Intensive Care Med (2011) 37:18161825

DOI 10.1007/s00134-011-2315-0

Etienne Gayat
Matthieu Resche-Rigon
Olivier Morel
Matthias Rossignol
Jean Mantz
Armelle Nicolas-Robin
Nathalie Nathan-Denizot
Jean-Yves Lefrant
Frederic J. Mercier
Emmanuel Samain
Yann Fargeaudou
Emmanuel Barranger
Marie-Jose`phe Laisne
Pierre-Henri Brechat
Dominique Luton
Ingrid Ouanounou
Patricia Appa Plaza
Claire Broche
Didier Payen
Alexandre Mebazaa

ORIGINAL

Predictive factors of advanced interventional

procedures in a multicentre severe postpartum
haemorrhage study

E. Gayat
M. Resche-Rigon
J. Mantz

E. Barranger
D. Luton
D. Payen

A. Mebazaa
University Paris Diderot, Sorbonne Paris
Cite, 75205 Paris, France

E. Samain
Pole danesthesie-reanimation chirurgicale,
Center hospitalier universitaire JeanMinjoz, Universite de Franche-Comte,
Besancon, France

Patent The SPPH score has been patented

by the Assistance Publique-Hopitaux de
Paris and the University Paris 7 Diderot;
A.M., E.G. and M.R.-R. were the main coinventors (Patent No. WO2010/015690 A1).

O. Morel
E. Barranger
Department of Obstetrics and Gynaecology,
Lariboisie`re University Hospital, AP-HP,
Paris Cedex 10, France

Y. Fargeaudou
Department of Radiology, Lariboisie`re
University Hospital, AP-HP,
Paris Cedex 10, France

Electronic supplementary material

The online version of this article
(doi:10.1007/s00134-011-2315-0) contains
supplementary material, which is available
to authorized users.

J. Mantz
Department of Anesthesiology and Critical
Care Medicine, Mobile Care Unit,
Beaujon University Hospital, AP-HP,
Clichy-La Garenne, France

P.-H. Brechat
Research Laboratory of Social and Health
Policies (LAPSS), Ecole des Hautes Etudes
en Sante Publique (EHESP),
Rennes Cedex, France

A. Nicolas-Robin
Department of Anesthesia and Intensive
Care, Mobile Care Unit, Pitie-Salpetrie`re
University Hospital, AP-HP, University
Paris 6, Paris Cedex 13, France

D. Luton
Etablissement Francais du Sang,
Center de transfusion, Lariboisie`re
University Hospital, Assistance Publique,
Hopitaux de Paris, Paris Cedex 10, France

Received: 28 January 2011

Accepted: 5 May 2011
Published online: 30 July 2011
Copyright jointly held by Springer and
ESICM 2011

N. Nathan-Denizot
A. Mebazaa ())
Department of Anesthesia and Intensive
Department of Anesthesiology and Critical
Care, Hopital Me`re-enfant, Limoges, France Care Medicine, Sorbonne Paris Cite,
Lariboisie`re Hospital, University Paris
J.-Y. Lefrant
Diderot, INSERM UMR 942, Universite
Division Anesthesie Reanimation Douleur
Paris 7, 2 rue Ambroise Pare, 75010 Paris
Urgences, Groupe Hospitalo-Universitaire
Cedex 10, France
Caremeau, CHU Nmes, Nmes, France
e-mail: alexandre.mebazaa@lrb.aphp.fr
E. Gayat
M. Resche-Rigon
Tel.: ?33-1-49958071
F. J. Mercier
Department of Biostatistics and Clinical
Fax: ?33-1-49958083
Department of Anesthesiology and Critical Care
Epidemiology, Saint-Louis University
Hospital, AP-HP, INSERM UMR 717, Paris Medicine, Antoine Becle`re hospital, AP-HP,
Cedex 10, France
University Paris Sud 11, Clamart, France
E. Gayat
M. Rossignol
M.-J. Laisne

I. Ouanounou
P. A. Plaza
C. Broche

D. Payen
A. Mebazaa
Department of Anesthesiology and Critical
Care Medicine, Mobile Care Unit,
Lariboisie`re University Hospital, AP-HP,
EA322, Paris Cedex 10, France

