Obstetric hemorrhage is the leading cause of maternal death worldwide and is among the top three causes of these deaths in the United States.1,2 Postpartum hemorrhage is the most common type of obstetric hemorrhage, and uterine atony accounts for more than 80% of postpartum hemorrhage.3 Prophylactic use of uterotonic agents prevents uterine atony and reduces the risk of postpartum hemorrhage by 40–50%.4 – 7 Compared with methylergometrine and misoprostol, oxytocin has a good safety profile and induces fewer, if any, side effects.6,8,9 In the United States, oxytocin is the uterotonic routinely used for prophylaxis. Despite its widespread use, the optimal prophylactic oxytocin dose regimen is unknown. Twenty units of oxytocin administered in 1 L of crystalloid solution “at a rate of 10 mL per minute for a few minutes to get an adequate uterine tone, then reduced to 1–2 mL per minute during postpartum recovery in the delivery suite and then discontinued before transfer to the postpartum suite” is a commonly recommended dose regimen.10 The dose regimen corresponds to that routinely used after vaginal delivery at our institution: 10 units of oxytocin in 500 mL of crystalloid solution administered over 1 hour. For cesarean delivery, a higher dose regimen (80 units oxytocin in 500 mL of crystalloid) is used at our institution based on positive findings from a randomized trial that included 321 women who underwent cesarean delivery.11 In that study, compared with women who received the higher dose regimen, the standard dose (10 units in 500 mL) was associated with a twofold increase in the risk of uterine atony or postpartum hemorrhage requiring treatment with uterotonics (including additional oxytocin) and nearly a fivefold increase in the need for second-line uterotonics such as methergine and hemabate.11 Although vaginal deliveries account for more than two thirds of all births,12 it remains unknown whether a higher dose of prophylactic oxytocin is similarly more effective among women who deliver vaginally. If this were so, we could have a single, dedicated oxytocin-dose concentration to prevent postpartum hemorrhage. Therefore, we compared the effectiveness of two higher-dose prophylactic oxytocin regimens to the standard dose regimen for vaginal deliveries. We hypothesized that higher doses of oxytocin (80 units or 40 units) as compared with the standard 10-unit dose would safely reduce uterine atony or postpartum hemorrhage requiring treatment.
MATERIALS AND METHODS
We conducted a single-center, double-blind, randomized controlled trial that included women with viable pregnancies undergoing vaginal delivery at 24 weeks of gestation or more at University Hospital, Birmingham, Alabama. Those who underwent cesarean delivery or who had a fetal demise, a diagnosis of coagulopathy, pulmonary edema, or cardiomyopathy were excluded. Eligible women were approached and consented at the time of admission for delivery (spontaneous labor or induction). The Institutional Review Board of the University of Alabama at Birmingham approved the study.
Women were randomized to one of the three study arms according to a confidential computer-generated block randomization algorithm. The algorithm randomly allocated three women to each study dose in blocks of nine, thus ensuring equal allocation among the study arms. The randomization scheme was sequentially numbered and delivered to the pharmacy. The investigational drug pharmacists prepared identical oxytocin bags by adding 10 units, 40 units, or 80 units of oxytocin into a malleable bag of 500 mL of normal saline. The bags were prepared in advance of patient randomization, were numbered according to the randomization scheme provided, and were stored at room temperature in a dedicated and secure research study closet on the labor and delivery unit. Only the investigational pharmacist and one statistician, who had no role in patient enrollment or outcome ascertainment, had knowledge of the code matching the sequential number to oxytocin dose or the size of the randomized blocks. At the time of vaginal delivery of each consented patient, the next sequentially numbered oxytocin bag was dispensed to the nurse. The sequential drug number together with the patient's name and medical record number were entered into the randomization log. At this point, the patient was considered randomized. On delivery of the placenta, the study medication bag was then administered to the patient over 1 hour using an infusion pump for precision. During this hour, use of additional oxytocin to treat uterine atony or hemorrhage was avoided (second-line uterotonics such as hemabate or methergine were used). However, additional oxytocin could be utilized for treatment after completion of the prophylactic infusion (ie, after 1 hour).
