In a previous investigation,1 we evaluated a labor management protocol, which required that labor be arrested for a minimum of 4 hours (if uterine activity was greater than 200 Montevideo units) or 6 hours (if greater than 200 Montevideo units could not be sustained) before performing a cesarean for active phase labor arrest. We concluded that these objective criteria for labor arrest allowed, by virtue of continued labor augmentation, a substantial proportion of women eligible for cesarean because their labor was arrested for 2 hours, to ultimately achieve a safe vaginal delivery. That investigation was essentially an “intent-to-manage” analysis, which depended on the correct bedside interpretation of Montevideo units by the managing obstetricians, and their adherence to our minimum duration criteria for the diagnosis of labor arrest. In that analysis, we studied temporal aspects of labor in depth, but did not independently confirm labor adequacy. Rather, we relied on the reported Montevideo units recorded in the nursing charts.
Our purpose in designing this study was to provide contemporary, validated uterine activity and labor progress data from women receiving oxytocin for protracted or arrested active phase labor. We recognized that prior investigations had focused predominantly on “normal” labor,2 or were performed before the widespread use of epidural analgesia, which is known to slow labor progress.3 Our intent was also to revisit, with well-characterized uterine activity data, the question of whether 2 hours of active phase labor arrest with at least 200 Montevideo units are sufficiently rigorous criteria for the performance of cesarean delivery.4 Therefore, we prospectively assessed a new cohort of 501 women in spontaneous active labor who were managed by the same protocol, independently measured their uterine activity, and correlated labor adequacy with progress and outcome.
MATERIALS AND METHODS
From September 19, 1998, to November 12, 2000, we assessed 501 consecutive women on the Maternal-Fetal Medicine service of the University of Alabama at Birmingham Hospital who were managed according to our previously reported labor management protocol.1 To be eligible, women were at or beyond 36 weeks' gestation; in spontaneous active phase labor, defined as a cervical dilation of at least 3 cm and complete effacement, or 4 cm and at least 75% effacement, with regular uterine contractions (at least two in 10 minutes); and receiving oxytocin for slowly progressive or arrested labor. Women with a nonvertex presentation, previous cesarean delivery, multiple gestation, or, before oxytocin initiation, a nonreassuring fetal heart rate tracing or chorioamnionitis, were not consistently managed according to this protocol, and were not included in this study population. All women had one-on-one nursing support throughout labor and were managed by Obstetrics and Gynecology residents under direct in-house Maternal-Fetal Medicine faculty supervision. None ambulated after oxytocin initiation, but instead labored in the lateral recumbent position. Throughout the period of study, a risk-based approach5 to the prevention of early onset neonatal group B streptococcal disease was employed. Because the labor protocol represented (and represents) our institutional standard, individual patient consent for the study was not sought. However, Institutional Review Board approval was obtained to review the medical records of women managed by protocol.
Maternal and neonatal characteristics and outcomes were defined and abstracted as described previously,1 with the addition of clinically diagnosed shoulder dystocia and neonatal brachial plexus injury. Uterine activity recordings were reviewed, starting from intrauterine pressure catheter insertion (which generally occurred before or concomitant with oxytocin initiation) and ending at complete cervical dilation or active phase cesarean delivery. For each consecutive 10-minute window, Montevideo units were calculated (as the sum of peak pressures above baseline of all contractions). Contractions spanning two consecutive 10-minute windows were counted in the window where the peak occurred.
Uterine activity recordings were reviewed by one of three research nurses, one of whom (KGS) reviewed approximately 80% of the recordings. Before instituting the study, archived recordings were assessed (by DJR, JO, and KGS) until greater than 90% agreement in interpretation was consistently achieved. One of the authors (KGS) then repeated this process with the other two research nurses. Over the course of the study, no specific measures were taken to blind the uterine activity reviewers to the labor or pregnancy outcome. However, in many cases, uterine activity review and maternal and neonatal outcome assessment occurred in temporal isolation, and thus, for practical purposes, were unlinked.
