Ultrasound of the Postpartum Uterus
Prediction of Retained Placental Tissue
Barbara S. Hertzberg, MD#James D. Bowie, MD
We reviewed ultrasound images an 53 postpartum
patients referred for possible retained products of con-
ception and correlated specific ultrasound patterns
with clinical and pathologic follow-up. The most com-
mon finding in patients with retained placental tissue
was an echogenic mass in the uterine cavity, seen in 9
of 11 patients with pathologically proven retained pla-
cental tissue. In the remaining 2 patients with patho-
logically confirmed retained placenta, a heterogeneous
mass was seen in the uterine cavity at some point
during the course of serial sonography. Retained pla-
cental tissue was unlikely when ultrasound demon-
strated a normal uterine stripe (n = 18), endometrial
fluid (n = 6), or hyperechoic foci in the uterine cavity
U
ltrasound is often used to evaluate for possible
retained products of conception in the patient
with excessive postpartum hemorrhage. Cu-
rettage is the treatment of choice for retained placental
or fetal tissue. However, conservative management is
preferred in the absence of retained placental or fetal
parts since curettage of the postpartum uterus is
fraught with potential complications including perfo-
ration and synechia formation. 1- 3 Prior studies of the
ultrasound evaluation for retained products of concep-
tion resulted in conflicting conclusions regarding the
sonographic appearance of retained placental tissue
and the ability of ultrasound to differentiate between
retained placenta and blood clots or necrotic debris. 4 9
In an attempt to determine if currently available trans-
abdominal ultrasound equipment detects any findings
Received September 9, 1990, from the Department of Radiology,
Duke University Medical Center, Durham, North Carolina. Revised
manuscript accepted for publication February 8, 199};
Address: correspondence and reprint requests to Dr. Barbara S.
Hertzberg, Duke University Medical Center, Department of Radiol-
ogy, Box 3808, Durham, NC 27710.
@ 1991 by the American Institute of Ultrasound in Medicine JUltrasound Med 10:451- 456, 19910278; 4297/ 91/ $3.50
without an associated mass (ri = 17). The latter finding
was often associated with recent uterine instrumenta-
tion. The sonographic appearance of retained placental
tissue is variable, but detection of an echogenic mass
in the uterus strongly supports the diagnosis. A heter
ogeneous mass is sometimes caused by retained pla-
centa, but can also be secondary to blood clots or
infected or necrotic material in the absence of placental
tissue. Sonographic evaluation for retained products of
conception is best done before uterine instrumentation
to avoid confusion with iatrogenically introduced air.
JCEY woaos: Placenta, ultrasound studies; Uterus, ultra-
sound studies; Postpartum uterus; Retained placenta .
UUltrasound Med 10:451, 1991)
characteristic of retained products of conception, we
retrospectively reviewed the sonograms of patients re-
ferred for evaluation for retained products and corre-
lated the ultrasound findings with patient outcomes.
METHODS AND MATERIALS
The study group was comprised of 53 consecutive
patients referred for ultrasound examinations to eval-
uate for retained products of conception (RPC) after
second and third trimester deliveries during the period
from January 1985 through October 1988. Of this
group, 27 patients had undergone vaginal delivery of
a living fetus, 10 had cesarean sections, 9 had sponta
neous miscarriages or stillbirths, and 7 had therapeutic
abortions. The clinical problems leading to ultrasound
examinations included bleeding, pain, fever, and
cramps. The time interval between delivery and eval-
uation by ultrasound ranged from 1 day to 5 months
tp tu
pos ar m.
Ultrasound images were reviewed retrospectively
without knowledge of patient outcome. For each case,
452 ULTRASOUND OF POSTPARTUM UTERUS
the thickness of the uterine cavity was measured in
anteroposterior (AP) dimension, perpendicular to the
endometrial stripe, and the sonographic appearance of
the uterine cavity was classified into one of the follow-
ing five categories:
1. Normal uterine "'stripe"': linear pattern of endo-
metrial echoes without endometrial fluid, in-
tensely echogenic foci, or endometrial thickening
(thickness of uterine cavity less than 1 5 cm)
(Fig. lA).
