The surgical approach to postpartum haemorrhage

by **mandible** Fri Mar 05, 2010 2:40 am

Introduction
Postpartum haemorrhage (PPH) has always been a
major cause of maternal mortality and morbidity;
its incidence is rising for reasons that remain
obscure.1 Fortunately, techniques for dealing with it
have improved so that mortality from this cause
continues to decline. Some PPHs are traumatic
(vaginal and uterine lacerations). The primary
approach to dealing with these is surgical correction
of the defect; the techniques required are as varied as
the lacerations themselves. This article deals with
the surgical approach to the more generic causes of
PPH, namely uterine atony, and the less common
conditions of uterine inversion, placenta praevia
and placenta accreta. It should be noted that, as with
many emergency surgical procedures, there are very
few systematic studies of their use and no
randomised trials of efficacy. Long-term follow-up
data are also very few. Inevitably, therefore, this
article relies substantially on anecdote and the
author’s personal experience and this should be
borne in mind if any of the techniques are adopted
by the reader.
Uterine atony
Failure of the uterus to contract effectively
following the delivery of the baby is the commonest
cause of massive PPH. There is no accepted
definition of massive PPH; for the purposes of this
article I have defined it as any case with continuing
haemorrhage despite the ‘usual’ treatment, such as
intravenous oxytocin (Syntocinon,Alliance
Pharmaceuticals Ltd, Chippenham,Wilts, UK)
10 iu, ergometrine 0.5 mg 2, carboprost 0.25 mg
intramuscularly (up to 6) and misoprostol
200 micrograms 5 rectally. (The use of activated
VIIa is controversial and currently not supported
by controlled trials.2) The surgical techniques that
can then be employed are listed in Box 1, in the
order in which they are commonly tried.
Bimanual compression,with one hand (made into
a fist) in the vagina and the other compressing the
uterus using the other hand to press downwards
onto the uterus through the mother’s abdomen,
is often effective at staunching the flow, at least
temporarily. It allows a respite during which blood
can be crossmatched and other resources
marshalled. It is tiring to maintain adequate
compression and it is usually necessary for the
surgeon and their assistants to take turns at
5-minute intervals, if satisfactory compression is to
be maintained. If this is insufficient, compression of
the lower abdominal aorta against the spinal
column at the level of L2–4 can produce an
additional reduction in bleeding by reducing blood
flow to the uterus. Such compression can be
produced by an additional assistant, providing the
mother is not grossly obese. Special ‘anti-shock’
garments have been produced which combine
aortic and uterine compression with compression
of the lower limbs, both to reduce bleeding and to
maintain venous return (Figure 1 and Figure 2).3–5
If bimanual compression appears effective, but
bleeding recommences when compression is
stopped, a traditional approach is to pack the uterus.
Although its effectiveness has been questioned, a
recent review6 has concluded that, performed
properly, this can work well.The key to most
effective use is to insert wide ribbon gauze firmly,
making sure that it is placed initially at the fundus
using a sponge holder and then fed systematically
into the uterus. Each layer must be pressed firmly
home before the next layer is placed.However,
probably more convenient than packing with gauze
is the use of an intrauterine balloon.This technique
was described independently in 2001 by Johanson
et al.7 and Bakri et al.8 The capacity of the balloon
needs to be up to 500 ml, so small balloons such as
those found on Foley catheters are insufficient.
Bakri balloons are now commercially available
in the UK and are manufactured by Cook Ireland
Ltd (Limerick,Republic of Ireland). They contain a
central lumen which ends above the balloon, so that
any blood still being lost above the level of the
uterine tamponade can drain and be measured. In
the absence of a balloon specifically designed for the
purpose, similar tamponade can be obtained using
the stomach balloon of the Sengstaken catheter,
which is stocked in many hospitals for the
management of bleeding oesophageal varices
(although the Sengstaken catheter is effective, the
Bakri balloon is cheaper and simpler to use).Once
inserted fully into the uterus, the balloon should
be inflated with sterile saline until the bleeding is
controlled; commonly, ~300 ml is needed.There
have been no randomised trials of balloon use,
but in a series of 23 cases unresponsive to medical
therapy reported by Dabelea et al.,9 bleeding was
arrested in 21,with only two needing to proceed
to hysterectomy.
