- Top of page
Secondary postpartum hemorrhage is defined as any abnormal bleeding from the uterus occurring between 24 h and 12 weeks postpartum1 and occurs in 1–2% of deliveries2. The most common causes of secondary postpartum hemorrhage are abnormal involution of the placental site in the uterine cavity, endometritis or retention of placental tissue3. Arteriovenous malformations of the uterine artery have been described recently as rare causes of severe secondary postpartum bleeding4. Secondary postpartum hemorrhage is an important cause of maternal morbidity worldwide. In developing countries, it is a major contributor to maternal death1. In developed countries, more than half of postpartum women admitted to hospital with this condition undergo uterine surgical evacuation1, 5, 6. Histological confirmation of residual placental tissue is obtained in only 30% of these cases1, 5, 6. Moreover, puerperal curettage might traumatise the implantation site and provoke additional bleeding, which can be life-threatening and can require a hysterectomy. In a recently published review concerning secondary postpartum hemorrhage6, uterine curettage was performed in 63% of cases. Perforation of the uterus after curettage occurred in 3% and hysterectomy in about 1% of patients. Histological confirmation of residual placental tissue was obtained in 37% of cases following sonographic diagnosis of retained placental tissue. Thus, even though ultrasound technology has become more powerful, demonstration of retained placental tissue is still difficult.
Westendorp et al.7 examined prospectively 50 women undergoing either a repeat removal of placental remnants after delivery or a repeat curettage for incomplete miscarriage. At a later hysteroscopy, 20 of the 50 women had intrauterine adhesions. Alexander et al.1 identified 45 papers on the management of secondary postpartum hemorrhage and concluded that there is little information available from randomized trials to guide clinicians in the management of this condition.
Curettage in the postpartum period may be hazardous. It is indicated only if placental tissue remains in the uterine cavity. Therefore, a tool that demonstrates retained placental tissue would be advantageous. There are conflicting data on the sonographic diagnosis of retained placental tissue and ultrasound has been reported to have limited diagnostic accuracy concerning secondary postpartum hemorrhage8–18. As echogenic masses have been found in asymptomatic women19, 20, the previously held opinion that an echogenic mass in the cavity represents retained placental tissue responsible for postpartum bleeding8–18 has been challenged.
In our longitudinal study of the normal puerperium21, echogenic masses were not found in any of the 42 women. However, a mixed-echo pattern (echogenic material mixed with fluid components) occurred frequently, particularly during the first 2 weeks of the puerperium. In the majority of studies, distinction between an echogenic mass and a mixed-echo pattern is not well established and there is confusion as to how ultrasound images of the intracavitary content should be interpreted8–18.
The purpose of this study was to describe sonographic findings associated with secondary postpartum hemorrhage of various etiologies, particularly retained placental tissue. Another aim was to compare these findings with those from a study of women with a normal puerperium.
- Top of page
Ninety women with secondary postpartum hemorrhage were invited to participate in this study conducted at Uppsala University Hospital, Sweden, between 1996 and 2004. The women had delivered between the 24th and 42nd week of pregnancy. Eight women declined to participate and three women were excluded because of large fibromyomas. Patients with predominant clinical symptoms of puerperal endometritis were not included. Informed consent was obtained from the seventy-nine women remaining in the study and the ethics committee of the medical faculty at Uppsala University approved the study. Demographic and obstetric data were obtained from the medical records. The sonographers were informed about the clinical data but were not responsible for the management of the patients. The physician responsible for the patient received descriptive information of the ultrasound findings.
All ultrasound scans were performed with commercially available real-time machines: Acuson Sequoia 512 (Acuson, Mountain View, CA, USA) or Siemens sonoline SI-400 (Siemens, Erlangen, Germany) equipped with a 3.5- or 4-MHz transabdominal convex probe or a 5-MHz vaginal probe. In order to compare results with the reference values from a normal population21, a similar study design was used and the examinations were scheduled for postpartum days 1, 3, 7, 14, 28 and 56. The first day of examination was the day when clinical symptoms started. The examinations were continued until uterine surgical evacuation was performed or until the bleeding stopped. If a woman presented with secondary postpartum hemorrhage, for example on day 9, the findings were compared with reference findings from day 7 and the scans were repeated on days 14, 28 and 56 postpartum. The uterus was assessed in the longitudinal section. During the first 2 weeks postpartum, a transabdominal probe was used and for the later examinations a transvaginal probe was used. A few exceptions were made in cases when the uterus was too large and out of the focal range of the vaginal probe, in which case the transabdominal approach was chosen even later than 2 weeks postpartum. The urinary bladder was filled moderately during the transabdominal examinations and was empty during the transvaginal examinations.