2. Endometrial fluid (Fig. 1B).
3. Echogenic mass: echogenic material enlarging
uterine cavity (AP measurement 1.5 cm or
greater), sometimes with a '"stippled pattern of
scattered punctate intensely echogenic foci (Fig.
JC).
4. Hyperechoic foci/no mass: scattered intensely
echogenic foci in _the endometrial cavity; no evi
dence of a uterine cavity mass (cavity less than
1.5 cm in AP diameter) (Fig. 10).
5. Heterogeneous mass: endometrial cavity enlarged
(1.5 cm or greater) by heterogeneous material
exhibiting a mixed pattern of echogenic and echo-
penic areas (Fig. 1E).
When more than one study was available for a
patient, all postpartum ultrasound examinations were
reviewed. Such cases were categorized based on find-
ings from the first postpartum study, and interval
changes occurring on subsequent examinations were
noted.
All studies were done on commercially available
real-time equipment (Acuson 128, Mountain View, CA;
Diasonics ADA400, Milpitas, CA; and ATL Mark 3,
Bothell, WA) with 3.5- and 5.0-MHz sector and linear
transducers using a transabdominal approach. None of
the scans in the series was done via a vaginal approach.
Patients were initially scanned with a full urinary blad-
der. Postvoid views were obtained in selected patients
when an overdistended urinary bladder compressed
the uterus, interfering with sonographic evaluation of
the uterine contents.
The medical records of all patients were reviewed
independent of ultrasound findings to determine clin-
ical course and pathologic correlation. RPC were con-
sidered pathologically proven only when material ob-
tained at the time of curettage or hysterectomy revealed
chorionic villi or fetal parts. Identification of blood
clots, decidua, or necrotic material was not considered
sufficient to diagnose RPC. Additionally, for each case,
we recorded whether uterine instrumentation (curet-
tage, insertion of ring forceps, uterine sounding, or
endometrial culture) had been performed in the 2 days
immediately before ultrasonography. Subsequently the
ultrasound findings were correlated with the clinical
course and pathologic findings.
J Ultrasound Med 10:451- 456, 1991
RESULTS
The ultrasound appearance of the uterine cavity in the
53 patients referred for evaluation for RPC was cate-
gorized as follows: normal uterine stripe, 18 cases;
endometrial fluid, 6 cases; echogenic mass, 10 cases;
hyperechoic foci/no mass, 17 cases; and heterogeneous
mass, 2 cases.
Histologic findings were available for a total of 16
patients: 2 with a normal uterine stripe, 1 with endo-
metriaJ fluid, 9 with an echogenic mass, 3 with hyper
echoic foci/no mass, and 1 with the heterogeneous
mass pattern. Of the 16 patients in whom histologic
correlation was available, 11 had pathologically proven
RPC.
Ultrasound Patterns Correlated with Outcome
Table 1 summarizes the correlation between ultrasound
patterns and patient outcome, subdividing patients into
the following categories: (1) patients likely to have had
RPC based on either (a) pathologic proof of RPC or (b)
persistent clinical symptomatology suggestive of RPC
(no histologic data available) and, (2) patients unlikely
to have had RPC based on (a) histologic material
revealing the absence of RPC or (b) subsequent clinical
course not suggestive of RPC. Details of specific cases
are discussed below.
1. Normal Uterine Stripe. None of the 18 patients
exhibiting a normal uterine stripe had pathologically
proven RPC. The histologic diagnoses for the 2 patients
with pathologic proof were '"blood clot, fibrin, and
necrotic tissue (11 - 1) and "'inactive endometrium (11
- 1). Of the 16 patients without histologic data, 12 had
a completely benign clinical course and the remaining
4 had the following postpartum complications: a rec
tovaginal fistula (n = 1), systemic lupus erythematosus
(11 = 1), pseudomembranous colitis (u - 1), and a
nonspecific collagen vascular disease (n =' 1). None of
these patients had persistent clinical symptoms sug-
gesting RPC.
