Acta Obstetricia et Gynecologica.

2009; 88: 909
913

ORIGINAL ARTICLE

Surgical versus medical treatment for secondary post-partum

hemorrhage

TOMER FEIGENBERG, YAEL EITAN, HEN Y. SELA, URIEL ELCHALAL,

ASSAF BEN-MEIR & NATHAN ROJANSKY

The Departments of Obstetrics and Gynecology, Hadassah Ein-Kerem Medical Center, The Hebrew University,
Jerusalem, Israel

Abstract
Background. Secondary post-partum hemorrhage (PPH) is defined as any abnormal bleeding from the birth canal occurring
between 24 hours and 12 weeks postnatally. Treatment usually falls into one of the two categories: surgical evacuation of the
uterus or medical treatment. Objective. To compare the two different clinical approaches and the implications on future
fertility. Study design. A retrospective study. Setting. From 1990 to 2002, 168 women diagnosed with late PPH were admitted
to the Hadassah Medical Centers in Jerusalem. The cases were divided into two groups according to the planned initial
treatment: primary surgical treatment vs. primary medical treatment. Results. Primary surgical treatment was associated with
significantly more primary negative events (p0.01). After the primary event, primary surgical treatment was associated
with fewer future deliveries (p0.04) and resulted in increased rate of secondary infertility of borderline significance
(p0.06). Conclusions. Our results show that secondary PPH is related to high rates of immediate and long-term
complications. It is possible that a conservative medical approach for secondary PPH may be superior to surgical treatment.

Key words: Late post-partum hemorrhage, surgical management, medical treatment, infertility

Introduction 40 years. Secondary PPH has received little attention

Secondary post-partum hemorrhage (PPH) is
defined as any abnormal bleeding from the birth
canal occurring between 24 hours and 12 weeks
postnatally (1). Since blood loss is difficult to
assess, the diagnosis is subjective and based on
clinical observation. Normal post-partum blood
loss varies in amount and duration, with a median
span of 30 days (2). The loss may be intermittent
and generally diminishes in 3
6 weeks. Usually,
secondary PPH is defined as heavy bleeding with
an abrupt onset, lasting from 7 to 14 days following
delivery.
The incidence of secondary PPH in developed
countries ranges in different studies from 0.5 to 2%
(3
5). Despite a dramatic improvement in perinatal
care, its occurrence has not changed over the last
or research study, although cases of severe morbidity
and even mortality are still reported (6). Treatment
usually falls into one of two categories: surgical
evacuation of the uterus or medical treatment with
various drugs. In 2005, the Cochrane database
published its review on treatment for secondary
PPH (7). No prospective randomized controlled
trials addressing the issue were found, emphasizing
the lack of research focusing on this problem. In
addition, complications of secondary PPH and its
impact on future fertility have not been sufficiently
investigated.
The aim of our study was to compare initial
surgical intervention and medical treatment with
regard to immediate complications and future
reproductive performance.

Correspondence: Tomer Feigenberg, Department of Obstetrics and Gynecology, Hadassah-Hebrew University Medical Center, Ein-Kerem, POB 12000,
Jerusalem 91120, Israel. E-mail: feigenberg@013.net
Presentation information: The 2nd Asia Pacific Congress on Controversies in Obstetrics Gynecology & Infertility, The Academy of Clinical Debates &
Controversies in Medicine, Shanghai, China, November 8
11, 2007.

(Received 3 September 2008; accepted 2 June 2009)

