Delivery of the placenta is usually a very rapid event after vaginal delivery of the neonate. In an investigation by Dombrowski,1 50% of the placental deliveries occurred within 5 minutes, and 90% had been delivered by 15 minutes. Other large investigations have confirmed the rapid delivery of the placenta, with a mean delivery time of 8.3 minutes in the World Health Organization study2 and 9 minutes in another study.3 Failure of the placenta to deliver spontaneously is an important cause of postpartum hemorrhage.1,4 Despite the observation that failure of the placenta to deliver promptly can be linked with postpartum hemorrhage, there are few investigations that have assessed the morbidity associated with a prolonged third stage of labor. A study by Combs4 observed that the risks for hemorrhage, blood transfusion, and a dilation and curettage were constant for a third stage of labor lasting less than 30 minutes, but rose progressively after 30 minutes. Hemorrhage frequency peaked at a third stage of labor at 40 minutes in another investigation.1 This has resulted in the recommendations for manual removal of the placenta when it has been retained for longer than 30 minutes after delivery.1 However, the practice of removing placentas that have not been spontaneously expelled within 30 minutes of the delivery of the neonate has been poorly evaluated. The purpose of this investigation was to estimate the optimal time after delivery for removal of the placenta to prevent postpartum hemorrhage.
MATERIALS AND METHODS
From July 1, 2000, to June 30, 2002, the blood loss at all vaginal deliveries was measured and recorded at a single tertiary obstetric hospital. All deliveries were managed actively (the administration of oxytocin with the delivery of the anterior fetal shoulder). The third stage of labor was defined as the period of time from birth of the neonate until the delivery of the placenta. If the placenta was not spontaneously expelled, the length of the third stage of labor was calculated from the delivery of the neonate until the placenta was manually removed. Placentas were manually extracted if they had not spontaneously delivered after 30 minutes following the delivery of the neonate. The vast majority of the women had epidural anesthesia, and the placental was manually extracted under epidural anesthesia; other extractions were done under intravenous sedation or removed under conduction of general anesthesia in the adjacent operating room, depending on the clinical circumstances. This study was initiated with the consent and the approval of the hospital and ethics committee to assess the rate of postpartum hemorrhage and compare it with hospitals delivering similar numbers of patients with comparable maternal demographics.
The hospital was evaluating their incidence of postpartum hemorrhage and for that reason was meticulously assessing the blood loss at each vaginal delivery. The blood lost at each delivery was calculated by the delivering midwife by collecting and measuring the blood in collection devices used specifically for vaginal deliveries and then weighing sheets and drapes used at the delivery. All deliveries were done by midwives, except for operative vaginal or abdominal deliveries. Nevertheless, the midwife was still present at the operative vaginal deliveries and measured the blood loss. Postpartum hemorrhage was defined as blood loss in excess of 1,000 mL or need for a red blood cell transfusion because of maternal anemia or hemodynamic instability. The practice of classifying excessive blood loss as postpartum hemorrhage by the drop in the hematocrit was not used in this study because not all of the women had both predelivery and postdelivery hematocrit. All vaginal deliveries beyond 20 weeks of gestation were evaluated.
Logistic regression analysis was used to construct a receiver operating characteristics curve to determine the optimal cutoff point for the length of the third stage of labor to predict a postpartum hemorrhage, by simultaneously maximizing the specificity with the sensitivity. The area under the receiver operating characteristic (ROC) curve, which can be used as a measure of the accuracy of the test, was also calculated by a point-to-point trapezoidal method of integration.
Between July 1, 2000, and June 30, 2002, there were 6,588 vaginal deliveries at this tertiary obstetric hospital. postpartum hemorrhage occurred in 335 (5.1%) of the vaginal deliveries. The women delivering vaginally were aged 28.1 ± 6 years, 69.2% white, 10.5% African American, 10.2% Asian, and 10.05% were of other races. Most of the women were in their first (28.7%) or second pregnancies (26.5%), with the remainder in their third or greater pregnancy (44.8%). Nearly one half of the women were primiparous (43.6%). The median gestational age at the time of delivery was 39 weeks, with an interquartile range from 37 to 40 weeks. Eighty-three percent of the deliveries were spontaneous, 10.6% were vacuum, 2.6% used forceps, and 0.8% were unsuccessful vacuum applications followed by forceps deliveries, whereas 2.9% were vaginal breech births.
