A prolonged third stage of labor has traditionally been defined as one lasting greater than 30 minutes. This definition is based on a 1991 report that demonstrated risks of maternal morbidities, including postpartum hemorrhage and the need for blood transfusion, began rising after duration of the third stage exceeded 30 minutes,1 yet both the modern obstetric population and standard obstetric practice have evolved since publication of this study. One very significant change in obstetric practice is the shift toward routine active management of the third stage after vaginal deliveries. This involves administering an uterotonic agent, clamping the cord early, and providing controlled cord traction until the placenta is delivered. These active management strategies have consistently been shown to decrease postpartum hemorrhage by 50–70%.2
Active management has also been shown to decrease the average duration of the third stage, although this has not been studied as rigorously.3,4 Recent studies suggest the risk of postpartum hemorrhage may increase earlier than the commonly referenced 30 minutes with some suggesting that risk increases as early as 10–15 minutes after a vaginal delivery.5,6 Together these data suggest that the definition of a prolonged third stage of labor in modern obstetrics needs to be reexamined.
The aim of this study was to characterize the duration of the third stage of labor in a modern obstetric cohort and examine the association between increasing duration and postpartum hemorrhage.
MATERIALS AND METHODS
This was a secondary analysis of a cohort of all term laboring women at Washington University Medical Center in St. Louis, Missouri, between April 2010 and August 2014. Washington University in St. Louis Human Research Protection Office approved this study. To be included in this study, patients were required to be in labor (either spontaneous or induced) at or beyond 37 weeks 0 days of gestation by best obstetric estimate. Women with multifetal gestations, preterm gestational ages at delivery, and cesarean deliveries were excluded.
Our institution uses the standard practice of active management of the third stage of labor; thus, it was used during the entire study period. The umbilical cord was clamped and cut immediately after the neonate was delivered. A segment of cord was then reclamped and cut for umbilical artery blood gas measurements. Oxytocin administration was started immediately after delivery of the neonate and the umbilical cord clamping. The oxytocin was delivered as either 30 units in 500 mL of lactated Ringer's solution through an intravenous bolus or as 10 units administered intramuscularly if the patient did not have intravenous access. Placentas generally were delivered by controlled cord traction with external fundal massage performed. These were both begun as soon as the cord was clamped and the cord gas segment was removed or after cord blood was collected for banking. Our institution does not use standard delay of cord clamping for term deliveries. However, if this practice was requested by a patient and used at the time of delivery, the cord clamping was delayed by approximately 30–45 seconds and traction was started as outlined previously. All women then received an additional intravenous infusion of 30 units of oxytocin in 1,000 mL of lactated Ringer's solution administered over 8 hours.
Trained obstetric research nurses abstracted detailed demographic information; obstetric, gynecologic, prenatal, medical, and surgical histories; antepartum history; and labor and delivery course. The third stage of labor duration was calculated in minutes as the time from neonatal delivery to placenta delivery. Postpartum hemorrhage was defined as an estimated blood loss of 500 mL or greater or if patients had a postpartum hemorrhage within the first 24 hours of delivery as documented by the treating obstetrician.