The primary outcome was a composite outcome of uterine atony or hemorrhage requiring treatment, including treatment with any uterotonic, uterine tamponade (typically with a Bakri balloon), interventional radiology for uterine or other arterial embolization, surgery, or blood transfusion. Key prespecified secondary outcomes included individual outcomes in the primary composite, a decline of 6% unit or more in hematocrit after delivery, clinically estimated blood loss, endometritis, hospital stay, and safety outcomes (need for fluid bolus, vasopressor treatment, and fluid overload requiring diuretic therapy). Need to treat uterine atony or postpartum hemorrhage with uterotonics is recommended as a priority outcome measure for postpartum hemorrhage by a World Health Organization international panel.13 At our institution, the protocol for postpartum hemorrhage requires routine application of fundal massage after delivery; bimanual palpation is used if there is uterine atony or ongoing hemorrhage. The absolute decline in hematocrit was calculated by subtracting the first postpartum hematocrit (typically collected within 8–24 hours) from the most recent predelivery hematocrit (typically drawn at the time of admission for delivery). A 6% unit decline in hematocrit (eg, from 35% to 29%) was chosen because it corresponds on average to a 2-unit blood loss, which we consider to be clinically significant in the context of a vaginal birth. Need for blood transfusion was based on actual administration of whole or packed red blood cells before discharge. The need for fluid bolus and need for pressor treatment were as determined and ordered by the obstetric or anesthetic provider. Endometritis was based on a clinical diagnosis by the obstetric providers and the use of antibiotics for treatment. All outcomes were ascertained by chart abstraction until discharge from hospital by trained research nurses.
Two separate primary oxytocin dose comparisons were specified: 80 units compared with 10 units and 40 units compared with 10 units. As a secondary comparison, we planned to evaluate for trend in outcomes across 10-unit, 40-unit, and 80-unit dose groups. For both primary pair-wise comparisons, we estimated a sample size of 600 per group, or a total of 1,800, based on an assumed primary outcome rate of 18% in the 10-unit group (alpha of 0.05 for each comparison, 80% power, and a hypothesized 33% reduction in the primary outcome, ie, 18% to 12% for 80 units compared with 10 units and 18% to 12% for 40 units compared with 10 units). The baseline rate of 18% in the 10-unit group was estimated from a review of outcomes among vaginal deliveries conducted over 1 month at our institution. A single interim analysis was planned at enrollment of 1,200 women, approximately two thirds of the total planned sample size. The Lan-DeMets spending function approximation to O'Brien-Fleming stopping boundaries14 was used to adjust the level of significance of each primary analysis at both the interim review (significance level 0.017) and at study termination (significance level 0.033) to preserve the overall 0.05 level of significance. At the planned interim review (n=1,201 randomized) by the data safety and monitoring board, boundaries for early termination were not exceeded. An investigation for futility concluded that the conditional power for the 40 units compared with 10 units was less than 1%. The 40-unit arm thus was stopped for futility, and enrollment was continued in the 10-unit and 80-unit dose arms to the original total sample size of 1,800.
Analyses were by intent-to-treat. Baseline characteristics including risk factors for postpartum hemorrhage15,16 and outcomes were compared between groups. The χ2, analysis of variance, and Kruskal-Wallis tests were used as applicable for baseline and outcome comparisons among the three treatment groups. The χ2 tests and analysis of variance were used for two group tests. Fisher exact and Wilcoxon rank-sum tests were used when appropriate. We computed relative risks and 95% confidence intervals for pair-wise comparisons of outcomes. Tests for trends in dichotomous outcomes across groups were based on the Mantel Haenzsel test and tests for ordered differences in quantitative measures were based on the nonparametric Jonckheere-Terpstra test.17 All statistical tests, with the exception of the primary outcome as previously described, were evaluated at a 0.05 level of significance. SAS 9.2 was used for all statistical analyses.