2. E11do111etrial Fluid. None of the 6 patients in this
category had evidence of RPC. One had pathologic
confirmation (histologic diagnosis = decidualized en-
dometrium with necrosis/no chorionic villi), 4 had a
benign clinical course, and the 6th patient did not
return for follow-up.
3. Ecl1oge11ic Mass. There were 10 patients in this
category, 9 of whom had pathologically confirmed
RPC. The 10th had dysfunctional uterine bleeding and
pelvic pain for 10 months postpartum and was classi-
fied as having persistent clinical symptomatology
suggestive of RPC, without histologic proof.
In 8 of the 10 patients in this category, the echogenic
mass in the uterine cavity displayed a stippled"' ap-
pearance, containing scattered punctate highly echo-
J Ultrasound Med 10:451-456, 1991 HERTZBERG AND BOWIE 453

\ ..._
~ .....,

- '"
-
. _ /~ l .
._ ....
~
\
'I
~ {- ir
- ,..,.
. ,_ '4- ~ . Z~
.....
..
..t
~~~.t. )

A
~
~
.. .II}
""-
.-AJI'

E
Figure 1 Classification of ultrasound appearance of the
uterine cavity. A, Nonna! uterine stripe: sagittal monogram
reveals a linear pattern of endometrial echoes (arrows) with ~
out fluid or hyperechoic foci. 8, Endometrial fluid: arrows
depict a small amount of fluid in the uterine cavity. C,
Echogenic mass: echogenic material enlarges the uterine cav-
ity (arrows). D, Hyperechoic foci/ no masses: intensely echo-
genic foci are seen in the uterine cavity (arrows), without an
accompanying mass. E, Heterogeneous mas&: axial sonogram
reveals heterogeneous material enlarging the uterine cavity
(arrows) .

c interval increase in echogenidty of the mass in the

genie foci. One of the patients with a stippled uterine
cavity mass underwent serial sonography before uter-
ine curettage. Ultrasound scans at both 4 and 7 weeks
after the initial examination revealed a generalized
uterine cavity, obscuring the stippled appearance seen
on the initial scan, and suggesting progressively in
creasing calcification or fibrosis (Fig. 2, A and B). The
curettage specimen consisted of "'extensively hyalinized
and calcified placental fragments:
454 ULTRASOUND OF POSTPARTUM UTERUS J Ultrasound Med 10:451-456, 1991

Table 1: Ultrasound Patterns Correlated with Outcome in Patients with Suspected RPC
RPC Likely Based on RPC Unlikely Based on
Persistent Follow-up
Clinical course
Pathologic symptomatology NoRPC Unavailable
(no histology
proof (no histology pathologically available)
available)
Normal uterine stripe 0 0 2 16 0
(11 =18)
Endometrial fluid 0 0 1 4 1
(11 - 6)
Echogenk mass 9 0 0 0
(11 ... 10)
Hyperechoic foci/ no mass 1 (converting 0 2 14 0
(11 17) to heteroge-
neous mass)
Heterogeneous mass 1 0 0 1 0
(11 = 2)

4. Hyperechoic Foci/No Mass. Of the 17 patients
exhibiting hyperechoic foci in the uterine cavity with-
out evidence of a mass, only 1 had pathologically
proven RPC. The initial scan on this patient was done
immediately after uterine instrumentation, and a son-
ogram performed 2 days later revealed a heterogeneous
mass pattern without the hyperechoic foci seen on the
earlier study.
Histologic correlation was obtained for 2 other pa-
tients in this group. The pathologic diagnoses in these
2 patients were acute inflammatory exudate, microab-
scesses, and degenerating decidua .. (11 - 1) and de-
=
cidua with extensive necrosis/ no villi.. (11 1). In one
additional patient a computed tomographic scan re-
vealed air in the uterus, endometrial cultures were
positive for Clostridium, and symptoms resolved with
antibiotic therapy. All other patients in this group
improved with antibiotic or Methergine therapy and
did not exhibit persistent symptomatology of retained
products.