Baseline characteristics of the cohort were estimated for the entire cohort and compared between nulliparous and multiparous women using a two-sample Student's t test or Mann-Whitney U test for continuous variables and χ2 for categorical variables. Descriptive analysis of the third stage of labor duration was characterized by calculating the mean, median, interquartile range, 90th percentile, 95th percentile, and 99th percentile. Subsequently, these same parameters were described stratified by parity. Incidence of postpartum hemorrhage was calculated and compared by parity and then length of the third stage. First, comparisons were made between those with third-stage duration in the upper percentiles compared with those with shorter durations (upper 90th percentile was compared with 90th percentile or less, upper 95th percentile was compared with 95th percentile or less, upper 99th percentile was compared with 99th percentile or less). Next, to further delineate the threshold for increased postpartum hemorrhage risk, third-stage length was stratified by 5-minute intervals. Rates of postpartum hemorrhage were compared between each 5-minute interval group and the women who had shorter third-stage durations than the indicated group. Odds ratios (ORs) and 95% confidence intervals (CIs) were used to estimate the association between postpartum hemorrhage and third-stage duration. Multivariable logistic regression was used to adjust for parity, prolonged first stage, prolonged second stage, and induction of labor, factors previously shown to affect rates of postpartum hemorrhage or duration of the third stage.1,7,8 Model fit was assessed with the Hosmer-Lemeshow goodness of fit test.9
An a priori sample size calculation was not performed because the sample size was fixed and all women in the cohort and who met our inclusion criteria were included in these analyses. Statistical significance was defined as a P value of <.05. All analyses were completed using STATA 12.0.
There were 7,121 consecutive term deliveries meeting inclusion criteria in our cohort. Of these, 2,763 were nulliparous and 4,358 were multiparous (Table 1). Several baseline clinical characteristics differed by parity. On average, nulliparous women were younger, less likely to be obese, and had a higher incidence of prolonged second stage of labor and chorioamnionitis. Nulliparous women were also more likely to have an induction of labor, an augmentation of labor, or an operative vaginal delivery.
There was no significant difference in duration of the third stage of labor between vaginal and operative vaginal deliveries (5.46 minutes compared with 5.44 minutes, P=.94). When further analyses were performed with vaginal deliveries and operative vaginal deliveries combined, there was no significant difference in duration of the third stage for nulliparous compared with multiparous women (Table 2). Within the total cohort, 90% of women had a third stage lasting 9 minutes or less, 95% had one lasting 13 minutes or less, and 99% of women completed placenta delivery by 28 minutes.
Compared with multiparous women, nulliparous women were more likely to experience a postpartum hemorrhage (12.5% compared with 6.4%, P<.001). Women with increasing duration of the third stage of labor after vaginal delivery had increasing risk for postpartum hemorrhage. Those with a third stage greater than the 90th percentile (greater than 9 minutes) compared with those with a third stage duration 90th percentile or less (9 minutes or less) had an increased risk for postpartum hemorrhage (13.2% compared with 8.3%; OR 1.68, 95% CI 1.33–2.12; Table 3). This association remained significant after adjusting for labor induction, prolonged first or second stages, and parity (adjusted OR 1.82, 95% CI 1.43–2.31; Table 3). To further delineate the threshold for increased postpartum hemorrhage risk, we stratified the third-stage duration into 5-minute intervals (Fig. 1). This divided our cohort into seven groups, five of which contained the women in the top 90th percentile (greater than 9 minutes' duration). A significant rise in postpartum hemorrhage risk was noted starting at 20 minutes (15.9% at 20–24 minutes compared with 8.5% at less than 20 minutes, adjusted OR 2.38, 95% CI 1.18–4.79; Fig. 1). This risk continued to increase and patients with a third-stage duration of 30 minutes or greater had a postpartum hemorrhage risk of 35.1%. There was no significant association between the need for blood transfusion and duration of the third stage of labor (1.0% compared with 0.84% for third-stage duration greater than 90th percentile compared with 90th percentile or less, adjusted OR 1.18, 95% CI 0.53–2.60; Table 3).
After vaginal delivery, 54 placentas (0.76%) were manually extracted and 40 of these were extracted before 30 minutes' duration. Of the placentas that were manually extracted before 30 minutes, 17 were not associated with a postpartum hemorrhage. Because this could introduce additional bias, as a result of premature termination of the third stage, we performed two additional analyses. To evaluate the potential bias on the length of the third stage, we made the assumption that all 17 of the prematurely extracted placentas were left in place until the classically recommended 30 minutes. This assumption did not significantly alter the mean duration of the third stage (5.51 minutes compared with 5.46 minutes, P=.60). The 90th percentile was slightly higher at 10 minutes (compared with 9 minutes), 95th percentile was 14 minutes (compared with 13 minutes), and 99th percentile was 30 minutes (compared with 28 minutes). To address the risk of bias on the calculations assessing risk for morbidity, we performed a sensitivity analysis excluding these patients. The association between postpartum hemorrhage and duration of the third stage did not change if these cases of early placental extraction were removed from analysis.