ISSN 0001-6349 print/ISSN 1600-0412 online # 2009 Informa UK Ltd. (Informa Healthcare, Taylor & Francis AS)
DOI: 10.1080/00016340903093559
910 T. Feigenberg et al.
Material and methods
Secondary PPH was defined as any excessive vaginal
bleeding occurring between 24 hours after the end of
the third stage of labor and up to 12 weeks later, in
an amount sufficient to prompt hospitalization. Data
were collected from the computerized archives and
medical records were reviewed for relevant data.
From 1990 to 2002, 168 women diagnosed with late
PPH were admitted to the Hadassah Ein-Kerem and
Mount Scopus Hebrew University hospitals, two
major referral medical centers in Jerusalem, Israel.
The patients were divided into two groups accord-
ing to the planned initial treatment: 118 women were
initially treated medically, while 50 had, as their
primary treatment, surgical evacuation of the uterus.
The mean time between delivery and the day of
admission was 16.8 days for the group treated
conservatively, and 27.9 days for the group treated
surgically (t 3.8, df 68, p0.0003). The mean
age was 28.5 years for the women treated medically
and 29.9 for the women treated surgically (t 1.4,
df 166, p 0.16). Women who were treated medi-
cally had a mean of three pregnancies prior to the
event of PPH, while women who were treated
surgically had a mean of 2.7 pregnancies (t0.57,
df 166, p 0.57). There was no difference between
the groups in mean number of deliveries or
miscarriages prior to the event of PPH (t0.42,
df 166, p 0.68; t 1.01, df 166, p 0.32).
A negative primary outcome was defined as any of
the following severe events:
1. need for blood transfusion for women whose
hemoglobin levels were higher than 80 g/L
upon admission and dropped during hospitali-
zation;
2. hysterectomy;
3. perforation of the uterus during primary or
secondary evacuation of the uterus; and
4. need for broad-spectrum antibiotics due to
systemic infection.
A negative secondary outcome was defined as any
of the following events:
1. need for second evacuation of the uterus or any
evacuation of the uterus if one was not initially
planned;
2. re-admission to the hospital after discharge;
3. hospitalization for more than three days; and
4. a drop in hemoglobin level of more than 20 g/L
for those who did not receive blood.
A telephone questionnaire was completed for 131
(78%) of the women, i.e. 90 women from the
primary medical treatment (76.3%) and 41 (82%)
from the primary surgical group (Fisher’s exact
p0.54). The mean interval from the primary event
to the interview was 88.3 months for the conserva-
tive group, and 81.6 months for the surgical group
(not significant). The women were questioned about
side effects of the primary treatment, a need for
further intervention after discharge, the desire for
future pregnancy after the event, the number of
future pregnancies, any secondary infertility after the
primary event, and need for fertility treatment.
Fisher’s exact test was used for comparisons of
variables with two categories between treatment
groups. For other categorical variables the chi-
squared test was applied. Numeric variable differ-
ences were tested by Student’s t-test. Analysis of
co-variance was performed to control for various
confounders. Microsoft Excel 2003 and the SPSS/
PC package were used for statistical evaluation.
A p B0.05 was considered significant. The local
institutional review board (IRB) approved exemp-
tion of consent for the use of retrospective files.
Results
Any primary negative outcome (hysterectomy, blood
transfusion, perforation of the uterus, and systemic
infection) was identified in 16.5% of the women who
had medical treatment compared to 37.5% of the
women who had primary surgical intervention (Fish-
er’s exact, p 0.01). One woman needed a hyster-
ectomy because of uncontrollable bleeding during
curettage. Perforation of the uterus was identified
after primary surgical evacuation in two additional
cases. Of the subgroup of women who were admitted
to the hospital with a hemoglobin level higher than
80 g/L and normal blood pressure, as many as 20%
of the surgically treated women vs. 9.3% of the
women who were treated medically required blood
transfusion after admission (Fisher’s exact, p 
0.07). Primary negative outcomes are shown in
Table I.
Secondary negative outcomes were observed in
59.1% of the women who had conservative medical
treatment, and in 53.1% of the women who had
primary surgical treatment. The secondary negative
outcomes are displayed in Table I. Evacuation after
failure of medical treatment was more common than
secondary evacuation after surgical intervention
(26.3% vs. 8.0%; Fisher’s exact, p0.01).
Re-admission to the hospital after discharge was
also more common in the conservative medical
group, although this was not significant (15.5% vs.
8.2%; Fisher’s exact, p 0.32). Almost half of the
 Secondary post-partum hemorrhage 911
Table I. Comparison of primary and secondary negative outcomes of women who were treated medically vs. surgically.

Primary medical (n118) Primary surgical (n 50)

Event No. Percentage (%) No. Percentage (%) p-Value

Blood transfusiona 11 9.3 10 20.0 0.07

Broad-spectrum antibioticsa 10 8.5 9 18.4 0.11
Hysterectomya 0 0 1 2.0 0.30
Perforationa 0 0 2 4.1 0.09
Any negative primary outcomea 19 16.5 18 37.5 0.01
Re-admissiona 18 15.5 4 8.2 0.32
Hospitalizationtwo daysa 48 41.0 22 44.0 0.73
Secondary surgical evacuationa 31 26.3 4 8.0 0.01
Hemoglobin drop20 g/La 16 13.6 5 10.0 0.62
Any negative secondary outcomea 68 59.1 26 53.1 0.49
a
Fisher’s exact test.