If the cervix is fully dilated, there is sometimes
insufficient resistance in the lower segment and
vagina for a pack or balloon to be retained when it
is fully inserted/inflated.This can be countered by
putting in a cervical cerclage (using Prolene or
Mersilene [both made by Ethicon Ltd., Livingston,
UK]) and tightening it to a diameter of ~3 cm; this
provides a platform which maintains the
Box 1
Surgical techniques for controlling
postpartum haemorrhage
• Uterine compression and massage
• Packing/balloon
• Uterine compression suture
• Uterine artery ligation
• Hysterectomy
• Logethotopulos pack
• Internal iliac ligation
• Arterial embolisation
pack/balloon securely in the body of the uterus so
that it can compress the uterus effectively against
its elastic limit. Balloons can also be used for
tamponade in the vagina when there is bleeding
from multiple vaginal lacerations.10
Uterine compression sutures
If packing or balloon tamponade are ineffective, the
next step is to consider direct uterine compression
suturing.The first suggestion of this approach was
by Christopher B-Lynch, ofMilton Keynes Hospital
in the UK,who in 1997 published an account of five
cases11 where compression of the uterus was
achieved following caesarean section using the
technique shown in Figure 3. It requires that the
uterus is opened; the suture compresses the upper
segment but the lower segment remains open. If the
uterus has not previously been opened (e.g. at
caesarean section), a simplified suture can be
inserted, such as square suturing (Figure 4).12
However, there is concern that the square suture
may completely occlude the blood supply to the
uterine muscle within the square, leading to
ischaemic necrosis and subsequent complications
(see below).An important principle is, therefore,
to avoid sutures that apply compression both
vertically and horizontally, but instead use sutures
that are compressive,whether transversely, e.g.
multiple horizontal sutures as recently described
by Hackethal et al.13 (Figure 5) or horizontally as
with the simpler loop suture inserted through the
lower segment and tied at the fundus, as described
by Hayman et al.14 (Figure 6).
As with balloons, there are no randomised
controlled trials of compression sutures, but in a
recent series of 11 cases where the Hayman suture
was used, hysterectomy was only necessary in
one.15 In another series of 31 519 births, uterine
compression sutures were applied in 28 cases;
they were successful in 23 whereas 5 still required
hysterectomy.16
A particular problem is dealing with bleeding from
the lower segment of the uterus. This can be dealt
with by square suturing,12 by a simple horizontal14
or vertical17 loop suture, opposing the anterior to
the posterior walls of the lower segment.An
ingenious variant of this, if the cervix is not fully
dilated, is to invert the lower segment upon itself
before suturing it, thus compressing the bleeding
surfaces without occluding the uterine cavity18
(Figure 7).
Another possibility is to combine the compression
suture with an intrauterine balloon.19 The suture
must be inserted first: clearly, inserting a suture
after the balloon risks puncturing it.Moreover,
once the suture has been inserted, the balloon can
be used to apply counter pressure more effectively.
A series of five such cases was reported by Nelson
and O’Brien20 and this method was effective in all
cases without complications.
All effective interventions have complications and
these are now being reported with all the
approaches described above.An important
practical point is that all compression sutures
should be absorbable.21 The reason for this is that as
the uterus involutes, the sutures will become loose
and, if they are nonabsorbable and do not produce
an inflammatory reaction making them adhere to
the uterine surface, there is always the risk that
233
The Obstetrician & Gynaecologist 2009;11:231–238 Review
2009 Royal College of Obstetricians and Gynaecologists
Figure 1
Non-inflatable garment for the
control of postpartum haemorrhage.
Reproduced with permission from
Miller et al.5
Figure 2
Noninflatable garment for the
control of postpartum haemorrhage.