The median length of the third stage of labor was similar in women without a postpartum hemorrhage (7 minutes) and in women with a postpartum hemorrhage (9 minutes), but the third quartile is much higher for the latter (30 minutes compared with 11 minutes), and the difference became more marked at higher percentiles (for 90%, 1 hour 48 minutes compared with 18 minutes). The minimal time was 2 minutes for both groups, and the maximum was 4 hours for the postpartum hemorrhage group and 2 hours for the nonpostpartum hemorrhage group. For third stages of labor more than 10 minutes compared with less than 10 minutes there was a greater risk of postpartum hemorrhage (OR 2.1, 95% CI 1.6–.6, P < .001). For a third stage of labor more than 20 minutes compared with less than 20 minutes there was a greater risk of postpartum hemorrhage (OR 4.3, 95% CI 3.3–5.5, P < .001). Similarly for deliveries more than 30 minutes compared with deliveries less than 30 minutes there was a greater risk of postpartum hemorrhage (OR 6.2, 95% CI 4.6–8.2, P < .001). The ROC curve determined that 18 minutes, with a sensitivity of 31% and a specificity of 90%, was the best cutoff for the prediction of a postpartum hemorrhage. (Fig. 1). Logistic regression analysis showed several factors that also affected the prediction of postpartum hemorrhage. Although multivariate analysis would simultaneously consider a number of risk factors for postpartum hemorrhage, the complexity of this analysis would limit the clinical usefulness of such information. Its use would require a complex assessment rather than a decision based on time alone, which is easily measured. The area under the ROC curve using length of stage 3 alone was 0.60, and this would have been increased to 0.71 with all factors included.
Traditionally, a third stage of greater than 30 minutes has been considered abnormal and has warranted manual extraction of the placenta even though there is scant evidence to support intervention at 30 minutes.4 Investigations of the length of the third stage of labor have primarily assessed an active compared with an expectant management5–7 or a comparison of different active management plans.5,8 Active management of the third stage of labor involves the administration of an oxytocin or other ecbolic agent to effect uterine contractions and placental expulsion as compared with allowing the placenta to deliver spontaneously. Active management of the third stage of labor is associated with less blood loss;5,6 however, it has more unpleasant side effects such as nausea and pain as well as potential increased blood pressure when ergotamines are used.9 The average length of the third stage of labor was prolonged in the passive compared with the active management group5 and also in those whose third stage of labor exceeded 30 minutes.6 Despite the best strategies to prevent blood loss at a vaginal delivery, approximately 3% of all deliveries will still lose in excess of 1,000 mL.2,10
All of the women in this investigation were actively managed with the administration of oxytocin upon delivery of the anterior shoulder. This active management strategy not only reduces the amount of blood loss at vaginal delivery5,6 but also decreases the number of women whose placenta remains undelivered 30 minutes after the delivery.6 Nevertheless, the observed risk of postpartum hemorrhage in this patient population (5.1%) is similar to what is reported in the literature for all vaginal deliveries (3.9%).11
This study is noteworthy for the number of consecutive vaginal deliveries assessed in a single institution, the meticulous evaluation of blood loss at each delivery, the rigid criteria for the diagnosis of a postpartum hemorrhage, active management of all third stage labor, and the similarity of management techniques used for the second and third stage of labor. This study evaluated the length of the third stage of labor by ROC curves for an optimal management time of the third stage to prevent postpartum hemorrhage. Other investigators4 have relied on fixed cutoffs such as 30 minutes and have compared the risk of postpartum hemorrhage in those pregnancies less than and more than 30 minutes rather than determining whether that 30 minute cutoff is the optimal interval for placental delivery to prevent a postpartum hemorrhage.