The maximum anteroposterior (AP) diameters of the uterus and uterine cavity in the longitudinal section were measured. Morphological findings in the uterine cavity were recorded. If only an echogenic central line was visualized from the fundus to the lower uterine segment, the cavity was defined as empty. Fluid in the cavity was defined as a space separating the anterior from the posterior wall. A mixed-echo pattern was defined as echogenic material mixed with fluid components of varying proportions. An echogenic mass was defined as a well-circumscribed mass, often with a lobulated appearance and calcifications, without any fluid components.
Statistical analyses were performed with the SAS statistical package (SAS Institute Inc., Cary, NC, USA, 1997). Descriptive statistics were used for the entire study population, expressed as median and range. Frequencies were presented as percentages. Quantitative values from pathological pregnancies were plotted on the reference curves21, with the 10th, 50th and 90th percentiles denoted.
We used t-tests to compare mean values of demographic and obstetric variables. For categorical data, chi-square tests, or in cases of sparse data, Fischer's exact test, were used to compare group proportions. A P-value < 0.05 was considered to be statistically significant.
- Top of page
Seventy-nine patients were enrolled into the study and were assigned to one of two groups according to choice of treatment as follows. Group 1 (n = 18, 22.8%) consisted of patients who underwent surgical evacuation. All sonographic measurements were performed before surgical evacuation. Tissue was sent for histological analysis in 16 of these 18 cases but was inconclusive for two patients. Three patients required two evacuations. Group 2 (n = 61, 77.2%) was treated conservatively, consisting of expectant management alone or in combination with uterotonic drugs and/or antibiotics.
Twenty-five of the 79 (31.6%) patients had both primary and secondary postpartum hemorrhage, nine from Group 1 and 16 from Group 2. The median time intervals between primary and secondary postpartum hemorrhages were 10 (range, 1–60) days and 9 (range, 1–95; only one case > 84) days for Groups 1 and 2, respectively. Ten of the 79 (12.6%) patients had undergone primary manual evacuation of the placenta, four from Group 1 and six from Group 2. All except one patient (Case 15), who underwent manual evacuation of the placenta, had primary postpartum hemorrhage. The initial treatment for this, apart from manual evacuation of the placenta, was conservative, with uterotonic drugs.
Table 1 summarizes the demographic and obstetric data of the patients. More patients from Group 1 received blood transfusions and antibiotics. The blood transfusions were treatment for secondary postpartum hemorrhage in eight patients from Group 1 and in three patients from Group 2. Blood transfusions were performed in the other eight cases in Group 2 for treatment of primary postpartum hemorrhage. There were no other differences.