In the first portion of this investigation, we analyzed nulliparous and parous women separately. We calculated the mean and range of Montevideo units after the initiation of oxytocin, maximum oxytocin infusion rate, length of augmentation, and rate of active phase cervical dilation for women delivered vaginally compared with those delivered by cesarean. Because the achievement and maintenance of at least 200 Montevideo units is a recommended goal of our labor-augmentation protocol, we further analyzed women by whether they ever achieved at least 200 Montevideo units. If they did, we determined whether this level of uterine activity was sustained (ie, at least 200 Montevideo units confirmed in at least 50% of 10-minute windows after the first occurrence of 200 Montevideo units). Using these definitions, we categorized our patients into three uterine activity groups for comparison of delivery methods and indications: never achieved 200 Montevideo units; achieved but did not sustain 200 Montevideo units; and sustained Montevideo units of at least 200.
For the second part of this investigation, we identified the subset of women who experienced labor arrest for at least 2 hours despite sustained uterine activity of at least 200 Montevideo units. To do this, we identified women who achieved a sustained uterine contraction pattern of at least 200 Montevideo units, and had at least two informative cervical examinations from which to determine whether labor was arrested and, if so, for how long. We required that an index cervical examination be recorded within no more than one-half hour of the first 10-minute window with at least 200 Montevideo units, followed by a subsequent examination performed from 1.5–4.0 hours (median = 2.06 hours) after the index examination during which the dilation increased by no more than 1 cm. Mode of delivery and maternal and neonatal complications in this subgroup were then compared with the remainder of the study cohort.
For selected outcomes, exact 95% confidence intervals (CI) were calculated. Proportional data were compared by χ2 or Fisher exact test as appropriate, or relative risks (RR) and 95% CI. Normally distributed data were compared using the Student t test, and non-normal continuous data were compared with the Wilcoxon rank-sum test. P < .05 was considered significant. Data were analyzed using the SAS system version 8.0 for Windows (SAS Institute, Cary, NC). We established our sample size of at least 500 women based on the pragmatic and statistical considerations that a cohort of this size could be assembled in our institution in 2 years, and that this would enable us to achieve high precision (narrow CI) around most observed event rates. For example, with a sample of 500, the upper limit of the exact binomial 95% CI for an event with an observed rate of 1% is less than 3%.
Two hundred eighty-six (57%) of the 501 women were nulliparous, and 215 (43%) were parous. The mean age (years) of nulliparas and parous women was 20 and 25, respectively, and mean weight 80 kg and 84 kg. Seventy-two percent of nulliparas and 76% of parous women were black; 21% and 20% were white. The mean gestational age at delivery was 40 weeks for both groups; the mean birth weight was 3339 g and 3337 g, respectively; and 97% and 90% received lumbar epidural analgesia for labor. Thirty infants (6%) had birth weights over 4000 g, and one infant (.2%) had a birth weight over 4500 g. Fifteen infants (3%) weighed less at birth than the 10th percentile for gestational age according to the standard of Brenner et al.6
The cesarean rate for nulliparas was 18% and for parous women it was 5%. Neither the average number of Montevideo units, nor the maximum oxytocin infusion rate varied substantially between nulliparous and parous women, or between women who were delivered vaginally versus those who were delivered by cesarean (Table 1). However, nulliparous women who underwent cesarean received oxytocin for a median 3.5 hours longer than nulliparas who were delivered vaginally (P < .001), and parous women who underwent cesarean received oxytocin for a median 4.6 hours longer than those who were delivered vaginally (P < .001, Table 1). When compared by maternal age, weight, diabetic status, rupture of membranes before oxytocin administration, dilation at first cervical examination, and infant birth weight, only the last variable differed between women delivered by cesarean and those delivered vaginally: median 3510 g versus 3296 g, respectively (P < .001), and % over 4000 g, 13% versus 5%, respectively (P = .01).
After the initiation of oxytocin, nulliparas who were delivered vaginally dilated at a median rate of 1.4 cm/hour, compared with 1.8 cm/hour for parous women who were delivered vaginally (Table 1). Women who were delivered by cesarean dilated at much slower rates: 0.3 cm/hour for nulliparas (P < .001) and 0.4 cm/hour for parous women (P < .001, Table 1). For both nulliparous and parous women who were delivered vaginally, the 10th percentile of cervical dilation rate after the initiation of oxytocin was 0.6 cm/hour, and the fifth percentile was 0.5 cm/hour.