women in both groups were hospitalized for more
than two days.
There was no difference between the two groups
in the desire to conceive after the primary event:
74.4% of the women from the primary medical
group vs. 65% of the women from the primary
surgical group (Fisher’s exact, p0.3).
Of those who tried to conceive, 12.1% of the
women from the medical group, and 30.8% of the
surgical group suffered from secondary infertility
(Fisher’s exact, p 0.06).
Various infertility treatments to achieve pregnancy
were required in 10.8% of the women from the
conservative group vs. 27.8% of the women in the
surgical group (Fisher’s exact, p 0.13). In those
women who tried to conceive after the event, the
mean number of deliveries following the event of
secondary PPH varied between the two groups.
Women after primary medical treatments had a
mean of 2.8 deliveries, while women after primary
surgical treatments had a mean of only 1.5 deliveries
(t 2.95, df 68.5, p 0.004). Because of differ-
ences in the mean elapsed time between the event
and the telephone questionnaire, an analysis of
co-variance was performed. For the average time of
86.4 months, women from the conservative group
had 2.23 deliveries, while those from the surgical
group had 1.6 (p0.05). There was no difference in
the number of miscarriages between the two groups.
An adhesiolysis procedure was needed in 2.5% of
the women after medical treatment, and in 16% of
the women after surgical treatment (Fisher’s exact,
p0.003). Table II shows the main findings of the
telephone questionnaire.
Mode of delivery leading to the secondary PPH
did not differ between the two groups. History of
primary PPH after birth, manual separation of the
placenta, and manual exploration of the uterine
cavity after birth were more common in the surgical
group. Table III shows the main characteristics of the
deliveries preceding the event.
Ultrasonographic examination was performed
before the procedure in 45 out of 50 women in the
surgical intervention group. A suspicion of retained
placental tissue was raised in 38 (84.4%). Histolo-
gical examination was obtained in all of these cases
and confirmed placental remnants in 35 (77.8%;
85.7% sensitivity, 20% specificity).
Discussion
To the best of our knowledge this may be the first
study comparing medical and surgical approaches
for the treatment of late PPH. Our results demon-
strate that medical treatment is related to less
primary negative outcomes. Furthermore, medical
treatment resulted in less secondary infertility and
better reproductive outcomes.
Secondary PPH is associated with serious mor-
bidity and even mortality (6). In our study 22% of
the women who were hospitalized had any one of
four life-threatening primary negative effects. The
incidence of uterine perforation after curettage was
4% in those who were treated surgically, consistent
with the reported incidence (5). Minor secondary
complications were observed in over half of the
women, with no differences between the groups,
although women who were treated medically needed
more secondary evacuations of the uterus and had
higher rates of re-admission to the hospital.
The etiology of secondary PPH is variable and
includes sub-involution of uteroplacental vessels,
retained placental tissue, placenta accreta, infection,
trauma, and coagulopathies (2). Establishing the
correct cause is difficult and sometimes only possible
after evacuation of the uterine cavity. It seems that
the most common causes seen at pathological
examination of tissue collected after curettage or
912 T. Feigenberg et al.
Table II. Results of telephone questionnaire regarding the impact of the two different treatment modalities on fertility, after discharge from
the hospital.

Primary medical Primary surgical

Event n Percentage (%) n Percentage (%) p-Value

a
Answered the questionnaire 90 76.3 41 82.0 0.54
Tried to conceive after the eventa 67 74.4 26 65.0 0.30
Secondary infertilitya 8 12.1 8 30.8 0.06
Need for adhesiolysisa 3 2.5 8 16.0 0.03
Infertility treatmentsa 4 10.8 5 27.8 0.13
Mean no. pregnanciesb 2.24 1.85 0.20
Mean no. deliveriesc 2.28 1.5 0.04
Mean no. miscarriagesd 0.4 0.42 0.89
a
Fisher’s exact.
b
t 1.30, df69.9.
c
t 2.95, df68.5.
d
t 0.13, df91.

hysterectomy, are sub-involution of uteroplacental
vessels and retained placental fragments (8). The
usefulness of ultrasound examination in order to
distinguish retained placental fragment from blood
clots is controversial. Edwards and Ellwood (9) tried
to define the ultrasonographic appearance of the
uterine cavity in women with a normal puerperium.
Their work revealed an echogenic mass in 51% of
the women with normal post-partum bleeding after
seven days and 21% after 14 days, casting doubt on
the significance of the finding of an echogenic mass
in the uterus during the post-partum period.
Although some investigators have found ultrasono-
graphic evaluation to be a useful and accurate tool to
distinguish retained placental remnants from the
situation of an atonic uterus in such cases (10),
others have found it to have a limited diagnostic
accuracy, not superior to clinical assessment (11). In
our study ultrasonographic examination had a high
sensitivity for the detection of placental remnants,
but the specificity was low.
Considering the difficulties a clinician faces in
establishing the correct cause of secondary PPH, and
the lack of data on the efficacy of each treatment
modality, it is not surprising that choosing the
appropriate therapy is difficult. Complications of
surgical evacuation of the uterus include short-term
complications as well as jeopardize to future fertility.
In 1948, Joseph G. Asherman (12) described the
syndrome of intra-uterine adhesions. Causes of
Asherman’s syndrome include trauma to the uterine
cavity by curettage, especially after birth, and infec-
tion (13). In their work on the incidence of intra-
uterine adhesions, Westendrop et al. found that 40%
of women had intra-uterine adhesions after second-
ary removal of placental remnants after birth, or
repeated curettage after incomplete abortion (14).
We found that 16% of the women who later tried to
conceive after surgical intervention, and 2.5% of
those who had primary medical treatment, needed
hysteroscopy or curettage for adhesiolysis. Compli-
cations of medical management protocols are less
well studied, and the impact on future fertility is not
known. This observational prospective-historical
study indicates that both medical and surgical
treatments for secondary PPH have high rates of