From November 2008 through June 2010, 2,869 women were screened and 1,798 were randomized as follows: 658 to 80 units of prophylactic oxytocin, 481 to the 40-unit group (discontinued) and 659 to the 10-unit group (Fig. 1). The baseline characteristics of women were similar except the incidence of chorioamnionitis was higher, and the frequency of spontaneous membrane rupture and prolonged second stage of labor was lower in the 10-unit group (Table 1).
Overall, the primary composite outcome of treatment for hemorrhage or atony occurred in 6–7% of the study sample. Compared with the 10-unit group, higher doses of oxytocin did not significantly decrease the unadjusted risk of the primary outcome; there was no linear dose-response trend across groups (Table 2). Higher doses of oxytocin did not decrease treatment of uterine atony or obstetric hemorrhage with any uterotonics. However, 80 units but not 40 units of oxytocin compared with 10 units significantly decreased the need for treatment with additional oxytocin, corresponding to a decreased need for treatment after the first hour or in the postpartum suite. There was also a significant decreasing trend in the need for treatment with oxytocin (3% to 2% to 1%) with increasing dose of prophylactic oxytocin. All other components of the primary outcome, including the rare need for tamponade, surgery, interventional treatment, or blood transfusion, did not differ by dose of prophylactic oxytocin.
Mean change in hematocrit after delivery was not significantly different between groups. However, fewer women in the 80-unit group, but not in the 40-unit group, compared with the 10-unit group had a 6% or greater decline in hematocrit (Table 3). The incidence of this clinically important decline in hematocrit decreased modestly but significantly from 28% to 23% as prophylactic oxytocin dose increased from 10 units to 80 units (P<.05). Other secondary outcomes including estimated blood loss (mean and clinically estimated blood loss more than 500 mL), fluid bolus or pressor treatment for hypotension, fluid overload, endometritis, and prolonged hospitalization (4 or more days) did not differ by dose of prophylactic oxytocin. Because need for fluid bolus or pressor treatment in labor was primarily the result of epidural, we restricted the study population to women who did not receive an epidural; the incidences of fluid bolus by decreasing doses of oxytocin were 0%, 0%, and 0.3% (P for trend >.999). Results for pressor treatment were identical.
Relative risks (95% confidence interval) for the relationship between key study outcomes and higher doses of oxytocin compared with the 10-unit dose are given (Table 4). There were no significant differences between higher doses (80 or 40 units) and the 10-unit standard dose oxytocin for the primary composite outcome or need for any uterotonic to treat postpartum hemorrhage. However, there was a reduction in the need for oxytocin to treat uterine atony or postpartum hemorrhage, primarily on the postpartum floor in the postrecovery period. Results for a 6% or more decline in hematocrit suggested a lower incidence with 80 units but not with 40 units compared with 10 units of oxytocin (Table 4).
We conducted additional (post hoc) analyses to further evaluate our findings. The mean times (±SD) between predelivery and postdelivery hematocrit revealed no differences by group: 25.0±11.3, 25.2±12.1, and 25.3±11.5 hours, respectively, for 10-unit, 40-unit, and 80-unit groups (P=.891). The respective mean times (±SD) from delivery to postdelivery hematocrit were also similar: 15.5±8.0, 16.3±9.0, and 16.0±8.8 (P=.351). We compared the incidence of hematocrit decline greater than the prespecified 6% cut-off: 80 units compared with 10 units of prophylactic oxytocin was associated with a lower incidence of an 8% or greater decline in hematocrit (9% compared with 13%; P=.013) but not with a 10% or greater decline (4% compared with 5%; P=.301). Finally, results of analyses adjusting for the baseline differences were consistent with our main findings.
Overall, higher doses of prophylactic oxytocin (80 units or 40 units), as compared with the standard dose of 10 units of oxytocin when administered in 500 mL of crystalloid over the course of 1 hour after vaginal delivery, did not significantly reduce the incidence of the primary composite outcome of uterine atony or hemorrhage requiring any treatment. However, 80 units of oxytocin reduced the frequency of two prespecified secondary outcomes: hemorrhage requiring treatment after the first postpartum hour and a decline in hematocrit more than 6% units. There was a significant dose-response trend in these outcomes (reducing incidence with increasing dose of prophylactic oxytocin). Additionally, higher dose regimens were not associated with an increase in adverse events such as hypotension or fluid overload.