1817

Abstract Purpose: Severe postpartum haemorrhage (SPPH) is the

leading cause of peripartum hysterectomy and maternal death. There are
no easily measurable parameters that
indicate the failure of medical therapy
and the need for an advanced interventional procedure (AIP) to stop
genital tract bleeding. The aim of the
study was to define factors predictive
of the need for an AIP in the management of emergent PPH.
Methods: The study included two
phases: (1) an initial retrospective
study of 257 consecutive patients
with SPPH, allowing the determination of independent predictors of AIP,
which were subsequently grouped in

a predictive score, followed by (2) a

multicentre study of 239 patients
admitted during 2007, designed to
validate the score. The main outcome
measure was the need for an AIP,
defined as uterine artery embolization, intraabdominal packing, arterial
ligation or hysterectomy.
Results: Abnormalities of placental
implantation, prothrombin time
\50% (or an International Normalized Ratio [1.64), fibrinogen \2 g/l,
troponin detectable, and heart rate
[115 bpm were independently predictive of the need for an AIP. The
SPPH score included each of the five
predictive factors with a value of 0 or
1. The greater the SPPH score, the

Introduction
Postpartum haemorrhage (PPH) is the leading cause of
maternal death in many developing and high-income countries and is also the leading cause of peripartum hysterectomy,
which is known to cause many physical and psychological
sequelae in young women of fertile age [15]. Severe PPH
(SPPH), defined as a blood loss exceeding 1 l [6], is often
unrecognized because not only is it often inaccurately estimated during delivery [7], but it is also sometimes not even
externalized. Furthermore, persistence of bleeding can also go
unrecognized as the initial bleeding leads to volume resuscitation of the patient including eventual blood transfusions so
that parameters such as haemoglobin level do not reflect the
severity of the bleeding [8].
International societies recommend the early introduction of a uterotonic agent as first-line therapy in PPH,
simultaneously with massaging of the uterus, bladder
catheterization and bimanual uterine compression [6, 9
12]. Although in the case of failure of initial medical
management, advanced interventional procedures (AIP),
i.e. uterine artery embolization and/or haemostatic surgery,
are recommended, unfortunately there is no unequivocal
evidence as to when, or based on what parameters, a more
invasive therapy should be initiated [13]. This lack of
accepted (and easily measurable bed-side) parameters
indicating the failure of conservative therapy and the need
for an AIP, has undoubtedly led in many cases to delays in
initiation of aggressive therapy and/or transfer to a specialized tertiary care centre, and it has been proven that
delay in initial care is an independent predictor of severe
blood loss in women with PPH [14].
Accordingly, the aim of our study was to define the
factors that indicate failure of the initial medical and/or
surgical management in resuscitated patients and also to

greater the percentage of patients

needing an AIP (11% for SPPH 0, to
75% for SPPH C2). The AUC of the
ROC curve of the SPPH score was
0.80. Conclusions: We identified
five independent predictors of the
need for an AIP in patients with
SPPH and persistent bleeding. Using
these predictors in a single score
could be a reliable screening tool in
patients at risk of persistent genital
tract bleeding and needing an AIP.
Keywords Post-partum hemorrhage

Hemorrhagic shock
Predictive score

predict the need for an AIP to halt blood loss via the
genital tract in parturient patients with PPH.

Materials and methods

The present work consisted of two phases. In an initial
retrospective study data were collected from 257 consecutive parturient patients admitted between 1 January
2004 and 31 December 2005 to Lariboisie`re Hospital
(Paris, France) for PPH. This resulted in the identification
of independent predictors of the need for an AIP. The
identified predictors were subsequently grouped to create
a predictive score, called the severe post-partum hemorrhage (SPPH) score, to predict the need for an AIP in the
management of parturient patients with PPH. This phase
was followed by a second multicentre study designed to
validate the SPPH score. This study included 239 parturient patients with PPH admitted during 2007 to seven
French referral centres. Data collection was approved by
the Institutional Review Board (Comite de Protection des
Personnes Ile-de-France IV and CEERB GHU Nord) that
waived the need for patient signed consent.
Initial study to identify independent predictors
of the need for an AIP
Details of PPH management in Lariboisie`re tertiary centre
are given in the Electronic supplementary material 1.
The 20042005 initial cohort of Lariboisie`re Hospital
Patients with PPH were identified on the computerized
system of Lariboisie`re Hospital on which any parturient