5. Heterogeneous Mass. One of the 2 patients who
initially displayed the heterogeneous mass pattern had
pathologically proven RPC. The other patient had a
benign clinical course and did not undergo uterine
evacuation. A third patient (described above in the
section entitled Hyperechoic Foci/ No Mass..) exhibited
the heterogeneous mass pattern on a follow-up scan
but was not included in this group because the initial
ultrasound study had been classified as hyperechoic
foci/no mass. This latter patient also had histologically
proven retained placental tissue.
Uterine Cavity Thickness
As a group, patients with pathologically proven RPC
showed greater AP thickness of the uterine cavity
(mean 21.4 mm) compared with patients without his-
tologically confirmed RPC (mean 8.3 mm), and this
difference was statistically significant (P < 0.001 using
a two-sample t-test). However, the range of uterine
cavity thicknesses for the 2 groups overlapped, varying
from 11 to 29 mm in the group with proven RPC and
from 3 to 29 mm in the group without histologic
confirmation of RPC.
Recent Uterine Instrumentation
Even in the absence of pathologically confirmed re-
tained placenta, hyperechoic foci were commonly seen
in the uterine cavity of patients who had undergone
recent uterine instrumentation. Uterine instrumenta-
tion had been performed on 13 patients during the 2
days immediately before ultrasonography. The appear-
ance of the uterine cavity in these patients fell into the
following categories: hyperechoic foci/no mass, 11 8;
hyperechoic foci/no mass, converting to heterogeneous
mass pattern, 11 = 1; normal uterine stripe, 11 - 3; and
echogenic mass, 11 = 1. Of the 8 patients with recent
instrumentation who exhibited hyperechoic foci/no
mass, none had pathologically confirmed RPC.
Discussion
Excessive postpartum hemorrhage occurs secondary to
a variety of etiologies, the most common being uterine
atony, an undetected laceration of the cervix, vagina,
or perineum, and retained products of conception.2. 10
When conservative forms of management fail to slow
blood loss to expected levels, ultrasound is often re-
quested to evaluate for the presence of retained placen-
tal tissue. However, data on the sonographic appear-
J Ultrasound Med 10:451- 456. 1991
A
B confirmed retained placenta. We believe that the son
Figure 2 A. Initial ultrasound study on a patient with path-
ologically proven retained placental tissue demonstrates an
echogenic mass enlarging the upper portion of uterine cavity
(arrows) and exhibiting a stippled pattern of scattered hy-
perechoic foci. 8, Scan perfonned 7 weeks later reveals a
generalized increase in the echogenicity of the placental tissue
retained in the uterine cavity (arrows). obscuring the previ-
ously seen scattered hyperechoic foci . Extensively hyalinized
and calcified placental fragments were removed by curettage.
ance of RPC in the patient with postpartum hemor
rhage are limited. The few studies available were done
mainly in the 1970s and early 1980s and presented
conflicting data regarding the appearance of RPC and
the ability of ultrasound to differentiate between re-
tained placenta and blood clots or necrotic debris. 4 "9
In 1972, Robinson 4 diagnosed retained intrauterine
tissue in 29 of 112 postpartum patients based on -a
HERTZBERG AND BOWIE 455
large volume of echoes extending towards the margin
of the uterine outline. Curettage specimens for 25 of
6
these patients revealed chorionic villi in 16, and he
concluded that a positive ultrasound report is sufficient
indication for exploration of the uterus. The following
year, Malvern et a1 5 described a series which included
28 patients who were thought to have RPC because of
detection of circumscribed echoes lying in the line of
the uterine cavity. Retained placenta was found in
only 8 of the 25 patients who underwent curettage.