Table 1. Demographic and obstetric characteristics of the study population
|Characteristic||Group 1||Group 2||P|
|n (% of total)||18 (22.8)||61 (77.2)|| |
|Age (years, median (range))||30 (22–40)||30 (20–43)||0.683|
| Primiparous||7/18 (38.9)||27/61 (44.3)|| |
| Multiparous||11/18 (61.1)||34/61 (55.7)|| |
|GA at delivery (weeks, median (range))||39 (24–41)||39 (24–42)||0.169|
|Smoker (n (%))||5/18 (27.8)||10/61 (16.4)||0.312|
|History of prior postpartum bleeding (n (%))||7/11 (63.6)||17/34 (50.0)||0.491|
|Mode of delivery (n (%))|| ||0.437|
| Spontaneous vaginal||17/18 (94.4)||51/61 (83.6)|| |
| Ventouse||1/18 (5.6)||6/61 (9.8)|| |
| Cesarean section||0/18 (0)||4/61 (6.6)|| |
|Birth weight (g, median (range))||3540 (798–6200)||3470 (553–5660)||0.995|
|Breast-feeding (n (%))||14/18 (77.8)||50/61 (81.9)||0.736|
|Estimated bleeding at delivery (mL, median (range))||1200 (200–3500)||500 (100–2800)||0.118|
|Blood transfusion (n (%))||8/18 (44.4)||11/60* (18.3)||0.019|
|Uterotonic drugs (n (%))||18/18 (100)||54/61 (88.5)||0.341|
|Antibiotics (n (%))||8/16† (50.0)||13/61 (21.3)||0.030|
|Surgical treatment (n)||18/18||0/61|| |
|Histology available (n (%))||16/18 (88.9)||0/61|| |
Table 2 shows ultrasound findings in Group 1 at presentation. The AP diameters of the uterus and uterine cavity are shown in Figure 1, with the reference values as percentile curves. Uterine size tended to be larger compared with the reference group. Sixteen of the patients had cavity measurements above the 90th percentile (Figure 1b) and 17 had measurements above the 50th percentile (Table 2).
Figure 1. Maximum uterine (a) and uterine cavity (b) anteroposterior diameters (dashed lines) in postpartum women with secondary postpartum hemorrhage who underwent surgical evacuation (Group 1). Reference values21 (solid lines) from uncomplicated postpartum periods (10th, 50th and 90th percentiles) are given for comparison.
Download figure to PowerPoint
Table 2. Ultrasound findings in Group 1 on the postpartum day on which patients presented with symptoms; all patients underwent a uterine surgical evacuation
|Case||Day*||Day†||AP: uterus (mm)||Normal references||AP: uterine cavity (mm)||Normal references (mm, median (range))||Mixed echo (n)||Echogenic mass (n)||Histological confirmation (n)|
|1||1||1||99.1||92.7 (70.0–109.3)||25.3||14.2 (3.5–45.2)||0||1||+|
|2||15||14||73.8||62.7 (50.0–86.0)||32.2||7.9 (2.5–22.0)||0||1||+|
|3||7||7||65.1||74.0 (58.3–94.0)||29.3||13.5 (3.2–44.0)||0||1||+|
|4||10||7||105.5||74.0 (58.3–94.0)||79.1||13.5 (3.2–44.0)||0||1||+|
|5||2||3||74.3||84.9 (68.5–109.0)||28.1||10.6 (2.5–36.0)||0||1||+|
|6||10||7||74.6||74.0 (58.3–94.0)||27.5||13.5 (3.2–44.0)||0||1||+|
|7||55||56||64.5||38.1 (29.4–50.0)||28.6||2.9 (6.1–10.0)||0||1||?|
|8||13||14||76.7||62.7 (50.0–86.0)||29.1||7.9 (2.5–22.0)||0||1||+|
|9||3||3||95.8||84.9 (68.5–109.0)||37.7||10.6 (2.5–36.0)||0||1||?|
|10||53||56||54.4||38.1 (29.4–50.0)||35.9||2.9 (6.1–10.0)||0||1||+|
|11||60||56||63.8||38.1 (29.4–50.0)||32.2||2.9 (6.1–10.0)||0||1||+|
|12||6||7||74.9||74.0 (58.3–94.0)||30.1||13.5 (3.2–44.0)||0||1||+|
|13||7||7||70.1||74.0 (58.3–94.0)||12.6||13.5 (3.2–44.0)||1||0||0|
|14||13||14||81.6||62.7 (50.0–86.0)||47.7||7.9 (2.5–22.0)||0||1||+|
|15||10||7||68.9||74.0 (58.3–94.0)||35.9||13.5 (3.2–44.0)||0||1||0|
|16||8||7||70.0||74.0 (58.3–94.0)||23.1||13.5 (3.2–44.0)||0||1||+|
|17||30||28||73.4||49.7 (36.8–61.0)||53.5||5.0 (1.4–21.0)||0||1||+|
|18||1||1||69.3||92.7 (70.0–109.3)||35.1||14.2 (3.5–45.2)||0||1||+|
The most common qualitative ultrasound finding from patients in Group 1 was a well-circumscribed, lobulated echogenic mass, which was observed in 17 (94.4%) patients (Figures 2a, 3a and 4a). The reference ultrasound scans are presented as Figures 2b, 3b and 4b. One patient had a mixed-echo pattern and a cavity width within the normal range (Case 13, Table 2). Tissue from this patient was not sent for histological analysis. In 14 cases with a demonstrated echogenic mass, histology revealed placental tissue. Thirteen of these patients had an AP diameter of the cavity above the 90th percentile.