The majority of women (54%) achieved a sustained uterine contraction pattern of at least 200 Montevideo units after the initiation of oxytocin (Tables 2 and 3). Augmentation to achieve this level of uterine activity was associated with low rates of intervention for fetal indications: the rate of cesarean performed solely for fetal intolerance of labor was 3% among nulliparas (Table 2), and only one parous woman underwent cesarean solely for this indication (Table 3). There were no significant differences in delivery type or indication across the three contraction pattern groups among nulliparous (P = .32) or parous women (P = .64).
No woman experienced uterine rupture or underwent hysterectomy. Rates of chorioamnionitis and endometritis for nulliparas were 17% and 7%, respectively, and for parous women, 5% and 3%, respectively. Three percent of women experienced postpartum hemorrhage, and 1% received a red blood cell transfusion.
Four infants (1%) had serious congenital malformations and were excluded from the analysis of neonatal outcomes. No stillbirths or neonatal deaths occurred among the remaining 497 infants. No infants were diagnosed with or had sequelae attributed to asphyxia.7 Five infants (1%) suffered a serious neonatal complication: pneumonia requiring mechanical ventilation (one), sepsis (one), nonasphyxial seizures (two), and persistent fetal circulation requiring mechanical ventilation (one). None sustained a brachial plexus injury.
Thirty-eight women could be identified who experienced labor arrest for at least 2 hours despite a sustained uterine contraction pattern of at least 200 Montevideo units (30 nulliparous and eight parous women). Of these 38, 23 (61%, 95% CI 42, 77%) achieved a vaginal delivery (60% of nulliparas and 63% of parous women). Although the rates of chorioamnionitis (26%) and endometritis (26%) experienced by these women were significantly higher than the rates experienced by the rest of the cohort (10% and 4%, respectively, RR, 95% CI 2.5, 1.4, 4.6, and 6.8, 3.4, 13.6, respectively), none of the 38 experienced a postpartum hemorrhage. The deliveries of three of the 23 women (13%) who delivered vaginally were complicated by shoulder dystocia, compared with four of 412 (1%) vaginally delivered women who did not experience a sustained labor arrest (RR, 95% CI 13.6, 3.2, 57.1). All shoulder dystocias were described as “mild” by the managing physicians, responded promptly to McRobert's maneuver alone or in combination with suprapubic pressure, and were not associated with brachial plexus injury. None of the 38 infants (95% CI 0, 10%) born to these women sustained a serious neonatal complication such as birth asphyxia, seizures, requirement for mechanical ventilation, pneumonia, or sepsis. All had 5-minute Apgar scores of at least 6, and none had an umbilical artery cord pH of less than 7.1.
These data demonstrate several points about contemporary augmented labor in gravidas with a term gestation. First, oxytocin-stimulated labor, which eventuates in vaginal delivery, proceeds at substantially slower rates than spontaneous labor. Friedman observed that the 5th percentile rate of dilation for nulliparas in spontaneous active labor was 1.2 cm/hour, compared with 1.5 cm/hour for parous women.2 We observed 5th percentile rates of 0.5 cm/hour for both nulliparous and parous women, data which are consistent with those of Kelly et al, who recently reported a 5th percentile active labor dilation rate for nulliparas of 0.7 cm/hour (Kelly G, Peaceman AM, Colangelo L, Rademaker A. Normal nulliparous labor: Are Friedman's definitions still relevant? Am J Obstet Gynecol 2000;182:S129 [abstract]).
There are two plausible explanations for the slower rates observed both in this study and in the study of Kelly et al. First, a high percentage of women in both reports (84% and 96%, respectively) received epidural analgesia for the relief of labor pain. In contrast, few if any of the women on whom Friedman reported received it.2 Epidural analgesia slows labor progress. In the randomized trial of Sharma et al,3 allocation to epidural analgesia (as opposed to intravenous narcotics) increased the subsequent duration of first-stage labor by 1 hour. Second, but equally important, is the fact that although Friedman reported on normal labor (ie, labor not requiring oxytocin augmentation), we have characterized abnormal labor: all women in our cohort, and 63% of those in Kelly et al's report, received oxytocin for labor augmentation. Thus, our data, and the data of Kelly et al, suggest that the parameters of labor progress described by Friedman do not apply to oxytocin-augmented (or epidural-modified) labor, and that interventions predicated on those parameters will result in unnecessary cesarean delivery.