This investigation also raises additional questions about blood loss in the third stage of labor. What is the contribution of the method of placental removal on postpartum hemorrhage considering that manual removal is linked with greater blood loss than spontaneous removal12? Is the time from neonatal to placental delivery the primary determinant of total blood loss or a combination of length of the third stage of labor and method of placental removal or some additional unique condition of the uterus or pregnancy that influences the risk of excessive postpartum blood loss?
This observational study strongly suggests that a placenta that has not been delivered by 18 minutes should be removed to reduce the incidence of postpartum hemorrhage. The investigation is limited, as are other studies, by retrospectively evaluating the length of the third stage of labor and determining the interval for placental delivery to minimize the number of placentas manually extracted compared with the risk of a postpartum hemorrhage. The best evidence to confirm the optimal time for placental delivery would be a prospective trial. Women, in whom the placenta had not delivered in a set period of time, could be randomly assigned to placental removal or no intervention, and the rate of postpartum hemorrhage could be calculated. We would encourage further investigations to identify modifiable risk factors and techniques to decrease the risk of postpartum hemorrhage, because it remains as one of the leading causes of maternal death in both developed and nondeveloped countries.13
1. Dombrowski MP, Bottoms SF, Saleh AA, Hurd WW, Romero R. Third stage of labor: analysis of duration and clinical practice. Am J Obstet Gynecol 1995;172:1279–84.
2. Gulmezoglu AM, Villar J, Ngoc NN, Piaggio G, Carroli G, Adetoro L, et al.; WHO Collaborative Group To Evaluate Misoprostol in the Management of the Third Stage of Labour. WHO multicentre randomised trial of misoprostol in the management of the third stage of labour. Lancet 2001;358:689–95.
3. Bugalho A, Daniel A, Faundes A, Cunha M. Misoprostol for prevention of postpartum hemorrhage. Int J Gynaecol Obstet 2001;73:1–6.
4. Combs CA, Laros RK. Prolonged third stage of labor: morbidity and risk factors. Obstet Gynecol 1991;77:863–7.
5. Prendiville WJ, Elbourne D, McDonald S. Active versus expectant management in the third stage of labour (Cochrane Review). In: The Cochrane Library, Issue 3, 2000. Oxford: Update Software.
6. Rogers J, Wood J, McCandlish R, Ayers S, Truesdale A, Elbourne D. Active versus expectant management of the third stage of labour: the Hinchingbrooke randomized controlled trial. Lancet 1998;351:693–99.
7. Jackson KW, Allbert JR, Schemmer GK, Elliot M, Humphrey A, Taylor J. A randomized controlled trial comparing oxytocin administration before and after placental delivery in the prevention of postpartum hemorrhage. Am J Obstet Gynecol 2001;185:873–7.
8. Gerstenfeld TS, Wing DA. Rectal misoprostol versus intravenous oxytocin for the prevention of postpartum hemorrhage after vaginal delivery. Am J Obstet Gynecol 2001;185:878–82.
9. McDonald S, Prendiville WJ, Elbourne D. Prophylactic syntometrine versus oxytocin for delivery of the placenta. (Cochrane Review). In: The Cochrane Library, Issue 2, 2000. Oxford: Update Software.
10. Kundodya TW, Majoka F, Rusqkaniko S. Misoprostol versus oxytocin in the third stage of labor. Int J Gynaecol Obstet 2001;75:235–41.
11. Combs CA, Murphy EL, Laros RK. Factors associated with postpartum hemorrhage with vaginal birth. Obstet Gynecol 1991;77:69–76.
12. Magann EF, Dodson MR, Albert JR, McCurdy CM Jr., Martin RW, Morrison JC. Blood loss at time of cesarean section by method of placental removal and exteriorization versus in situ repair of the uterine incision. Surg Gynecol Obstet 1993;177:389–92.
13. Revised 1990 estimates of maternal mortality: a new approach by WHO and UNICEF. Geneva: World Health Organization; 2000.