Reproduced with permission from
Miller et al.5
Figure 3
B-Lynch suture. Reproduced with
permission from Lynch et al.11
loops of free suture will result.This can allow bowel
to become entangled in the loops, resulting in
obstruction. Square suturing results in tight
apposition of the anterior and posterior walls of
the uterus,which can impede drainage of lochia,
resulting in pyometra22 or, in the longer term, in the
formation of synechiae.23Moreover, if the sutures
are placed too tightly, or result in an area of the
uterus being totally deprived of blood supply (for
example, if there is placement of both vertical and
horizontal compression sutures), ischaemic
necrosis will result.24 Even with the B-Lynch suture,
which does not occlude the uterine cavity, necrosis
of the entire uterine corpus has been reported25
and reports of partial necrosis are becoming more
common.26–28 The outcome in subsequent
pregnancies has been little studied, but in seven
reported pregnancies following prior use of uterine
compression sutures, pregnancy and birth was
uncomplicated.16
The needles and suture material used vary
according to the report.The first paper by
B-Lynch11 describes the use of a 70 mm roundbodied
hand needle with a number 2 chromic
catgut suture.Cho et al.12 describe the use of
number 7 or 8 straight needles with number 1
atraumatic chromic catgut.However, catgut is now
rarely used in obstetrics because of its relative lack
of strength and durability.Hayman et al.14 report
the use of either polyglycolic acid (Dexon,
Covidien, Gosport, UK) or Vicryl (Ethicon Ltd.,
Livingston, UK) (number 1 or 2 sutures). They also
mention the use of a straight needle; in fact I
usually bend this manually to a shallow curve,
which makes it easier to insert in the depths of the
pelvis while avoiding puncture of the structures
immediately behind the lower segment. The needle
should ideally be 6 cm long so as to exceed the
combined thickness of the anterior and posterior
lower segment. A shallow curved needle with this
dimension is available commercially. On the other
hand,Hackethal et al. describe the use of an XLH
needle (in conjunction with 0 Vicryl) in which the
curve had been straightened! Ghezzi et al.15 also
recommend using a straight needle with a number
2 polyglactin suture.
Uterine artery ligation
If use of a simple compression suture is
unsuccessful, then ligation of the uterine arteries
can be tried next29 and is often effective. Indeed,
one suspects that uterine artery ligation is
sometimes performed inadvertently when a
lower segment incision extends during a difficult
delivery (for example, of a large baby) and
extensive suturing into the broad ligament is
necessary to control the resultant bleeding.
There appear to be no consequences for future
pregnancies of such ligation, presumably because
a collateral circulation develops from other
vessels (particularly the ovarian arteries) to
compensate.
234
Review 2009;11:231–238 The Obstetrician & Gynaecologist
2009 Royal College of Obstetricians and Gynaecologists
Figure 4
Cho square haemostatic suture12
Figure 5
Multiple U-suture. Reproduced with
permission from Hackethal et al.13
Figure 6
Hayman suture. Reproduced with
permission from Ghezzi et al.15
Internal iliac artery ligation
and aortic compression
There has long been controversy about when ligation
of the internal iliac artery should be attempted.30 It is
a difficult manoeuvre because of the proximity of the
internal iliac vein, which can be torn during
mobilisation of the artery and is difficult to repair,
and the external iliac artery,which if ligated in error
results in an ischaemic leg.A practical point is that
when the artery is mobilised using an artery clamp,
this should be done laterally to medially, so that the
tip of the clamp points away from,rather than into,
the internal iliac vein. In the hands of experts who
perform the procedure regularly, the results can be
good.31 In the UK, it should probably not be
undertaken by the obstetrician who performs it, for
example, only once every 5 years, but instead the
assistance of a gynaecological oncologist or vascular
surgeon should be sought. If there is a delay in
obtaining assistance from such an expert, direct
compression of the aorta against the spinal column
can reduce bleeding by ~40% and this can be lifesaving
in some cases.Complete occlusion of the
aorta by clamping below the renal arteries is even
more effective and flow to the legs can be completely
stopped for 4 hours or more without irreversible
damage.However, analogous to the problem with
ligating the internal iliac artery, damage to the vena
cava can be catastrophic and so such clamping
should only be applied by an experienced vascular
surgeon.
Hysterectomy
In women wishing to retain their fertility,
caesarean hysterectomy is the procedure of last
resort; but, as has been repeatedly emphasised in
the Confidential Enquiries into Maternal and
Child Health, it should not be left until the woman
is in extremis, but instead should be carried out
promptly if the previously described procedures
prove to be ineffective and there are signs of
impending cardiovascular decompensation.
Anaesthetists will be the people most in touch with
the woman’s condition and if they declare that the
pulse rate is continuing to rise and the blood
pressure to fall despite conservative measures,
hysterectomy becomes inevitable. The precise
timing of this intervention must, of course, always
remain a matter of clinical judgment.