Our data also confirm that the majority of women (61%) who experience 2 hours of labor arrest despite a sustained uterine contraction pattern of at least 200 Montevideo units will achieve vaginal delivery if oxytocin augmentation is continued. Moreover, the infants of such women were no more likely to experience complications than the infants of women whose labor progressed in a more normal fashion. Although shoulder dystocia was more common in the infants of the 23 women who met these arrest criteria and delivered vaginally than in the 412 who did not (13% compared with 1%), these shoulder dystocias were mild, easily reduced, and did not result in neonatal injury. For these reasons, and the fact that we could not exclude clinical bias in the diagnosis of shoulder dystocia (eg, that clinicians may have been more likely to diagnose shoulder dystocia if labor was protracted), further study is needed to increase the precision of this estimate and assess its clinical implications.
In addition to shoulder dystocia, peripartal infection was the only other complication, which was increased in women whose labor was arrested for 2 hours with at least 200 Montevideo units. Twenty-six percent of these women were diagnosed with chorioamnionitis, and an equal percentage with endometritis, compared with rates of 10% and 4%, respectively, among women in the remainder of the cohort. However, because 61% of these women were delivered vaginally, even in the unlikely event that prompt cesarean at the 2-hour mark would have prevented many or even all of their infections, more cesareans would have been performed than infections prevented.
Our data also demonstrate that oxytocin augmentation of labor is characterized by remarkable biologic variation. Overall rates of progressive labor and vaginal delivery did not vary appreciably by either the infusion rate of oxytocin or the uterine response to it (in terms of measured Montevideo units). Across parity groups, and irrespective of route of delivery, the median and range of oxytocin infusion rate and Montevideo units were broadly overlapping (Table 1). Furthermore, the achievement and maintenance of 200 Montevideo units, although a goal of this protocol1 and generally recognized as an adequate level of uterine activity,4,8 was poorly predictive of delivery route (Tables 2 and 3). Thus, in addition to criteria for uterine activity, guidelines for the management of dysfunctional labor must include criteria for the duration of augmentation across a range of uterine activity levels actually achieved.
From our data, we cannot establish the optimal maximum duration of augmented adequate labor with continued arrest, but we have shown that an intent to exceed the arbitrary bound of 2 hours was safe and effective in achieving vaginal delivery in the majority of women whose duration of arrest exceeded this bound.
1. Rouse DJ, Owen JO, Hauth JC. Active-phase labor arrest: Oxytocin augmentation for at least 4 hours. Obstet Gynecol 1999;93:323–8.
2. Friedman EA. Labor. Clinical evaluation and management. New York: Appleton-Century-Crofts, 1967.
3. Sharma SK, Sidawi JE, Ramin SM, Lucas MJ, Leveno KJ, Cunningham FG. A randomized trial of epidural versus patient-controlled meperidine analgesia during labor. Anesthesiology 1997;87:487–94.
4. American College of Obstetricians and Gynecologists. Dystocia and the augmentation of labor. ACOG technical bulletin no. 218. Washington, DC: American College of Obstetricians and Gynecologists, 1995.
5. American College of Obstetricians and Gynecologists. Prevention of early-onset group B streptococcal disease in newborns. ACOG committee opinion no. 173. Washington, DC: American College of Obstetricians and Gynecologists, 1996.
6. Brenner WE, Edelman DA, Hendricks CH. A standard of fetal growth for the United States of America. Am J Obstet Gynecol 1976;126:555–64.
7. American College of Obstetricians and Gynecologists. Inappropriate use of the terms fetal distress and birth asphyxia. ACOG committee opinion no. 197. Washington, DC: American College of Obstetricians and Gynecologists, 1998.
8. Hauth JC, Hankins GDV, Gilstrap LC, Strickland DM, Vance P. Uterine contraction pressures with oxytocin induction/augmentation. Obstet Gynecol 1986;68:305–9.