Findings from the few available studies examining various outcomes in relation to dose regimens suggest that both dose and rate of administration (including intravenous bolus) play a role.11,18 – 20 Typically, these studies have associated higher doses of prophylactic oxytocin with beneficial effect on outcomes such as estimated blood loss, decline in hematocrit, or need for additional uterotonics among women who underwent cesarean delivery.11,18 – 20 In two studies, women undergoing scheduled cesarean delivery received a 5-unit intravenous bolus compared with 35 units (5 unit bolus + 30 units over 4 hours) of prophylactic oxytocin.18,20 Women in the higher dose group had significantly lower mean estimated blood loss and lower frequencies of blood loss more than 500 ml or more than 1,000 mL, need for uterotonic treatment, or blood transfusion.18,20 In another small study of cesarean deliveries, higher dose of oxytocin was associated with a higher uterine tone and a nonsignificant reduction in need for additional uterotonic.19 Finally, in the previous trial at our institution, 80 units as compared with 10 units of prophylactic oxytocin reduced the need for any uterotonic treatment, as well as the need for treatment with second-line agents among women who underwent cesarean delivery after labor.11 It is noteworthy that women in both arms also received prophylactic low dose oxytocin (20 units in 1L) over 8 hours postpartum. Our current study is one of the largest randomized trials comparing different doses of oxytocin to prevent postpartum hemorrhage but focuses on women who underwent vaginal delivery. Although contrary to the previous study, we did not observe a significant reduction in need for any uterotonic treatment, the findings for prespecified secondary outcomes (hematocrit decline of 6 units or more and need for oxytocin after the first hour) do suggest potential benefits of higher dose regimens among women after a vaginal birth. Randomization provides balance across groups for baseline hematocrit and times to postdelivery hematocrit (a proxy for hydration during labor). Considering labor hydration, the amount of blood loss needed for a 6-unit decline in hematocrit may be higher than the postulated 2 units. Support for the safety of higher doses of oxytocin is evident from previous studies, including those of concentrated oxytocin protocols for mid-trimester pregnancy termination or induction of delivery11,21,22; these regimens have been generally reported to be safe.
The discrepancy in our primary outcome result could be attributed, at least in part, to differences in the risk profile of the study population and the higher rate of infusion of prophylactic dose regimens in the previous study (500 mL administered over the course of half an hour for cesarean deliveries compared with over the course of 1 hour for vaginal delivery in our study).11 The discrepancy may also be the consequence of our study's limited power to discern differences in the primary outcome. Our original estimated sample size was based on an 18% incidence of the primary outcome in the low-dose group. This included use of additional oxytocin to treat atony occurring while the prophylactic infusion was ongoing. However, based on pharmacy recommendations, we did not administer additional oxytocin during the first hour (concurrent with the prophylactic infusion). Instead, second-line uterotonics were administered if treatment was indicated in the first hour. The smaller than expected incidence of uterotonic treatment likely reflects a higher threshold to use methergine or prostaglandins as first-line treatment for hemorrhage or atony. The outcome rate of 7% would have required approximately double our current sample size to be able to detect a 33% reduction in the incidence of the primary composite outcome. At interim review, considering the 80% reduction in need for second-line agents to treat from a baseline of 9% in the previous trial and the futility of the intermediate study dose for the primary outcome, we opted to continue the enrollment in the two remaining arms to the original total sample size. We estimated that the sample size cumulated in these two groups would provide 80% power to detect a 50% reduction in the primary outcome from the new baseline of 7%. Apart from power considerations, our study had other limitations. Given the inherent difficulty in validly estimating postpartum blood loss, we used clinical outcomes as a proxy for blood loss. In our protocol, oxytocin was administered only after placental delivery. Although timing of administration does not appear to make a difference,23,24 we cannot guarantee that our results with higher doses would be the same if we initiated prophylactic oxytocin before placental delivery.
