Professional Documents
Culture Documents
OBJECTIVE: To characterize the duration of the third third-stage durations (15.9% compared with 8.5%;
stage of labor and the association with postpartum adjusted OR 2.38, 95% CI 1.18–4.79). However, blood
hemorrhage in a contemporary cohort. transfusion was not associated with third-stage duration
METHODS: We performed a secondary analysis of (1.0% compared with 0.84% for third-stage duration
a cohort of 7,121 women who had a vaginal delivery at greater than 90th compared with 90th percentile or less,
or beyond 37 weeks 0 days of gestation at a single adjusted OR 1.18, 95% CI 0.53–2.60).
tertiary care center from April 2010 to August 2014. CONCLUSION: Our data show that postpartum hem-
Active management of the third stage of labor was orrhage risk increases significantly when the third stage
routinely used during the study period. The mean, of labor duration is 20 minutes or more, suggesting that
median, interquartile range, 90th percentile, 95th per- the definition of a prolonged third stage of labor being
centile, and 99th percentile of the third stage of labor 30 minutes or more may be outdated.
duration were calculated. Odds ratios were calculated to (Obstet Gynecol 2016;127:951–6)
estimate the association between increased duration of DOI: 10.1097/AOG.0000000000001399
third stage of labor and incidence of postpartum hem-
A
orrhage. prolonged third stage of labor has traditionally
RESULTS: The mean duration of the third stage of labor been defined as one lasting greater than 30 minutes.
among women who had a vaginal delivery was 5.46 This definition is based on a 1991 report that demon-
(standard deviation 5.4) minutes and median duration strated risks of maternal morbidities, including postpar-
was 4 minutes. The 90th, 95th, and 99th percentiles were tum hemorrhage and the need for blood transfusion,
defined by 9, 13, and 28 minutes, respectively. Women began rising after duration of the third stage exceeded
with a third stage above the 90th percentile (n5705) had
30 minutes,1 yet both the modern obstetric population
an increased risk for postpartum hemorrhage compared
and standard obstetric practice have evolved since pub-
with a third stage below the 90th percentile (13.2% com-
lication of this study. One very significant change in
pared with 8.3%; adjusted odds ratio [OR] 1.82, 95%
obstetric practice is the shift toward routine active man-
confidence interval [CI] 1.43–2.31). When the 90th per-
centile was further subdivided into 5-minute increments,
agement of the third stage after vaginal deliveries. This
risk for postpartum hemorrhage significantly increased involves administering an uterotonic agent, clamping
beginning at 20–24 minutes compared with shorter the cord early, and providing controlled cord traction
until the placenta is delivered. These active manage-
ment strategies have consistently been shown to
From the Department of Obstetrics and Gynecology, Washington University
School of Medicine, St. Louis, Missouri. decrease postpartum hemorrhage by 50–70%.2
Dr. Cahill is supported by the Eunice Kennedy Shriver National Institute of
Active management has also been shown to
Child Health and Human Development (R01HD061619-01, Principal Inves- decrease the average duration of the third stage,
tigator Cahill), which partially supported this work. although this has not been studied as rigorously.3,4
Corresponding author: Antonina I. Frolova, MD, PhD, Department of Obstetrics Recent studies suggest the risk of postpartum hemor-
and Gynecology, Washington University School of Medicine, 660 S Euclid rhage may increase earlier than the commonly refer-
Avenue, Box 8064, St. Louis, MO 63110; e-mail: frolovaa@wudosis.wustl.edu.
enced 30 minutes with some suggesting that risk
Financial Disclosure
The authors did not report any potential conflicts of interest. increases as early as 10–15 minutes after a vaginal
© 2016 by The American College of Obstetricians and Gynecologists. Published
delivery.5,6 Together these data suggest that the defi-
by Wolters Kluwer Health, Inc. All rights reserved. nition of a prolonged third stage of labor in modern
ISSN: 0029-7844/16 obstetrics needs to be reexamined.