The topic of caesarean hysterectomy really requires
an article to itself, but the experience of this author
over the years suggests that it is often a good idea
to do subtotal hysterectomy first. This is often
sufficient to arrest the bleeding if the main cause is
an atonic corpus, because the two major pedicles
clamped, cut and tied include both the ovarian and
the uterine arteries.Even if there is continuing
bleeding, removing the body of the uterus improves
access to and visibility of the pelvic floor. It allows
identification of the cervix and therefore reduces the
chance of taking a pedicle too low and including the
ureter.Once the bleeding is controlled, any
temptation to remove more tissue, for example, the
cervix, should be resisted, as this may simply restart
the bleeding.Any specific bleeding sites should be
oversewn, even if it seems possible that the ureter
may be obstructed.This can always be rectified at a
later date, once the woman is no longer at risk of
death from haemorrhage.Even complete occlusion
of the ureter for several days will not result in
permanent damage to renal function,which will
resume once the obstruction is relieved. If bleeding
continues following hysterectomy, it becomes
mandatory to include surgeons with additional
experience of dealing with major haemorrhage,
such as a gynaecological oncologist or vascular
surgeon. In the meantime, pelvic tamponade with a
Logethotopulos pack32will usually staunch the flow
(Figure . The principle is straightforward.A
flexible plastic bag larger than the pelvic cavity is
filled with gauze swabs or anything similar to hand.
The neck is firmly tied to a length of tubing,which is
passed from the pelvis out through the vagina and
then attached to a litre bag of fluid which is allowed
to hang freely over the end of the bed. This applies a
steady tamponade which moulds itself to the pelvic
cavity and will stop all but the most major arterial
bleeding (especially as the woman is likely to be
quite hypotensive by this stage). I have had personal
communications from obstetricians who have
found this manoeuvre to be life-saving in extremis.
Special situations
Uterine inversion
This is a rare cause of PPH, but it is important to
recognise it promptly as the situation will not be
235
The Obstetrician & Gynaecologist 2009;11:231–238 Review
2009 Royal College of Obstetricians and Gynaecologists
Figure 7
Dawlatly suture for control of bleeding
from the lowersegment ofthe uterus.
Reproduced with permission from
Dawlatly et al.18
resolved until the inversion is corrected. If the
woman has had adequate analgesia, prompt
manual correction of the inversion is feasible and
will be effective in many cases. If the placenta is still
adherent to the uterus, it should be left in situ until
the uterus has been replaced. If there is a delay while
the woman is resuscitated and anaesthesia
provided, then hydrostatic replacement (the
O’Sullivan technique) may be necessary. Several
litres of warmed Hartmann’s solution instilled into
the vagina is usually enough to stretch the cervix
and generate enough pressure to push the uterus
back into a normal position.Traditionally, the
lower vagina was plugged with the accoucheur’s
hand, but a better seal can be obtained using a
silicone vacuum extractor (ventouse).33
More complicated methods have been described,
including applying manual upward pressure on the
cervix balanced by counter pressure on the uterus
via a laparoscopy probe34 and reducing the
inversion at laparotomy using a vacuum extractor
to suck out the fundus into its correct position.35
Placenta praevia and accreta
With the considerable rise in the rate of caesarean
section in recent years, the incidence of
placenta praevia and placenta accreta has risen
substantially.The risk of placenta praevia in a first
pregnancy is only about 1 in 400, but it rises to 1 in
160 after one caesarean section, 1 in 60 after two,
1 in 30 after three and 1 in 10 after four.36 If the
placenta is over the lower segment scar, then there
is an attendant risk that the placenta will invade
into (or occasionally through) the myometrium.
This risk is about 1 in 50 if there has been one
caesarean section, 1 in 6 after two, 1 in 4 after three,
1 in 3 after three or four and 1 in 2 after five.37 Thus,
the presence of a placenta praevia in a woman with
a previous caesarean section should always raise
the suspicion of a placenta accreta. This should be
investigated using ultrasound, supplemented if
possible with magnetic resonance imaging.
Ultrasound is probably the most sensitive method,
especially if a vaginal probe is used as well as an
abdominal probe, together with colour flow
(power) Doppler. This can reveal the presence
of large blood-filled spaces between the
fetus/amniotic fluid of the lower uterus and the
mother’s urinary bladder,with loss of the normal
myometrium. The presence of large blood vessels
with pulsatile flow in the bladder wall is a likely
indicator of placental invasion. In such cases,
operative delivery is necessary but is often
accompanied by profuse haemorrhage and
appropriate preparations must be made.The
likelihood of hysterectomy is significantly
increased to an odds ratio of 5.6 when there have
been five or more caesarean sections.38 Placenta
accreta is almost exclusively seen in association
with placenta praevia, it can sometimes develop as
pregnancy progresses and it never resolves with
advancing gestational age.