1818

patient admitted is registered with a main diagnosis. PPH

was coded as postpartum complication and haemorrhagic shock or acute anaemia or shock. These data
were cross-checked against the registries of the Department of Obstetrics, the Department of Anaesthesia and
Intensive Care and the Department of Radiology.
According to the data, 257 parturient patients with PPH
were managed in Lariboisie`re Hospital centre in 2004 and
2005. Patients were mostly transferred after delivery
(n = 227) from 82 primary care centres (academic and
non-academic, public and private) located within or
around Paris (Ile-de-France region) while a minority
delivered at Lariboisie`re Hospitals obstetric unit
(n = 30).
The 257 patients were divided into an advanced
interventional procedure (AIP) group (n = 110) and a
medically managed (MM) group (n = 147). The latter
group included those with no need of any of the abovementioned AIPs together with a few patients requiring
only instrumental genital tract examination.
Data used in the predictive model were recorded either
before admission (i.e. obstetric parameters and management details before transfer) or within 15 min of
presentation (i.e. haemodynamic and biological values).
Subsequently, independent factors necessitating AIP to
stop genital tract bleeding in patients with PPH of the
initial cohort were identified and included in the SPPH
score.
Multicentre validation study
The SPPH score was validated in the multicentre validation cohort, collected throughout 2007 by physicians
unaware of the SPPH score. This cohort included 239
patients: 150 hospitalized in Lariboisie`re Hospital and 89
in six other PPH referral centres located in various regions
of France (see Appendix).

Analyses were performed using the R statistical package

(available online at http://www.R-project.org). The statistical analysis is shown in detail in the Electronic
supplementary material 2.

Results
The 20042005 retrospective cohort
Patient characteristics
During 2004 and 2005 257 parturient patients with a
median age of 31 years (first to third quartile
2835 years), were consecutively admitted for PPH to
Lariboisie`re Hospital. Before admission to Lariboisie`re
Hospital, attempts were made to control genital tract
bleeding by continuous infusion of sulprostone and surgical repair (hysterectomy in 12 patients, arterial ligation
in 11 patients). Details of the patients peripartum management are shown in Table 1.
On admission, patients with PPH demonstrated the
following signs of severity (Table 2): median haemoglobin level 9.2 g dl/l and haematocrit 27% despite a median
transfusion of 2 U of red blood cells (RBC) and 1 U of
fresh frozen plasma prior to admission. Other signs of
haemorrhage included tachycardia (median heart rate
105 bpm), coagulopathy (median prothrombin time, PT,
67%) and thrombocytopenia (median platelets
118,000 mm-3).
The leading cause of PPH was uterine atony (69%)
followed by genital tract lacerations and abnormalities of
placental implantation, including placenta praevia, placenta accreta and placenta percreta. Of note, all placenta
accreta/percreta were unexpected and diagnosed at the
time of delivery.
Use of advanced interventional procedures

Statistical analysis
The results are expressed as means and standard deviation
(sd), median and first to third quartile or counts and percent. The study outcome was defined as parturient patients
undergoing at least one AIP defined as uterine artery
embolization, intraabdominal packing, arterial ligation or
hysterectomy. Multiple logistic regression was used to
determine a set of variables independently associated with
each outcome. The multiple model was finally used to
define a simple clinical prognostic score based on the
linear predictor obtained with a unit coefficient associated
with each of the five final selected variables. To assess its
external validity, the score was evaluated in the multicentre validation cohort using a logistic regression model.
All tests were two-sided at the 0.05 significance level.

After the initial evaluation, 110 patients underwent an

AIP due to suspected persistent active bleeding, and thus
were considered as the AIP group. Of note, the AIP group
included 14 patients in whom persistent active bleeding
was suspected on their arrival in Lariboisie`re Hospital,
despite surgical procedures (arterial ligation or hysterectomy) in primary institutions.
AIPs performed in Lariboisie`re Hospital included only
uterine artery embolization (n = 85), only open surgery
(n = 14), and combined embolization and surgery
(n = 11). Open surgery (n = 25) included peritoneal
packing, arterial ligation, hysterectomy, or a combination
of all three (Table 3).
Uterine artery embolization was performed a median
of 1.9 h [first to third quartile 1.22.5 h] after admission
and surgical procedures 2.1 h [1.44.6 h] after admission.

1819

Table 1 Patients peripartum management before admission in the initial cohort. The results are expressed as median [first to third
quartile] (minimummaximum range) or count (%)

Age (years)
Obstetric parameters
First delivery
First pregnancy
Previous postpartum haemorrhage
Uterine fibroma
Uterine scar
Pre-eclampsia
Term of pregnancy (weeks)
Twin pregnancy
Delivery in a university hospital
Details of delivery
Labour induced
Caesarean section
Labour duration (h)
Instrumental manoeuvre
Manual removal of placenta
Intrauterine examination
Obstetric anaesthesia
General
Epidural
Spinal
None
Cause of haemorrhage
Primary uterine atonia
Genital tract laceration
Abnormalities of placental implantation
Uterine rupture
Surgery before transfer
Instrumental genital tract examination
Hysterectomy
Intraabdominal packing
Arterial ligation
Medical management before admission
Sulprostone infusion
Catecholamine
Mechanical ventilation
Red blood cells (units)
Fresh frozen plasma (units)
Platelet concentrate unit (units)
Time between delivery and admission (h)
Transport duration (h)b
a
b