an induction of labor, an augmentation of labor, or an compared with those with a third stage duration 90th
operative vaginal delivery. percentile or less (9 minutes or less) had an increased
There was no significant difference in duration of risk for postpartum hemorrhage (13.2% compared with
the third stage of labor between vaginal and operative 8.3%; OR 1.68, 95% CI 1.33–2.12; Table 3). This asso-
vaginal deliveries (5.46 minutes compared with ciation remained significant after adjusting for labor
5.44 minutes, P5.94). When further analyses were induction, prolonged first or second stages, and parity
performed with vaginal deliveries and operative vag- (adjusted OR 1.82, 95% CI 1.43–2.31; Table 3). To
inal deliveries combined, there was no significant dif- further delineate the threshold for increased postpartum
ference in duration of the third stage for nulliparous hemorrhage risk, we stratified the third-stage duration
compared with multiparous women (Table 2). Within into 5-minute intervals (Fig. 1). This divided our cohort
the total cohort, 90% of women had a third stage into seven groups, five of which contained the women in
lasting 9 minutes or less, 95% had one lasting 13 mi- the top 90th percentile (greater than 9 minutes’ dura-
nutes or less, and 99% of women completed placenta tion). A significant rise in postpartum hemorrhage risk
delivery by 28 minutes. was noted starting at 20 minutes (15.9% at 20–24 minutes
Compared with multiparous women, nulliparous compared with 8.5% at less than 20 minutes, adjusted
women were more likely to experience a postpartum OR 2.38, 95% CI 1.18–4.79; Fig. 1). This risk continued
hemorrhage (12.5% compared with 6.4%, P,.001). to increase and patients with a third-stage duration of
Women with increasing duration of the third stage of 30 minutes or greater had a postpartum hemorrhage risk
labor after vaginal delivery had increasing risk for of 35.1%. There was no significant association between
postpartum hemorrhage. Those with a third stage the need for blood transfusion and duration of the third
greater than the 90th percentile (greater than 9 minutes) stage of labor (1.0% compared with 0.84% for third-stage
VOL. 127, NO. 5, MAY 2016 Frolova et al Third Stage of Labor 953
duration greater than 90th percentile compared with 13 minutes), and 99th percentile was 30 minutes (com-
90th percentile or less, adjusted OR 1.18, 95% CI pared with 28 minutes). To address the risk of bias on
0.53–2.60; Table 3). the calculations assessing risk for morbidity, we per-
After vaginal delivery, 54 placentas (0.76%) were formed a sensitivity analysis excluding these patients.
manually extracted and 40 of these were extracted The association between postpartum hemorrhage and
before 30 minutes’ duration. Of the placentas that were duration of the third stage did not change if these cases
manually extracted before 30 minutes, 17 were not of early placental extraction were removed from
associated with a postpartum hemorrhage. Because this analysis.
could introduce additional bias, as a result of premature
termination of the third stage, we performed two addi- DISCUSSION
tional analyses. To evaluate the potential bias on the We found that the third stage of labor after a vaginal
length of the third stage, we made the assumption that delivery is shorter than historically described. In their
all 17 of the prematurely extracted placentas were left 1991 study, Combs et al reported that 75% of
in place until the classically recommended 30 minutes. placentas were delivered by 10 minutes, whereas
This assumption did not significantly alter the mean 90% of our contemporary obstetric cohort had
duration of the third stage (5.51 minutes compared delivery of the placental by this time. Other recent
with 5.46 minutes, P5.60). The 90th percentile was reports also support this finding.5,7 We posit that
slightly higher at 10 minutes (compared with 9 mi- active management of the third stage is the most
nutes), 95th percentile was 14 minutes (compared with important contributor to this decreased duration.
