed whether or not the Bishop score was more than 6. In a pilot study,
primarily focusing on transvaginal ultrasound cervical assessment before induction of labor and only employing vaginal prostaglandin in patients with Bishop scores Q 6, the Bishop score was positively related to the proportion of women delivered vaginally [ 181. Furthermore, a system of reduced doses of vaginal PGE, with increasing Bishop score has been recommended [19]. The present study however, is the first large study evaluating in detail the relationship between preinduction pelvic findings and inducibility by locally administered PGE,. Applying local prostaglandins, our results resemble those of Lange et al. [9] who employed systemic oxytocics. Both the total Bishop score and even more the Lange score showed a significant inverse correlation to both the latency period (P < 0.01 and P < 0.0001, respectively) and the entire induction period (P < 0.001 and P < 0.0001, respectively). Further, the same Bishop subscores: fetal station, cervical dilatation and length of the cervix were significantly related to both the latency period and the induction period, whereas little could be learned from position and consistency of the cervix. In our study we found the impact of pelvic scores obtained at initiation of induction to differ little whether analyzing the latency period or the induction period (Table IV). As a matter of fact, in all respects we reached the same conclusions for both periods with lowest P values for the induction period, possibly reflecting delivery to be more accurately determined than the onset of
TABLE V The impact of maternal age, gestational age and fetal weight in primiparous women Item examined
Period examined “Latency”
Maternalage (d 23,24 - 27, 2 28 yr) Gestational age (< 38, 39 - 40, B 41 weeks) Fetal weight (Q 3ooO,3001 - 3799, > 3800 g)
“Induction”
“Labor”
P
P
P
0.1600 0.7864 0.9372
0.0579 0.3402 0.3292
0.5833 0.0002 * 0.0011 * *
P is the likelihood of “no trend” of the items on the periods. * Higher gestational age and * * larger babies were associated with longer labor. The material only comprises primiparous women; it is further stratified for site of prostaglandin medication and cervical ripeness.
22
labor. Accordingly, only little information was obtained from pre-induction pelvic findings regarding the labor period (onset of labor to vaginal delivery), the total Bishop score alone being significant 0’ = 0.0346). We have previously shown local PGE, (in vaginal pessaries) to be more efficient for induction of labor than i.v. oxytocin [13,14], especially within the first 24 h of induction [14]. However, disregarding such differences in the efficiency of the induction methods, our results considering the predictive value of pelvic scores on latency period/induction period are much like those of Friedman [6], Harrison et al. [7] and Lange et al. [91 using systemic oxytocics and/or amniotomy. This may’lead to two conclusions: (1) the impact of pelvic scores on induction hardly differs substantially whether local prostaglandin or systemic oxytocics and/or amniotomy are employed for induction, (2) the Bishop score as “the standard pelvic score” should be substituted by a scoring system, ignoring cervical consistency and cervical position as further discussed below. In this study, parity - crudely .adjusted for cervical ripeness - was highly significantly related to shorter latency period, shorter labor and thus sooner delivery (P < 0.0001, for al1 three periods). The influence of parity on labor is in .agreement with other studies using systemic oxytocics [5,9]. However, Lange et al. [9] in their study using multiple regression analysis found only a little impact of parity on the latency period. Our analysis of other inducibility factors in primiparae, showed, as reported by others [91, maternal age, gestational age and fetal weight to have little influence on onset of labor and induction of vaginal delivery. However, higher gestational age (P < 0.001) and larger babies (P < 0.01) in our study required longer labor. These observations are in contrast to those of Lange et al: [9], and we are not aware of other studies having examined these items adjusted for parity and “cervical ripeness”. As longer gestation was significantly correlated to larger babies (P < 0.00001, Spearman test), it is not possible to tell, if the observations simply reflect more labor being required to deliver larger babies. On the other hand, near term, our data do not support any
assumption of further improvement in uterine contractive potential with increasing gestational age. For almost 30 years, the Bishop score has been the world-wide approved pelvic score, irrespective of many other proposed pelvic scoring systems. However, including the present material, several large studies have shown cervical position [6,7,9] and cervical consistency [7,9] to contribute little to inducibility. Furthermore, all the studies [6,7,9] and the present one have shown cervical dilatation to be more important than the other scorings. Also, in the original Bishop score [S], effacement was expressed as a percentage of the (estimated) uneffaced cervix (see legend to Table I). However, as substantial variation exists in belief regarding the length of the uneffaced cervix near term [20], cervical length expressed in centimeters seems more appropriate [20]. A revision of the Bishop score as “the standard pelvic score” is thus relevant. We propose a pelvic score, essentially the same as proposed by Lange et al. ([93, Table II) but describing fetal position as in TabIe VI, since it seems more appropriate taking variations in size and shape of the pelvis into account. Furthermore, we suggest that the obvious consequences of doubling the weight of dilatation is taken and one point given for each centimeter of cervical dilatation (max. 4 points), rather than stepwise increasing the points by 2 point for every 2 cm. The proposed score (O-10 points) is shown in Table VI.