Figure 2. (a) Transabdominal ultrasound image of the uterus on day 13 postpartum in a patient with secondary postpartum hemorrhage (Group 1). An echogenic mass is seen in the cavity. Calipers indicate anteroposterior (AP) diameter of the uterine cavity. (b) A reference ultrasound image of the uterus of a woman with uncomplicated puerperium on day 14 postpartum is given for comparison. A mixed-echo pattern is visible as echogenic material combined with fluid. Calipers 1 and 2 indicate AP diameters of the uterus and uterine cavity, respectively.
Download figure to PowerPoint
Figure 3. (a) Transabdominal ultrasound image of the uterus on day 30 postpartum in a patient with secondary postpartum hemorrhage (Group 1). An echogenic mass is seen as a large well-defined mass with lobulated appearance and without fluid component. Calipers 1 and 2 indicate anteroposterior (AP) diameters of the uterus and uterine cavity, respectively. (b) A reference transvaginal ultrasound image of the uterus of a woman with uncomplicated puerperium on day 28 postpartum is given for comparison. An echogenic central line is visible from the fundus to the lower uterine segment. Calipers 1 and 2 indicate AP diameters of the uterus and uterine cavity, respectively.
Download figure to PowerPoint
Figure 4. (a) Transvaginal ultrasound image of the uterus on day 53 postpartum (Group 1). An echogenic mass is seen in the cavity. Calipers 1 and 2 indicate anteroposterior (AP) diameters of the uterus and uterine cavity, respectively. (b) A reference ultrasound image of the uterus of a woman with uncomplicated puerperium on day 56 postpartum is given for comparison. A central line indicates the empty cavity. Calipers 1 and 2 indicate AP diameters of the uterine cavity and uterus, respectively.
Download figure to PowerPoint
The conservatively treated Group 2 had somewhat higher measurements for uterine and cavity size compared with the reference values, but there was extensive overlap (Figure 5). Eighteen of the 61 patients from Group 2 had an empty cavity or a minor amount of fluid (Figure 6a). The remaining 43 had a cavity with mixed-echo patterns (Figure 6b). None had an echogenic mass in the cavity. All patients from Group 2 were treated conservatively and bleeding stopped spontaneously.
Figure 5. Maximum uterine (a) and uterine cavity (b) anteroposterior diameters in women with secondary postpartum hemorrhage, who were treated conservatively (Group 2). Reference values21 (red lines) from uncomplicated postpartum periods (10th, 50th and 90th percentiles) are given for comparison.
Download figure to PowerPoint
Figure 6. (a) Transabdominal ultrasound image of the uterus on day 14 postpartum in a patient who was treated conservatively (Group 2). A minor amount of fluid separating the anterior from the posterior wall is seen in the cavity. Calipers 1 and 2 indicate anteroposterior (AP) diameters of the uterus and uterine cavity, respectively. (b) Transabdominal ultrasound image of the uterus on day 16 postpartum in a patient from Group 2. A mixed-echo pattern is seen in the cavity. Calipers 1 and 2 indicate AP diameters of the uterus and uterine cavity, respectively.
Download figure to PowerPoint
- Top of page
Retained placental tissue in the uterine cavity postpartum is associated with a high risk of excessive bleeding, either immediate or delayed. In patients with secondary postpartum hemorrhage, sonography can help to verify or rule out retained placental tissue. Knowledge of the normal involution process, including sonographic findings of the changes of the uterus and uterine cavity throughout the puerperium is a pre-requisite for optimal interpretation of these findings in affected cases.
