Postpartum hemorrhage after cesarean delivery: an analysis of risk factors.

Objective: To determine the incidence and risk factors for postpartum hemorrhage (PPH) associated with cesarean delivery.

Methods: Blood loss at cesarean delivery was measured and defined as 1,000 to 1,499 mL or greater than 1,500 mL and/or the need for a blood transfusion. Variables were identified and evaluated to determine the factors associated with PPH.

Results: There were 1,844 elective and 2,993 nonelective cesarean deliveries over 4 years. The PPH rate in nonelective cesarean (6.75%) was greater than after elective cesarean (4.84%, P = 0.007). Risk factors for PPH after an elective operation included leiomyomata, blood disorders, placenta previa, antepartum bleeding, preterm birth, and general anesthesia. Nonelective cesarean PPH risk factors included blood disorders, retained placenta, antepartum transfusion, antepartum/intrapartum hemorrhage, placenta previa, general anesthesia, and macrosomia (odds ratio >1.5, confidence interval >1.5).

Conclusions: Nonelective cesarean deliveries have a higher risk of PPH than women delivered electively. Risk factor identification and prevention should be a priority.

Key Words: cesarean delivery, postpartum hemorrhage, risk factors

**********

The average blood loss at cesarean delivery has been calculated by chromium-labeled red blood cells and averages 1,000 mL. (1) The normal physiologic changes of pregnancy add 1,200 to 1,600 mL of plasma volume (2) to the maternal intravascular volume and a loss of 1,000 mL is generally well tolerated in low-risk uncomplicated pregnancies, with only a slight rise in the maternal pulse or drop of the maternal pulse pressure. Further losses, however, may be poorly tolerated, particularly in women with marginal iron stores at the onset of pregnancy or inadequate iron replacement therapy during the pregnancy, as well as those with a contracted plasma volume.

Approximately 17% of the cardiac output, 500 to 800 mL of blood per minute at term, perfuses the maternal uterus. (3) A cesarean delivery, by necessity, results in acute blood loss before the uterine musculature can contract around the spiral arteries of the uterus and the hysterotomy incision can be closed. Risk factors have been identified that increase the loss of blood after abdominal delivery and include general anesthesia, Hispanic/North American ancestry, preeclampsia, amnionitis, prolonged first stage of labor, prolonged second stage of labor, macrosomia, a history of postpartum hemorrhage, and the method of placental removal at cesarean delivery. (4-7)

Many studies have evaluated medical therapy for postpartum hemorrhage (PPH) as well as surgical techniques. Fewer have analyzed the risk factors associated with PPH in women undergoing an abdominal delivery. The investigations that have assessed factors associated with PPH in cesarean deliveries (4-6) have combined risk factors for elective and nonelective cesarean cases rather than stratifying for elective and nonelective cesarean operations individually.

The purpose of this investigation was to determine the risk factors for postpartum hemorrhage associated with elective cesarean deliveries and emergency cesarean deliveries.

Materials and Methods

An extensive obstetric database was created in 1998 at King Edward Memorial Hospital in Perth, Australia, which recorded multiple fields of information on each pregnancy including maternal demographics, obstetric history, current medical conditions, antenatal diagnostic procedures and fetal conditions, antepartum complications, intrapartum events, mode of birth, and postpartum events. After each delivery, pertinent intrapartum and postpartum information was recorded in the required fields of the obstetric database by the midwife assigned to the patient. This information was correlated with antepartum, postpartum, and neonatal records by the database coordinator. The accuracy of those records is also verified by the database coordinator. This study was determined to be exempt from Institutional Review approval by the Ethics Committee at King Edward Memorial Hospital.

Blood loss at cesarean delivery was calculated by measuring the blood collected in the suction apparatus and by weighing lap pads, towels, and drapes. This blood loss was calculated before irrigation of the abdominal cavity. An additional blood loss after irrigation would then be added to the measured blood loss without the inclusion of the irrigation fluid. A postpartum hemorrhage was defined for blood loss greater than 1,000 mL and greater than 1,500 mL and/or maternal hemodynamic instability or anemia necessitating a blood transfusion. Preoperative and postoperative hematocrits were not recorded on all women; therefore, measured blood loss and hemodynamic instability necessitating a transfusion were used as criteria to define PPH in this paper. All cesarean deliveries were evaluated consecutively. An elective cesarean delivery was defined specifically for this study as a cesarean delivery performed as a scheduled procedure without labor, or as an unscheduled procedure undertaken without labor. All women undergoing labor before their cesarean delivery were defined as undergoing nonelective cesarean.