Overall, higher doses of prophylactic oxytocin appear to be beneficial in preventing measures of postpartum hemorrhage among women delivered by cesarean.11,18 – 20 Therefore, doses as high as 80 units administered over the course of 30 minutes are used for cesarean deliveries. However, in our study of women undergoing vaginal delivery, the incidence of uterine atony requiring treatment was not significantly reduced when 80 units were administered over the course of 1 hour, although we observed a reduction in the need for oxytocin in the postpartum suite and a reduction in a decrease in hematocrit more than 6 units with higher-dose oxytocin. Based on the negative findings concerning our primary outcome, practitioners may opt not to use a high dose of oxytocin after vaginal delivery. Alternatively, others may choose to use 80 units considering the potential for benefits based on the positive findings for secondary outcomes. This is particularly applicable in settings in which the high dose is already being used for cesarean deliveries. This would enhance efficiency by allowing for a single premixed oxytocin dose bag for prophylaxis (as opposed to one for cesarean deliveries and another for vaginal deliveries) considering the low cost of oxytocin. Our results indicate that 51 women receiving 80 units of prophylactic oxytocin are needed to prevent one episode of use of additional oxytocin to treat hemorrhage after the first postdelivery hour; 21 are needed to prevent one episode of hematocrit decline more than 6% units. Nevertheless, ongoing monitoring, evaluation, and reporting of this use in larger populations will be necessary to further demonstrate the safety and the effectiveness (vis-à-vis outcomes such as our primary outcome and blood transfusion) of higher-dose regimens for vaginal delivery.
An important consideration for the efficient use of 80-unit oxytocin dose for postpartum prophylaxis concerns its stability when concentrated in 500 mL of crystalloid. Although concentrations of 80 units in 1,000 mL or less (ie, 40 units in 500 mL or less) have been demonstrated to be stable for at least 7 days (and therefore a premixed bag can have a shelf-life of 7 days), no such data are available for the 80-unit in 500 mL concentration.25,26 As a result, hospital pharmacies will typically accord no more than a 2-day shelf-life for premixed oxytocin concentration more than 40 units in 500 mL. Therefore, stability studies of high-dose oxytocin regimens are needed to facilitate its efficient clinical use and evaluation. Because both the dose and duration of administration may play a role, future evaluation should also assess the effect of varying duration of administration, particularly when administered within 30 minutes.
1. Berg CJ, Chang J, Callaghan WM, Whitehead SJ. Pregnancy-related mortality in the United States, 1991–1997. Obstet Gynecol 2003;101:289–96.
2. Ronsmans C, Graham WJ, Lancet Maternal Survival Series steering group. Maternal mortality: who, when, where, and why. Lancet 2006;368:1189–200.
3. Postpartum hemorrhage. ACOG Practice Bulletin No. 76. American College of Obstetricians and Gynecologists. Obstet Gynecol 2006;108:1039–47.
4. Prendiville WJ, Elbourne D, McDonald S. Active versus expectant management in the third stage of labour. The Cochrane Database of Systematic Reviews 2000. Issue 3, Art. No.: CD000007. DOI: 10.1002/14651858.CD000007.
5. Elbourne DR, Prendiville WJ, Carroli G, Wood J, McDonald S. Prophylactic use of oxytocin in the third stage of labour. The Cochrane Database of Systematic Reviews 2001, Issue 4. Art. No.: CD001808. DOI: 10.1002/14651858.CD00180.
6. McDonald S, Abbott JM, Higgins SP. Prophylactic ergometrine-oxytocin versus oxytocin for the third stage of labour. The Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD000201. DOI: 10.1002/14651858.CD000201.pub2.
7. Liabsuetrakul T, Choobun T, Peeyananjarassri K, Islam QM. Prophylactic use of ergot alkaloids in the third stage of labour. The Cochrane Database of Systematic Reviews 2007, Issue 2. Art. No.: CD005456. DOI: 10.1002/14651858.CD005456.pub2.