Table III. The main characteristics of the deliveries leading to the event of late PPH.

Primary conservative Primary surgical

Event n Percentage (%) n Percentage (%) p-Value

a
Primary PPH 15 12.8 14 28.0 0.03
Manual separation of the placentaa 5 4.2 8 16.0 0.02
Manual exploration of uterusa 6 5.1 7 14.0 0.06
Mode of deliveryb 0.58
Vaginal delivery 92 78.0 42 84.0
Cesarean section 16 13.6 4 8.0
Vacuum extraction 10 8.5 4 8.0
a
Fisher’s exact.
b 2
x 1.082, df2.

both immediate and late complications, and may
jeopardize future fertility. Due to the possibility of
selection bias between the two groups in our study,
and the lack of randomized trials comparing differ-
ent treatment modalities for secondary PPH, we
believe that a well-designed prospective study is
needed, but meantime medical treatment for sec-
ondary PPH might be the preferable option when-
ever possible.
Declaration of interest: The authors report no
conflicts of interest. The authors alone are respon-
sible for the content and writing of the paper.
References
1. Rome RM. Secondary postpartum hemorrhage. Br J Obstet
Gynecol. 1975;82:289
92.
/ /
2. Annemarie N, Thornton S. Secondary postpartum hemor-
rhage. J Obstet Gynecol. 2002;22:119
22.
/ /
3. King PA, Duthie SJ, Dong ZG. Secondary post partum
hemorrhage. Aust N Z J Obstet Gynecol. 1989;29:394
8.
/ /
4. American College of Obstetrics and Gynecology. Diagnosis
and management of postpartum hemorrhage. Inter J Gynecol
Obstet. 1990;36:159
63.
5. Fatemeh H, Mackenzie IZ. Secondary postpartum hemor-
rhage: incidence, morbidity and current management. Br J
Obstet Gynecol. 2001;108:927
30.
/ /
Secondary post-partum hemorrhage 913
6. Department of Health. Report on confidential enquiries into
maternal deaths in England and Wales 1991
1993. London:
HMSO, 1996. pp. 32
47.
7. Alexander J, Thomas P, Sanghrea J. Treatments for secondary
postpartum hemorrhage. Cochrane Database Syst Rev. 2002:
CD002867.
8. Khong TY, Khong TK. Delayed postpartum hemorrhage: a
morphologic study of causes and their relation to other
pregnancy disorders. Obstet Gynecol. 1993;8291:17
22.
/ /
9. Edwards A, Ellwood DA. Ultrasonographic evaluation of the
postpartum uterus. Ultrasound Obstet Gynecol. 2000;16: / /
640
3.
10. Mulic-Lutvica A, Axelsson O. Ultrasound finding of an
echogenic mass in women with secondary postpartum he-
morrhage is associated with retained placental tissue. Ultra-
sound Obstet Gynecol. 2006;28:312
9. / /
11. Neil AMC, Nixon RM, Thornton S. A comparison of clinical
assessment with ultrasound in the management of secondary
postpartum hemorrhage. Eur J Obstet Gynecol Rep Biol.
2002;104:113
5.
/ /
12. Asherman JG. Amenorrhea traumatica atretia. J Obstet
Gynecol Br Emp. 1948;55:23
30.
/ /
13. Shenker JG, Margaliot EJ. Intrauterine adhesions: an updated
appraisal. Fertil Steril. 1982;37:593
610.
/ /
14. Westendrop ICD, Ankum WM, Mol BWJ, Vonk J. Prevalence
of Asherman’s syndrome after secondary removal of placenta
remnants or a repeated curettage for incomplete abortion.
Hum Reprod. 1998;13:3347
50.
/ /