Practical aspects of preparation and care in the
operating theatre when placenta accreta is suspected
The average blood loss in cases of placenta accreta
is 3–51,39 so proper prior liaison with the
haematologist to ensure an appropriate supply of
crossmatched blood is essential. It is probably
advisable to have at least 4 units of packed red
blood cells in the operating theatre,with ready
access to further supplies, before commencing the
operation. It is also wise to arrange access to
supplies of clotting factors, including fresh frozen
plasma.Adequate intravenous access is important,
with two wide-bore venous lines inserted and an
arterial line to measure the blood pressure
accurately if there is major blood loss and
hypotension. In appropriate cases autologous
transfusion may be appropriate (e.g. some
Jehovah’s Witnesses will accept replacement of their
own blood, but will not accept it from other
people).Up to 1 unit per week can be removed for
storage during pregnancy without causing a
significant drop in haemoglobin concentration,
so up to 6 units can be collected in total: the bone
marrow can increase production of red cells to
compensate.Normovolaemic haemodilution
(taking off 250 ml of whole blood at a time and
replacing it with crystalloid) can also be used to
obtain a further 2 units immediately before
surgery.40 Cell savers can also be used to recycle
some of the woman’s own blood and they are now
routinely used in some units for this type of
surgery.41, 42
One needs to ensure adequate numbers of
experienced and well-trained supporting staff in
the operating theatre, plus appropriate equipment.
It is wise to have at least two suction devices with
bottles in reserve. The deleterious tissue perfusion
effects of blood loss are exacerbated by a drop in
body temperature, so the operating theatre should
be kept warm, as should the woman (using, for
example, a Bair Hugger warming blanket).
236
Review 2009;11:231–238 The Obstetrician & Gynaecologist
2009 Royal College of Obstetricians and Gynaecologists
Figure 8
Logethotopulos pack32
Instruments for bowel and bladder resection
should be available if needed, as should a vascular
surgery set. Preoperative cystoscopy and stent
placement is helpful to ascertain bladder
involvement and make any necessary bladder
surgery easier.Urological and vascular surgeons
should be available if needed. Packing of the vagina
with multiple gauze bandages to elevate the lower
uterine segment can make surgery easier if there is a
lot of bleeding and pelvic surgery becomes
necessary, as this elevates the pelvic floor and
facilitates identification of the cervix.43
For the delivery, general endotracheal anaesthesia is
preferred in combination with lumbar (thoracic)
epidural catheter placement preoperatively for
postoperative pain control. Intra-operative calf
compression (e.g.with Flowtron boots) helps to
guard against deep vein thrombosis if the operation
and recovery time is prolonged.
The most appropriate abdominal incision is a
midline, which gives the best access in case of heavy
bleeding (mass closure with a nylon suture gives the
lowest dehiscence rates). It is a good idea to scan
directly onto the uterus, using a sterile sleeve for the
transducer, to define the placental site precisely
before making the uterine incision. This incision
should be away from the placenta, often fundal, so
as to allow delivery of the baby before there is any
attempt at removing the placenta.We have
developed a technique in our unit of giving
oxytocics (such as an intravenous infusion of
oxytocin and 1000 micrograms of misoprostol
rectally) once the baby is safely delivered and then
waiting to see if the placenta separates. If it does,
and there is good uterine retraction with minimal
bleeding, then once the placenta is extruded by
uterine contraction the uterus can be closed. If the
placenta does not separate spontaneously within 10
minutes (this interval is arbitrary),we do not make
any attempt to separate it provided there is no
bleeding, but instead we close the uterus and wait
for the placenta to discharge spontaneously in the
puerperium (some authorities have suggested using
methotrexate to speed placental involution). If,
however, there is substantial bleeding, then we
proceed straight to hysterectomy without making it
worse by trying to remove the placenta piecemeal.
Persistent bleeding can often be arrested by arterial
embolisation but this technique is outwith the
scope of this article. The article by Boulleret et al.44
is recommended to readers as a good account of
recent techniques.
Conclusion
For most of the last century, the management of
major PPH relied upon the use of oxytocic agents,
followed by hysterectomy if these failed.However,
the last 10 years has seen the introduction of many
useful additional surgical procedures, in particular
uterine compression sutures and intrauterine
tamponade balloons. These are now widely used
and are effective in ~90% of cases.However, reports
of both short- and long-term complications are
now appearing and it is important not to reduce the
perfusion of the uterus so much that it becomes
devitalised.Uterine artery ligation can be carried
out safely by an obstetrician, but internal iliac
artery ligation should be carried out only by a
surgeon familiar with this procedure, for example,
a gynaecological oncologist or vascular surgeon.