Multiple variables were evaluated for a possible association with postpartum hemorrhage. Medical conditions evaluated include chronic hypertension, insulin-dependent or adult-onset diabetes, alcohol or drug addiction, maternal obesity, circulatory system disorders (history of heart surgery, heart valve prosthesis, pulmonary embolus, vascular disease, and heart disease), blood disorders (sickle cell anemia, thalassemia, thrombocytopenia, coagulation defects, and thrombophilia), vaginal or cervical procedures (cervical cone biopsy, cervical suture, colposcopy with multiple biopsies), and leiomyomata. Past obstetric problems assessed included uterine/vaginal anomalies, contracted maternal pelvis, antenatal procedures (chorionic villus sampling, amniocentesis, fetal reduction, fetal transfusion) and fetal anomalies, history of a previous fetal loss, and prior PPH. Antepartum factors evaluated included antenatal care, single versus multiple gestation, oligohydramnios (amniotic fluid index No. 5), hydramnios (amniotic fluid index >20), placenta location, antepartum bleeding (placenta previa, abruptio placentae, and bleeding after 20 weeks of gestation of unknown etiology), threatened abortion, preterm labor, preterm rupture of the membranes, and preeclampsia/eclampsia. Intrapartum events assessed include antenatal transfer to a tertiary hospital, presentation at delivery, onset of labor (spontaneous or induced), type of analgesia, location of delivery (labor and delivery or the attached family birthing center), length of the first stage of labor (latent labor >20 hours in nulliparous patients and >14 hours in parous women, dilation of <1.2 cm per hour in primigravida and <1.5 cm per hour in multiparous women), the length of the second stage of labor (>2 hours in women without an epidural and >3 hours in women with an epidural), use of oxytocin for induction or augmentation, chorioamnionitis (persistent temperature >38[degrees]C along with uterine tenderness/irritability), gestational age at delivery, excessive intrapartum bleeding (clinically excessive bleeding during labor), intrauterine fetal demise, genital tract lacerations, and neonatal birth weight.

Prepregnancy medical conditions, antepartum, intrapartum, and immediate postpartum events were evaluated by univariate analysis for an association with PPH. All of the identified variables associated with PPH were then entered into a logistic regression model in a hierarchical fashion. At each step, the Wald [chi square] was inspected, and only variables that had a high probability of association were retained before the next set of variables were added to the model. Once all the variables had been added, categorical variables with a Wald [chi square] probability greater then 0.3 were collapsed into the baseline category. Odds ratios and 95% confidence intervals were then calculated for all variables and categories that were retained in the model. A P value of less than 0.05 was considered significant.

Results

Between July 1, 1998, and June 30, 2002, 18,705 women gave birth to 19,476 babies; 4,837 women delivered by cesarean and 13,868 delivered vaginally. Of these women who delivered abdominally, 1,844 were delivered by an elective cesarean and 2,993 delivered by a nonelective cesarean. The PPH rate for the women undergoing an elective cesarean section with a measured blood loss of greater than 1,000 mL and blood transfusion cut off was 90/1,844 (4.84%), and the PPH rate was 35/1,844 (1.90%) for the women with a measured blood loss of greater than 1,500 mL and blood transfusion cutoff. There was no difference in the overall rate of PPH between the women undergoing an elective cesarean delivery and of those women giving birth vaginally (5.15%, P = 0.083 or 2.4%, P = 0.178, respectively). The women undergoing a nonelective abdominal delivery had a PPH rate of 202/2,993 (6.75%) in the 1,000 mL and blood transfusion category and 91/2,993 (3.04%) for the 1,500 mL/blood replacement cut off. The PPH rate was greater in women undergoing a nonelective cesarean delivery than for women giving birth vaginally (P = 0.001 or P = 0.043, respectively). The rate of PPH in the nonelective cesarean section group was also greater than the elective cesarean group (P = 0.007, and P = 0.015, respectively).

The maternal demographics of age, race, parity, and gestational age at delivery of the elective and nonelective cesarean deliveries are presented (Table 1). Ninety-eight maternal demographic, antepartum, and intrapartum variables were assessed to identify risks for a PPH. The identified risk factors for a PPH in the women undergoing an elective cesarean (Table 2) are presented for the >1,000 mL and 1,500 mL categories and included leiomyomata (P = 0.039 and 0.001), blood disorders in pregnancy (P = 0.030 and 0.003), threatened abortion (P = 0.024 and 0.078), placenta previa (P = 0.001 and 0.001), antepartum bleeding other than threatened abortion (P = 0.001 and 0.001), birth before 37 weeks (P = 0.001 and 0.015), and the use of general anesthesia (P = 0.001 and 0.006) (Table 2).