We found that the third stage of labor after a vaginal delivery is shorter than historically described. In their 1991 study, Combs et al reported that 75% of placentas were delivered by 10 minutes, whereas 90% of our contemporary obstetric cohort had delivery of the placental by this time. Other recent reports also support this finding.5,7 We posit that active management of the third stage is the most important contributor to this decreased duration. Although most trials looking at active management of the third stage aimed to determine its effects on postpartum hemorrhage, some also documented a significant decrease in third-stage duration.2 Therefore, in the setting of modern active third-stage management after a vaginal delivery, the definition of prolonged third stage and recommendations for intervention timing should be reconsidered.
The exact threshold for abnormal length of the third stage is unclear. The 95th percentile is often used statistically as the upper limit of normal. Based on our data, this cutpoint would suggest an abnormally prolonged third stage if it lasts greater than 13 minutes. However, arbitrary cutpoints based on population percentiles do not consider clinical sequelae. An alternative method would be to base the definition on risk of maternal morbidity, which is how the 30-minute upper limit was originally designated.1 These data show that the risk for postpartum hemorrhage nearly doubles by the time third-stage duration reaches 20 minutes. Additionally, the proportion of women experiencing a postpartum hemorrhage rose steadily with increasing third-stage duration, from 8.5% among those with a stage of 9 minutes or less to 15.9% at 20–24 minutes and as high at 35.1% when the third stage was 30 minutes or greater. At 20 minutes, the risk for postpartum hemorrhage was almost doubled compared with those who delivered the placenta within the first 9 minutes (durations encompassing the lower 90%).
Some epidemiologic studies have suggested that although rates of maternal death resulting from postpartum hemorrhage have been steadily dropping, the rate of postpartum hemorrhage in high-resource countries has been slowly increasing.10 The etiology of this rise is unclear. Another study previously used a receiver operator curve to determine the best duration cutoff to predict a postpartum hemorrhage in the setting of active management of the third stage.5 The study reported 18 minutes to be the duration most predictive of an impending postpartum hemorrhage with a specificity of 90%. However, the sensitivity was 31% and area under the curve was 0.60, indicating very poor predictability of the model and confirming that although prolonged duration of the third stage is associated with increased risk for postpartum hemorrhage, there are likely other contributory factors in its etiology. Alternatively, population changes may be responsible for the rise, as we have seen increases in maternal obesity and age, which are known risk factors of postpartum hemorrhage.10,11 These trends serve as reminders that despite our vast improvements in postpartum hemorrhage prevention and treatment, it continues to be a significant source of maternal morbidity and mortality that requires educated health care providers who are able to prevent, recognize, and treat postpartum hemorrhage.
There are limited data to guide interventions for a prolonged third stage of labor. Manual extraction of the placenta is commonly performed; however, there is no evidence to suggest whether this intervention decreases volume of blood loss. In fact, in the setting of cesarean deliveries, manual extraction was repeatedly shown to be associated with higher estimated blood loss and increased rates of endometritis.12,13 We noted that although rates of manual placental extraction after a vaginal delivery remain low in our cohort (0.76%), a high proportion of these extractions were performed before 30 minutes' duration and were not associated with a postpartum hemorrhage. This suggests that although the current analysis provides support for earlier intervention, obstetric practitioners may already be moving in that direction. Perhaps more important than manual extraction of a placenta, we need to be prepared for management of postpartum hemorrhage in the setting of a third stage that begins to approach 20 minutes.