In this study, the most common ultrasound finding associated with retained placental tissue was an echogenic mass. All 14 cases from Group 1 with histological confirmation of placental tissue had an echogenic mass in the cavity. The AP diameter was above the 90th percentile in all except one of these cases (Case 16, Table 2). Moreover, in our previous study no echogenic mass was identified in 42 women with an uncomplicated puerperium21.
Ultrasound images of retained placental tissue have been described as ‘a discrete solid mass and calcifications with acoustic shadows inside the mass’ (Lee et al.11), ‘an echogenic mass’ (Hertzberg and Bowie10), and ‘highly variable but most often an echogenic mass’ (Carlan et al.17). These descriptions are in good agreement with the results of our study. However, the Hertzberg study was retrospective, Lee et al. included cases of early postpartum hemorrhage and Carlan et al. evaluated asymptomatic women immediately after delivery of the placenta. In contrast, Edwards and Ellwood19 observed an echogenic mass in 51% of normal cases on day 7, in 21% on day 14 and in 6% on day 21, and questioned whether sonographic findings of an echogenic mass in the uterine cavity actually represent retained placental tissue. The appearance of the uterine cavity contents was classified into two categories: presence of an echogenic mass or not. However, it is unclear whether a mixed-echo pattern was included in the ‘echogenic mass’ group. Sokol et al.20 used the same classification and found ‘echogenic material’ in 40% of women 48 h after a normal delivery. However, 14 of the 16 cases demonstrated echogenic material in the lower uterine segment, while only two had such findings in the fundus. It is unclear if ‘echogenic material’ is the same as an ‘echogenic mass’, or if it might be a mixed-echo pattern as defined here.
In our study, only one of 18 (5.6%) patients who underwent a surgical evacuation had a mixed-echo pattern. Unfortunately, histological examination was not performed in this case. The patient also had an AP uterine cavity diameter within the normal range (Case 13, Table 2). Of 61 patients treated conservatively, 43 had a mixed-echo pattern but none required a surgical procedure.
In our previous study on the normal puerperium21, 33 of 42 (78.6%) women had some content in the cervical area on day 1 postpartum; in the majority of cases, there was fluid or a mixed-echo pattern. Twenty of 42 (47.7%) women displayed the same findings in the cervical area on day 3. Moreover, mixed-echo content was observed in the entire uterine cavity in 37 of 41 (90.2%) women on day 7, and in 31 of 41 (75.6%) women on day 14. It was concluded that a mixed-echo pattern is a common and insignificant finding of the involuting uterus. The findings were located in the lower uterine segment or cervical area in early puerperium, in the entire uterine cavity in the middle part of puerperium and became uncommon during the late postpartum period (being present in 10 of 39 women on day 28 and in none of 41 women on day 56)21.
Carlan et al.17 observed ‘false positive images’ in eight of 103 (7.7%) women and the common finding was a mixed pattern confined to the lower part of the uterus. Lee et al.11 reported that in all five cases with a mixed-echo pattern there was negative histology. These reports support the view that a mixed-echo pattern is compatible with a normal puerperium. However, differentiation between an echogenic mass and mixed-echo pattern might be difficult and subjective. It should be noted that none of our patients with histological confirmation of retained placental tissue had a mixed-echo pattern.
Ultrasound appears to be a valuable tool for confirming an empty cavity. Lee et al.11 found an empty cavity in 20 of 27 patients with secondary postpartum hemorrhage. Retained placental tissue could be demonstrated in only one case. Shen18 observed an empty cavity in 18 patients and a minimal quantity of residual trophoblastic tissue was found in just one. We found an empty cavity in 18 of the conservatively treated patients and all had an uneventful clinical course.
In conclusion, in patients with secondary postpartum hemorrhage, an echogenic mass in the uterine cavity combined with a cavity diameter above the 90th percentile is associated with retained placental tissue. An empty cavity and fluid in the cavity are consistent with the normal involution process of the puerperal uterus and predict a good clinical outcome. A mixed-echo pattern is also consistent with a normal involution process, and in the majority of cases these patients will have an uneventful course with conservative treatment. Although a mixed-echo pattern might be associated with retained placental tissue in a few cases, it is reasonable to use a conservative approach initially and reserve surgical procedures for when conservative treatment fails.