Risk factors for PPH after a birth by nonelective cesarean section included genital lacerations (P = 0.005 and 0.066), obesity (P = 0.010 and 0.473), circulatory system disorders (P = 0.019 and 0.436), blood disorders in pregnancy (P = 0.012 and 0.001), retained placenta in previous pregnancies (P = 0.370 and 0.047), vaginal or cervical operation (P = 0.004 and 0.690), antepartum maternal transfusion (P = 0.038 and 0.005), antepartum hemorrhage (P = 0.001 and 0.001), placenta previa without hemorrhage (P = 0.001 and 0.001), dystocia in labor (P = 0.004 and 0.104), intrapartum hemorrhage (P = 0.001 and 0.015), general anesthesia (P = 0.001 and 0.001), and neonatal birth weight greater than 4 kg (P = 0.001 and 0.017) (Table 3). The wide confidence intervals observed for placenta previa, antenatal transfusion, genital tract trauma, and circulatory disorders reflect the low number of events, and, as such, may denote a trend rather than an absolute event, although the odds ratio is significant.

Discussion

The combined PPH rate of the elective and nonelective cesarean deliveries in our study was 6% and is similar to the rate of 6.4% that was noted by Combs et al. (4) in their study evaluating the risk of a PPH rate in women who delivered by cesarean. Their PPH rate of 6.4% was significantly higher than the rate of 3.9% that was observed in their patients who underwent vaginal deliveries. (8) We observed no difference in the percentage of women with a PPH after a vaginal delivery or an elective cesarean delivery; however, the women undergoing a nonelective cesarean delivery had a significantly greater risk of PPH than the women who delivered vaginally or by elective cesarean for both the 1,000 and 1,500 mL cutoffs. A similar finding may have been present in the study by Combs et al., (4) but it cannot be determined because the deliveries were not allocated into elective and nonelective cesarean deliveries.

The maternal demographics of the women undergoing repeat elective cesarean delivery are different from the women undergoing a nonelective operation. The women undertaking a repeat cesarean were older and of greater parity than the women who underwent a nonelective cesarean. These differences are understandable and expected, as nearly all of the elective cesarean operations are repeat procedures, except for some limited abdominal deliveries done for fetal position and fetal and/or maternal conditions. The similarity in the gestational age at the time of delivery between both types of cesareans is surprising and may reflect that the overwhelming number of elective deliveries are at term and the number of women induced for post dates in the nonelective group, who then fail the induction, may balance the number of preterm deliveries in that group.

There are a number of risk factors for a PPH that are common to both elective and nonelective cesarean deliveries. Leiomyomata, by distorting the normal architecture of the uterus and preventing contraction of the uterine musculature, are easily understood as a reason for PPH, although there are very few published studies that have confirmed this association. (9) Those with blood disorders of pregnancy likewise were at greater risk of PPH. The use of general anesthesia, because of the uterine relaxing properties of some anesthetic agents, has been identified as a risk factor for PPH. (4,6) Placenta previa and its interference with contraction of the uterine musculature in the lower segment is a well known reason for PPH. (6) Antepartum bleeding beyond the 20th week of pregnancy, not related to a placenta previa or an abruption, was another risk factor for PPH in women who delivered abdominally and has not been previously reported. The reason for this link is unknown.

There were two additional predictive factors for a PPH exclusively after an elective cesarean delivery: threatened abortion and preterm delivery (<37 weeks). The more common use of a vertical or classic uterine incision in preterm deliveries and the delayed detachment of the placenta in preterm deliveries are the most likely reasons for this increased blood loss in preterm deliveries. (10) A threatened abortion in early pregnancy correlated with a PPH at the 1,000 mL limit but not at 1,500 mL.