8. McDonald SJ, Prendiville WJ, Blair E. Randomised controlled trial of oxytocin alone versus oxytocin and ergometrine in active management of third stage of labour [published erratum appears in BMJ 1993;307:1454]. BMJ 1993;307:1167–71.
9. Gulmezoglu AM, Villar J, Ngoc NT, Piaggio G, Carroli G, Adetoro L, et al.. WHO multicentre randomised trial of misoprostol in the management of the third stage of labour. Lancet 2001;358:689–95.
10. Cunningham FG, Leveno KJ, Bloom SL, Hauth J, Rouse D, Spong C. Williams Obstetrics. 23rd ed. New York (NY): McGraw Hill Medical; 2010. p. 399.
11. Munn MB, Owen J, Vincent R, Wakefield M, Chestnut DH, Hauth JC. Comparison of two oxytocin regimens to prevent uterine atony at cesarean delivery: a randomized controlled trial. Obstet Gynecol 2001;98:386–90.
12. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2009 Natl Vital Stat Rep 2010;59:1–29.
14. Friedman LM, Furberg CD, DeMets DL. Monitoring response variables. In: Fundamentals of clinical trials. 3rd ed. New York (NY): Springer; 1998. p. 246–83.
15. Combs CA, Laros RK Jr. Prolonged third stage of labor: morbidity and risk factors. Obstet Gynecol 1991;77:863–7.
16. Rouse DJ, Leindecker S, Landon M, Bloom SL, Varner MW, Moawad AH, et al.. The MFMU Cesarean Registry: uterine atony after primary cesarean delivery. Am J Obstet Gynecol 2005;193(3 Pt 2):1056–60.
17. Gibbons JD, Chakraborti S. Tests against ordered alternatives. In: Nonparametric statistical inference. 3rd ed. New York (NY): Marcel Dekker; 1992. p. 314–20.
18. Murphy DJ, MacGregor H, Munishankar B, McLeod G. A randomised controlled trial of oxytocin 5IU and placebo infusion versus oxytocin 5IU and 30IU infusion for the control of blood loss at elective caesarean section–pilot study. ISRCTN 40302163. Eur J Obstet Gynecol Reprod Biol 2009;142:30–3.
19. King KJ, Douglas MJ, Unger W, Wong A, King RA. Five unit bolus oxytocin at cesarean delivery in women at risk of atony: a randomized, double-blind, controlled trial. Anesth Analg 2010;111:1460–6.
20. Güngördük K, Asicioglu O, Celikkol O, Olgac Y, Ark C. Use of additional oxytocin to reduce blood loss at elective caesarean section: A randomized control trial. Aust N Z J Obstet Gynaecol 2010;50:36–9.
21. Winkler CL, Gray SE, Hauth JC, Owen J, Tucker JM. Mid-second-trimester labor induction: concentrated oxytocin compared with prostaglandin E2 vaginal suppositories. Obstet Gynecol 1991;77:297–300.
22. Owen J, Hauth JC, Winkler CL, Gray SE. Midtrimester pregnancy termination: a randomized trial of prostaglandin E2 versus concentrated oxytocin. Am J Obstet Gynecol 1992;167(4 Pt 1):1112–6.
23. Jackson KW Jr, 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.
24. Soltani H, Hutchon DR, Poulose TA. Timing of prophylactic uterotonic for the third stage of labour after vaginal birth. The Cochrane Database of Systematic Reviews 2010, Issue 8. Art. No.: CD006173. DOI: 10.1002/14651858.CD006173.pub2.
25. Gard JW, Alexander JM, Bawdon RE, Albrecht JT. Oxytocin preparation stability in several common obstetric intravenous solutions. Am J Obstet Gynecol 2002;186:496–8.
© 2012 by The American College of Obstetricians and Gynecologists.
26. Trissel LA, Zhang Y, Douglas K, Kastango E. Extended stability of oxytocin in common infusion solutions. Int J Pharm Comp 2006;10:156–8.