Hysterectomy still has an important place. If
bleeding continues after the uterus has been
removed, the Logethotopulos pack can be used to
stabilise the situation and arterial embolisation can
be life-saving.With the increasing incidence of
caesarean section, the possibility of placenta accreta
should always be considered in the next pregnancy
and ultrasound/magnetic resonance imaging are
important.Anticipation and careful preparation of
the operating theatre, facilities, blood products and
surgeons remain the key to successful management.
Acknowledgement
The author is grateful for the assistance of Professor
Michael Belfort of the Utah Valley Regional
Medical Center (USA) for helping with the
development of the lecture upon which this article
is based.
References
1 Joseph KS, Rouleau J, Kramer MS, Young DC, Liston RM, BaskettTF.
Investigation of an increase in postpartum haemorrhage in Canada.
BJOG 2007;114:751–9. doi:10.1111/j.1471-0528.2007.01316.x
2 Franchini M, Lippi G, Franchi M. The use of recombinant activated factor
VII in obstetric and gynaecological haemorrhage. BJOG 2007;114:8–15.
doi:10.1111/j.1471-0528.2006.01156.x
3 Hensleigh PA. Anti-shock garment provides resuscitation and
haemostasis for obstetric haemorrhage. BJOG 2002;109:1377–84.
doi:10.1046/j.1471-0528.2002.02197.x
4 Brees C, Hensleigh PA, Miller S, Pelligra R. A non-inflatable anti-shock
garment for obstetric hemorrhage. Int J Gynaecol Obstet
2004;87:119–24. doi:10.1016/j.ijgo.2004.07.014
5 Miller S, Hamza S, Bray EH, Lester F, Nada K, Gibson R, et al. First aid
for obstetric haemorrhage: the pilot study of the non-pneumatic
anti-shock garment in Egypt. BJOG 2006;113:424–9.
doi:10.1111/j.1471-0528.2006.00873.x
6 Hsu S, Rodgers B, Lele A, Yeh J. Use of packing in obstetric hemorrhage
of uterine origin. J Reprod Med 2003;48:69–71.
7 Johanson R, Kumar M, Obhrai M, Young P. Management of massive
postpartum haemorrhage: use of a hydrostatic balloon catheter to avoid
laparotomy. BJOG 2001;108:420–2. doi:10.1111/j.1471-0528.2001.00102.x
8 Bakri YN, Amri A, Abdul JF. Tamponade-balloon for obstetrical bleeding. Int J
Gynaecol Obstet 2001;74:139–42. doi:10.1016/S0020-7292(01)00395-2
9 Dabelea V, Schultze PM, McDuffie RS, Jr. Intrauterine balloon tamponade
in the management of postpartum hemorrhage. Am J Perinatol
2007;24:359–64. doi:10.1055/s-2007-984402
10 Tattersall M,.Braithwaite W. Balloon tamponade for vaginal lacerations
causing severe postpartum haemorrhage. BJOG 2007;114:64–8.
doi:10.1111/j.1471-0528.2007.01278.x
11 Lynch C, Coker A, Lawal AH, Abu J, Cowen MJ. The B-Lynch surgical
technique for the control of massive postpartum haemorrhage: an
alternative to hysterectomy? Five cases reported. Br J Obstet Gynaecol
1997;104:372–5.
12 Cho JH, Jun HS, Lee CN. Hemostatic suturing technique for uterine
bleeding during cesarean delivery. Obstet Gynecol 2000;96:129–31.
doi:10.1016/S0029-7844(00)00852-8
13 Hackethal A, Brueggmann D, Oehmke F, Tinneberg HR, Zygmunt MT,
Muenstedt K. Uterine compression U-sutures in primary postpartum
hemorrhage after Cesarean section: fertility preservation with a simple
and effective technique. Hum Reprod 2008;23:74–9.
doi:10.1093/humrep/dem364
14 Hayman RG, Arulkumaran S, Steer PJ. Uterine compression sutures:
surgical management of postpartum hemorrhage. Obstet Gynecol
2002;99:502–6. doi:10.1016/S0029-7844(01)01643-X