Risk factors for a PPH were documented in women undergoing a nonelective cesarean delivery but not in elective cesarean deliveries. Obesity, prolonged first and second stages of labor, dystocia in labor, as well as infant weight greater than 4 kg have previously been related to PPH. (4,6) Women who were transfused antenatally and those women with a placenta previa had an increased risk of PPH. Women with circulatory disorders (history of heart surgery, heart valve prosthesis, pulmonary disease, or cardiac disease) also had a greater risk of PPH; the reason for which is unclear. In addition, women who had undergone a cervical procedure (conization, biopsy, LEEP) were at greater risk for PPH at the 1,000 mL cutoff but not at the 1,500 mL cutoff, whereas those with a prior retained placenta did not have an increased risk for PPH at 1,000 mL but did at 1,500 mL.

Patients who had excessive intrapartum bleeding (excessive bleeding during labor without evidence of an abruption, previa or vasa previa) or those with genital tract lacerations (uterine rupture, bladder laceration, vaginal extension of uterine incision) were also at risk for excessive postpartum bleeding. Patients delivered by a nonelective cesarean section have a greater risk of having a PPH than women delivered by an elective cesarean or a vaginal delivery. This study had a power of 76.3% to identify the differences between elective and nonelective cesarean delivery PPH rates without the covariates and 99% with covariates, accounting for 25% of the variance. The recognition of this group of women who overall have a greater risk of hemorrhage, and the identification of the individual risk factors for hemorrhage in women undergoing an abdominal delivery, will assist in their management.

The importance of this study is emphasized by the large number of women evaluated, the meticulous measurement of blood loss, the separation of cesarean deliveries into elective and nonelective, and that few investigations other than the study by Combs (7) in 1991 have addressed the risk factors for bleeding at cesarean delivery. Prophylactic practices and surgical techniques, along with experienced providers, will help make certain that an optimal management team will be available to decrease the incidence of a PPH and thereby minimize maternal hemorrhage, which remains one of the leading causes of maternal death. (11)

References

1. Pritchard JA, Baldwin RM, Dickey JC, et al. Blood volume changes in pregnancy and the puerperium. II Red blood cell loss and changes in apparent blood volume during and following vaginal delivery, cesarean section, and cesarean plus total hysterectomy. Am J Obstet Gynecol 1962;84:1271-1282.

2. Lund CJ. Donavan JC. Blood volume during pregnancy. Am J Obstet Gynecol 1967;98:393.

3. Gant NF, Worley RJ. Measurement of Uteroplacental Blood Flow in the Human, in Rosenfeld (ed): The Uterine Circulation. Ithaca, Perinatalogy Press, 1989, pp 53-73.

4. Combs CA, Murphy EL, Laros RK. Factors associated with hemorrhage in cesarean deliveries. Obstet Gynecol 1991;77:77-82.

5. Naef RW, Chauhan SP, Chevalier SP, et al. Prediction of hemorrhage at cesarean delivery. Obstet Gynecol 1994;83:923-926.

6. Eggebo TM, Gjessing LK. Hemorrhage after cesarean. Tidsskrift for Den Norske Laegeforening 2000;120:2864-2866.

7. Morales M, Ceysens G, Jastrow N, et al. Spontaneous delivery or manual removal of the placenta during cesarean section: a randomized controlled trial. BJOG 2004;111:908-912.

8. Combs CA, Murphy EL, Laros RK. Factors associated with hemorrhage in vaginal deliveries. Obstet Gynecol 1991;77:77-82.

9. Grignon DJ, Carey MR, Kirk ME, et al. Diffuse leiomyomatosis: a case study with pregnancy complicated by intrapartum hemorrhage. Obstet Gynecol 1987;69:477-480.

10. Dombrowski MP, Bottoms SF, Saleh AA, et al. Third stage of labor: Analysis of duration and clinical practice. Am J Obstet Gynecol 1995;172:1279-1284.

11. Revised 1990 estimates of maternal mortality: a new approach by WHO and UNICEF. Geneva, World Health Organization, 2000.

Give what you have. To someone, it may be better than you dare to think.
--Henry Wadsworth Longfellow

Everett F. Magann, MD, Sharon Evans, PHD, Maureen Hutchinson, RN, RM, Robyn Collins, RN, RM, BAPPSC, CHC, Grainger Lanneau, MD, and John C. Morrison, MD

From the School of Women and Infants Health, King Edward Memorial Hospital, Perth, Australia; the Departments of Obstetrics and Gynecology, Bremerton Naval Hospital, Bremerton, WA; and the University of Mississippi Medical Center, Jackson, MS.

Reprints will not be available. Correspondence to Dr. John C. Morrison, Department of Obstetrics and Gynecology; University of Mississippi Medical Center; 2500 N. State St., Jackson, MS 39216-4505. Email: [e-mail]

Accepted January 13, 2005.