They postulated that the high incidence of false-posi
tive results was caused by the inability of ultrasound
to distinguish between retained placenta and thickened
decidua or organized blood clots. Subsequently, Lee et
al" diagnosed retained placenta in 9 of 56 patients (8
confirmed histologically) based on sonographic detec-
tion of a solid mass containing calcium or multiple
dense echoes. Other studies performed following first
and second trimester therapeutic abortions also pre-
sented conflicting findings.7- 9
Jn the current study, the ultrasound findings in the
11 patients with pathologically proven RPC were: ech-
ogenic mass (11 .,. 9), heterogeneous mass (11 = 1), and
hyperechoic foci/ no mass, converting to a heteroge-
neous mass pattern (u = 1). Detection of an echogenic
mass in the uterine cavity was strongly associated with
RPC, a finding similar to that of Lee et al. However, in
the latter series there was no evidence of RPC in 5 of
5 patients exhibiting a heterogeneous pattern of solid
and fluid components, leading to the conclusion that
such a pattern does not indicate retained placenta. In
contrast, although our numbers were limited to 3 cases
of heterogeneous uterine cavity masses (2 documented
on the initial scan and 1 at the time of a follow-up
study), 2 of these 3 were associated with histologically
ographic appearance of retained placenta is more vari-
able than previously recognized, probably depending
on the relative quantities of calcified and uncalcified
placental tissue, blood clots, and necrotic decidua.
Clinically significant retained placental tissue is un-
likely when ultrasound demonstrates a normally apA
pearing uterine stripe, isolated endometrial fluid, or
hyperechoic foci without an associated mass. PathoA
logic correlation was not available in many patients
exhibiting these latter findings since the potential for
complications from curettage of the postpartum uterus
precludes uterine evacuation unless there is a strong
suspicion of retained products. For those patients in
whom histologic correlation is unavailable, the absence
of retained placental tissue is not known with certainty.
Although it is possible that small clinically insignificant
fragments of retained tissue were occasionally missed,
as long as placental fragments are clinically insignifi-
cant and do not require therapy, it may be unimportant
456 ULTRASOUND OF POSTPARTUM UTERUS
if small fragments are present in some patients with
negative ultrasound examinations for RPC.
Echogenic foci seen in the uterus after instrumenta-
tion are frequently caused by air rather than by retained
placenta, and the significance of such foci may lie in
their tendency to obscure an existing mass in the uter-
ine cavity or be mistaken for echogenic fragments of
retained placenta or air produced by gasTforming or-
ganisms. Whenever possible, sonography should be
performed before uterine instrumentation in patients
with suspected RPC. If this is not possible, the physi-
cian interpreting the ultrasound study must be aware
of any recent uterine manipulations. If intrauterine air
poses a diagnostic dilemma, follow -up sonography can
be performed to determine if it resolves.
REFERENCES
L Madrazo BL: Postpartum sonography. /11: Sanders RC.
James AE (eds): The Principles and Practice of Ultraso-
nography in Obstetrics and Gynecology, 3td ed. Nor
walk, CT, Appleton-Century-Crofts, 1985, p 449
2. King PA, Duthie SJ, Dong ZG, Ma HK: Secondary post-
J Ultrasound Med 10:451-456, 1991
partum haemorrhage. Aust NZ J Obstet Gynaecol 29:394,
1989
3. Rome RM: Secondary postpartum haemorrhage. Br J
Obstet Gynaecol 82;289, 1975
4. Robinson HP: Sonar in the puerperium- A means of
diagnosing retained products of conception. Scot Med J
17:364, 1972
5. Malvern J, Campbell S, May P: Ultrasonic scanning of
the puerperal uterus following secondary postpartum
haemorrhage. J Obstet Gynaeco.I Br Commonw 80:320,
1973
6. Lee CY, Madrazo B, Drukker BH: Ultrasonic evaluation
of the postpartum uterus in the management of postpar
tum bleeding. Obstet Gynecol 58:227, 1981
7. Wilcox FL, Lawler L: Retained products of conception.
Letter to the Editor. NZ Med J 99:912, 1986
8. Sanders RC: Postpartum diagnostic ultrasound. In: Sand
ers RC, James AE (eds): Obstetrics and Gynecology, 1st
ed. New York, AppletonCenturyTCrofts, 1977, p 285
9. Stone M, Elder MG: Evaluation of sonar in the prediction
of complications after vaginal termination of pregnancy.
Am J Obstet Gynecol 120:890, 1974
to. Madrazo BL, Baker ME: Postpartum sonography. /II:
Putman CE, Ravin CE (eds) ~ Textbook of Diagnostic
Imaging. Philadelphia, WB Saunders, 1988, p 2034