Some strengths of this study are its large sample size and detailed data. This was a secondary analysis, which brings some inherent weaknesses such as a predefined sample size and only previously collected data, unable to be tailored to our specific question. This study was also limited by exclusion of preterm deliveries, which have previously been shown to have a prolonged third stage of labor.6 Delayed postpartum hemorrhage (within the first 24 hours after delivery) may be underrepresented in the cohort if it was not adequately recorded by the physician in the electronic medical record. Lastly, estimation of blood loss during deliveries was by the delivering physician. Although visual estimation of blood loss is the most commonly clinically used method for estimating blood loss, it has been shown to underestimate amount of blood loss and thus our rate of postpartum hemorrhage may also be underestimated. However, it is reassuring that recent work from our institution showed that this estimate was correlated with morbidity.14
In conclusion, we found that duration of the third stage of labor is shorter in a modern obstetric cohort relative to historically reported norms. Among term vaginal deliveries, duration of the third stage of labor of greater than 20 minutes is associated with increased risk of postpartum hemorrhage. This suggests that heightened awareness and preparation for the increased risk of postpartum hemorrhage should be considered earlier in the third stage of labor.
1. Combs CA, Murphy EL, Laros RK Jr. Factors associated with postpartum hemorrhage with vaginal birth. Obstet Gynecol 1991;77:69–76.
2. Begley CM, Gyte GM, Devane D, McGuire W, Weeks A. Active versus expectant management for women in the third stage of labour. The Cochrane Database of Systematic Reviews 2015, Issue 3. Art. No.: CD007412. DOI: 10.1002/14651858.CD007412.pub3.
3. Begley CM. A comparison of “active” and “physiological” management of the third stage of labour. Midwifery 1990;6:3–17.
4. Prendiville WJ, Harding JE, Elbourne DR, Stirrat GM. The Bristol third stage trial: active versus physiological management of third stage of labour. BMJ 1988;297:1295–300.
5. Magann EF, Evans S, Chauhan SP, Lanneau G, Fisk AD, Morrison JC. The length of the third stage of labor and the risk of postpartum hemorrhage. Obstet Gynecol 2005;105:290–3.
6. 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.
7. Magann EF, Doherty DA, Briery CM, Niederhauser A, Chauhan SP, Morrison JC. Obstetric characteristics for a prolonged third stage of labor and risk for postpartum hemorrhage. Gynecol Obstet Invest 2008;65:201–5.
8. Harper LM, Caughey AB, Roehl KA, Odibo AO, Cahill AG. Defining an abnormal first stage of labor based on maternal and neonatal outcomes. Am J Obstet Gynecol 2014;210:536.e1–7.
9. Hosmer DW, Lemeshow S, Sturdivant R. Applied logistic regression. 3rd ed. New York (NY): Wiley; 2013.
10. Knight M, Callaghan WM, Berg C, Alexander S, Bouvier-Colle MH, Ford JB, et al.. Trends in postpartum hemorrhage in high resource countries: a review and recommendations from the International Postpartum Hemorrhage Collaborative Group. BMC Pregnancy Childbirth 2009;9:55.
11. Weeks A. The prevention and treatment of postpartum haemorrhage: what do we know, and where do we go to next? BJOG 2015;122:202–10.
12. Atkinson MW, Owen J, Wren A, Hauth JC. The effect of manual removal of the placenta on post-cesarean endometritis. Obstet Gynecol 1996;87:99–102.
13. McCurdy CM Jr, Magann EF, McCurdy CJ, Saltzman AK. The effect of placental management at cesarean delivery on operative blood loss. Am J Obstet Gynecol 1992;167:1363–7.
© 2016 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
14. Conner SN, Tuuli MG, Colvin R, Shanks AL, Macones GA, Cahill AG. Accuracy of Estimated Blood Loss in Predicting Need for Transfusion after Delivery. Am J Perinatol 2015;32:1225–30.