RELATED ARTICLE: Key Points

* The postpartum hemorrhage (PPH) rate in nonelective abdominal birth (6.75%) is greater than after elective cesarean (4.84%, P = 0.007).

* Identified risk factors for PPH after an elective operation included leiomyomata, placenta previa, antepartum bleeding, preterm birth, and general anesthesia.

* Nonelective cesarean PPH risk factors include blood disorders, retained placenta, antepartum transfusion, antepartum/intrapartum hemorrhage, placenta previa, general anesthesia, and macrosomia.

Table 1. Demographic factors

                         Elective              Emergency          P
                         cesarean  %           cesarean   %       value

Maternal age*               31.5    5.5        29.7       6.14    0.001
Race
  White                  1,367     74.1        2,167      72.4    0.001
  Asian                    188     10.2          284       9.5
  Aboriginal               132      7.2          305      10.2
  Other                    157      8.5          237       7.9
Gravidity
  1                        266     14.4        1,020      34.0    0.001
  2                        515     27.9          771      25.8
  3+                     1,063     57.7        1,202      40.2
Gestational age[dagger]     38     37.0, 38.2     38      34, 38  0.641

*Mean, standard deviation.
[dagger]Median, interquartile range.

Table 2. Odds ratio and 95% confidence limits from logistic regression
analysis for association of factors with postpartum hemorrhage (blood
loss >1,000 mL or 1,500 mL and/or transfusion) for elective cesarean
deliveries

                              OR for blood loss
                                 >1,000 mL        95% Confidence
Effect                        and/or transfusion     limits

Placenta previa               6.65                (3,24, 13.09)
General anesthesia            4.81                (2.34, 9.35)
Antepartum hemorrhage         4.33                (2.36, 7.69)
Fibroids                      3.52                (0.92, 10.64)
Threatened abort              2.48                (1.06, 5.24)
Premature delivery <37 weeks  2.23                (1.27, 2.82)
Blood disorders               1.77                (1.04, 2.92)

                              OR for blood loss
                                  >1,500 mL       95% Confidence
Effect                        and/or transfusion     limits

Placenta previa                8.06               (2.98, 21.81)
General anesthesia             4.05               (1.49, 10.99)
Antepartum hemorrhage          4.91               (2.10, 11.51)
Fibroids                      11                  (3.01, 40.00)
Threatened abort               2.81               (0.89, 8.84)
Premature delivery <37 weeks   2.79               (1.22, 6.37)
Blood disorders                3.22               (1.50, 6.92)

Table 3. Odds ratios and 95% confidence limits from logistic regression
analysis for association of factors with postpartum hemorrhage (blood
loss >1,000 mL or 1,500 mL and/or transfusion) for nonelective cesarean
deliveries

                              OR for blood loss
                                 >1,000 mL        95% Confidence
Effect                        and/or transfusion     limits

Placenta previa                      6.00         (2.08, 16.72)
Antenatal transfusion                4.80         (1.02, 20.67)
Genital tract trauma                 4.80         (1.41, 13.74)
Circulatory system disorders         3.92         (1.12, 11.48)
Intrapartum hemorrhage               3.06         (1.89, 4.94)
Antepartum hemorrhage                2.92         (1.90, 4.45)
General anesthesia                   2.79         (1.88, 4.10)
Vaginal/cervical operation           2.66         (1.29, 5.08)
Birth wt >4 kg                       2.45         (1.60, 3.71)
Obesity                              2.28         (1.17, 4.16)
Dystocia in labor                    1.79         (1.19, 2.67)

                              OR for blood loss
                                 >1,500 mL        95% Confidence
Effect                        and/or transfusion     limits

Placenta previa                      11.21        (3.51, 35.79)
Antenatal transfusion                 9.01        (1.94, 41.94)
Genital tract trauma                  4.53        (0.90, 22.70)
Circulatory system disorders          2.11        (0.32, 13.71)
Intrapartum hemorrhage                2.27        (1.18, 4.39)
Antepartum hemorrhage                 3.64        (2.00, 6.60)
General anesthesia                    4.50        (2.70, 7.52)
Vaginal/cervical operation            0.77        (1.21, 2.85)
Birth wt >4 kg                        2.26        (1.16, 4.41)
Obesity                               1.50        (0.50, 4.49)
Dystocia in labor                     1.61        (0.91, 2.86)