TABLE VI The proposed pelvic score Points
Fetal station * Cervical dilatation (cm) length (cm)
0
1
2
3
> pi
= pi
= ps
4
* Fetal position in relation to ( > , above; =, at; < , below) the pelvic inlet (pi) and the pelvic spines (PSI. A maximum of 10 points is obtainable.
23
In conclusion we found: (1) The predictive value of pelvic scores differed little whether it is related to the latency period or to the induction-delivery period. (2) The predictive value of pelvic scores on inducibility when using locally administered prostaglandin-E does not differ substantially from what has been described using systemic oxytocics (oxytocin or oral PGE,) and/or amniotomy; (however, the period of time needed for induction may vary with the different induction methods 1141). (3) Among the individual subscores of the Bishop score, we found cervical dilatation to be the most significant followed by fetal station and cervical length, whereas cervical consistency and cervical position were of no importance. (4) The predictive value of the Lange score on inducibility was more significant than that of the Bishop score. Accordingly, we suggest a score (Table VI) based on the score proposed by Lange et al. [9]. (5) Parous women have shorter latency periods than primiparous women. (6) Larger fetal weight and longer gestation is related to longer duration of labor. Acknowledgements
We sincerely thank our co-authors from previous publications: Niels Jorgen Secher, Carsten Dahl, Torben Philipsen and Poul Sindberg Eriksen for letting us use original material, which they share part of. This study has been supported by grants from The Danish Medical Research Council. References Calkins LA. On predicting the length of labor. Am J Obstet Gynecol 1941;42:802-813. Friedman EA. Primigravid labor: a graphicostatistical analysis. Obstet Gynecol 1955;6:567-589. Bishop EH. Elective induction of labor. Obstet Gynecol 1955;5:519-527. Friedman EA, Sachtleben MR. Determinant role of initial cervical dilatation on the course of labor. Am‘J Obstet Gynecol 1962;84:930-935.
5 Bishop EH. Pelvic scoring for elective induction. Obstet Gynecol 1964;24:266-268. 6 Friedman EA, Niswander KR, Bayonet-Rivera NP, Sachtleben MR. Relation of prelabor evaluation to inducibility and the course of labor. Obstet Gynecol 1966;28:495-501. 7 Harrison RF, Flynn M, Craft I. Assessment of factors constituting an “inducibility profile”. Obstet Gynecol 1977;49:270-274. 8 Hughey MJ, McElin TW, Bird CC. An evaluation of preinduction scoring systems. Obstet Gynecol 1976;48: 635-641. 9 Lange AP, Secher NJ, Westergaard JG, Skovgird I. Prelabor evaluation of inducibility. Obstet Gynecol 1982; 60:137-147. 10 Keirse MJNC, Van Oppen ACC. Preparing the cervix for induction of labor. In: Chalmers I, Enkin M, Keirse MJNC, eds. Effective care in pregnancy and childbirth, vol 2. Oxford: Oxford University Press, 1989. 11 Ekman S, MaImstrom A, Uldbjerg N, Ulmsten U. Cervical collagen: an important regulator of cervical function in term labor. Obstet Gynecol 1986;67:633-636. I2 Legarth J, Guldbek E, Secher NJ. The efficiency of prostaglandin E, vaginal suppositories versus intracervical prostaglandin gel for induction of labor in patients with unfavorable inducibility prospects. Eur J Obstet Gynecol Reprod Biol 1988;27:93-98. 13 Legarth J, Lyndrup J, Dahl C, Philipsen T, Eriksen PS. Prostaglandin E, vaginal suppository for induction of labor: an efficient, safe and popular method. Eur J Obstet Gynecol Reprod Biol 1987;26:233-238. 14 Lyndrup J, Legarth J, Dahl C, Philipsen T Eriksen PS, Weber T. Induction of labor: the effect of vaginal prostaglandin or i.v. oxytocin - a matter of time only? Eur J Obstet Gynecol Reprod Biol 1990;37:11l-l 19. 15 Lyndrup J, Nickelsen C, Guldbaek E, Weber T. Induction of labor by prostaglandin E,: intracervical gel or vaginal pessaries? Eur J Obstet Gynecol Reprod Biol 1991;42: 101-109. 16 Lyndrup J, Legarth J, Dahl C, Philipsen T, Eriksen PS. Lamicel (R) does not promote induction of labor. A randomized controlled study. Eur J Obstet Gynecol Reprod Biol 1989;30:205-208. 17 Houghton DJ. An evaluation of the Bishop scoring system in relation to a method of induction of labor by intravaginal prostaglandin. Postgrad Med J 1982;58:403-407. 18 Paterson-Brown S, Fisk NM, Edmonds DK, Rodek CH. Preinduction cervical assessment by Bishop’s score and transvaginal ultrasound. Eur J Obstet Gynecol Reprod Biol 1991;40:17-23. 19 Grunstein S, Jaschevatzky OE, Shali A, Noy A, Davidson A, Ellenbogen A. A scoring system for induction of labor using prostaglandin E, vaginal Tablets. Int J Gynecol Obstet 1990;31:131-134. 20 Holcomb WL, Smeltzer JS. Cervical effacement: variation in belief among clinicians. Obstet Gynecol 1991;78:43-45.