Postpartum hemorrhage
All 626/7,121 (8.8)
90th %tile (greater than 9 min) 93/705 (13.2) 1.68 (1.33–2.12) 1.82 (1.43–2.31)
95th %tile (greater than 13 min) 57/356 (16.0) 2.08 (1.54–2.79) 2.22 (1.64–3.01)
99th %tile (greater than 28 min) 20/68 (29.4) 4.43 (2.61–7.52) 4.61 (2.68–7.93)
Greater than 30 min 15/44 (34.1) 5.47 (2.92–10.3) 5.94 (3.12–11.3)
Transfusion
All 61/7,121 (0.9)
90th %tile (greater than 9 min) 7/705 (1.0) 1.18 (0.54–2.61) 1.18 (0.53–2.60)
95th %tile (greater than 13 min) 2/356 (0.6) 0.64 (0.16–2.64) 0.62 (0.15–2.57)
99th %tile (greater than 28 min) 1/68 (1.5) 1.74 (0.24–12.7) 1.66 (0.23–12.2)
Greater than 30 minutes 1/44 (2.3) 2.72 (0.37–20.1) 2.83 (0.38–21.0)
OR, odds ratio; CI, confidence interval.
Bold indicates statistically significant result.
* Rate presented as number of events within indicated percentile, n/N (%).
†
Reference group for ORs includes all cases with third-stage durations below the indicated percentile.
‡
Adjusted for parity, induction, prolonged first stage, prolonged second stage.
Although most trials looking at active management of countries has been slowly increasing.10 The etiology
the third stage aimed to determine its effects on post- of this rise is unclear. Another study previously used
partum hemorrhage, some also documented a signifi- a receiver operator curve to determine the best dura-
cant decrease in third-stage duration.2 Therefore, in tion cutoff to predict a postpartum hemorrhage in the
the setting of modern active third-stage management setting of active management of the third stage.5 The
after a vaginal delivery, the definition of prolonged study reported 18 minutes to be the duration most
third stage and recommendations for intervention tim- predictive of an impending postpartum hemorrhage
ing should be reconsidered. with a specificity of 90%. However, the sensitivity was
The exact threshold for abnormal length of the 31% and area under the curve was 0.60, indicating
third stage is unclear. The 95th percentile is often used very poor predictability of the model and confirming
statistically as the upper limit of normal. Based on our that although prolonged duration of the third stage is
data, this cutpoint would suggest an abnormally associated with increased risk for postpartum hemor-
prolonged third stage if it lasts greater than 13 minutes. rhage, there are likely other contributory factors in its
However, arbitrary cutpoints based on population etiology. Alternatively, population changes may be
percentiles do not consider clinical sequelae. An responsible for the rise, as we have seen increases in
alternative method would be to base the definition maternal obesity and age, which are known risk fac-
on risk of maternal morbidity, which is how the 30- tors of postpartum hemorrhage.10,11 These trends
minute upper limit was originally designated.1 These serve as reminders that despite our vast improvements
data show that the risk for postpartum hemorrhage in postpartum hemorrhage prevention and treatment,
nearly doubles by the time third-stage duration rea- it continues to be a significant source of maternal
ches 20 minutes. Additionally, the proportion of morbidity and mortality that requires educated health
women experiencing a postpartum hemorrhage rose care providers who are able to prevent, recognize, and
steadily with increasing third-stage duration, from treat postpartum hemorrhage.
8.5% among those with a stage of 9 minutes or less There are limited data to guide interventions for
to 15.9% at 20–24 minutes and as high at 35.1% when a prolonged third stage of labor. Manual extraction
the third stage was 30 minutes or greater. At 20 mi- of the placenta is commonly performed; however,
nutes, the risk for postpartum hemorrhage was almost there is no evidence to suggest whether this inter-
doubled compared with those who delivered the pla- vention decreases volume of blood loss. In fact, in
centa within the first 9 minutes (durations encompass- the setting of cesarean deliveries, manual extraction
ing the lower 90%). was repeatedly shown to be associated with higher
Some epidemiologic studies have suggested that estimated blood loss and increased rates of endome-
although rates of maternal death resulting from post- tritis.12,13 We noted that although rates of manual
partum hemorrhage have been steadily dropping, the placental extraction after a vaginal delivery remain
rate of postpartum hemorrhage in high-resource low in our cohort (0.76%), a high proportion of these
VOL. 127, NO. 5, MAY 2016 Frolova et al Third Stage of Labor 955