MM group (n = 147)

AIP group (n = 110)

p valuea

31 [2735] (18; 44)

32 [3036] (16; 44)

0.02

72 (49)
57 (39)
6 (4)
9 (6)
14 (8)
18 (12)
39 [3840] (22; 41)
7 (6)
25 (15)

45 (41)
32 (29)
2 (2)
3 (3)
16 (16)
12 (12)
39 [3740] (22; 42)
8 (7)
32 (32)

NS
NS
NS
NS
NS
NS
0.04
NS
0.002

37 (25)
37 (25)
6 [48]
33 (22)
71 (48)
106 (72)

12 (13)
45 (35)
4 [26]
18 (19)
53 (53)
68 (68)

0.03
NS
0.02
NS
NS
NS
NS

13
98
18
20

14
64
12
11

(9)
(66)
(12)
(14)

(14)
(62)
(13)
(11)
0.003

109 (74)
34 (23)
4 (3)
0 (0)

69 (61)
22 (20)
16 (14)
3 (3)

61 (41)
4 (3)
0 (0)
7 (5)

29 (30)
8 (8)
5 (5)
4 (4)

NS
NS
0.01
NS

131 (89)
0 (0)
13 (9)
1.2 2.0 (0; 13)
0.6 1.7 (0; 10)
0.1 0.4 (0; 4)
5 [47] (0.96; 31.2)
0.5 [0.20.7] (0.24; 2.2)

86 (85)
5 (5)
22 (20)
2.8 4.1 (0; 23)
1.6 3.1 (0; 17)
0.4 1.9 (0; 10)
4 [37] (0.24; 28.6)
0.5 [0.20.7] (0.24; 2.4)

NS
0.01
0.03
0.0004
0.003
NS
NS
NS

Comparison between MM group and AIP group

Including only patients transferred from a primary centre to Lariboisie`re hospital (n = 227)

Two patients died in the AIP group, one from amniotic

fluid embolism and the other from refractory haemorrhagic shock. The median length of stay in the AIP group
(n = 110) was 3.2 days [2.36.2 days] in Lariboisie`re
Hospital and bleeding was stopped in all patients with
none was readmitted for PPH.
In the remaining 147 patients forming the MM group,
bleeding was considered as non-active both on admission
and during follow-up. No death occurred in this group.
Patients were hospitalized mostly in the high dependency
unit for 0.9 days [0.72.1 days] and transferred to the
obstetric ward. None was readmitted for PPH.

Factors predictive of the need for an advanced

interventional procedure
Univariate analysis. Although few differences in the
demographic and obstetric parameters collected prior to
admission were found between the AIP and MM groups,
differences were observed in the haemodynamic and
biological results (Tables 1 and 2). Patients who required
an AIP were more unstable haemodynamically: lower
systolic blood pressure (SBP) and diastolic blood pressure, higher heart rate, lower haemoglobin, higher plasma
troponin I, metabolic acidosis despite more RBC

1820

Table 2 Haemodynamic and biological status on admission in the initial cohort. The results are expressed as median [first to third
quartile] or count (%)

Haemodynamic status
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Heart rate (bpm)
Biological valuesb
pH (IU)
PaCO2 (mmHg)
PaO2 (mmHg)
Bicarbonate (mmol/l)
Proteins (g/l)
Creatinine (lmol/l)
Lactate (mmol/l)
Aspartate aminotransferase (IU/ml)
Alanine aminotransferase (IU/ml)
Troponin I (ng/ml)
Bilirubin (mmol/l)
Haemoglobin (g/dl)
Haematocrit (%)
Platelet (103/mm3)
Prothrombin time (%)
Fibrinogen (g/l)
Factor V (%)
Outcome
ICU stay
Length of stay in our unit (days)
Deaths
a
b

MM group (n = 147)

AIP group (n = 110)

p valuea

110 [100120]
60 [5065]
100 [90115]

100 [87115]
50 [4560]
115 [100130]

0.0007
0.0008
\0.0001

7.43 [7.47.45]
32 [2935]
138 [104187]
23 [2224]
45 [4051]
58 [5167]
1.91 [1.392.5]
23 [1930]
12 [1016]
0 [00.03]
7 [411]
9.5 [8.210.6]
27 [2431]
131 [98161]
73 [6485]
2.65 [2.083.46]
52 [3966]

7.39 [7.327.43]
33 [29.2538]
195 [144247]
22 [2023]
38 [32.543]
59 [4971]
2.50 [1.923.73]
21 [1631]
13 [1018]
0.06 [00.4]
8 [515]
8.7 [7.09.9]
25 [2030]
93 [64133]
58 [3873]
1.8 [1.092.52]
41 [2356]

\0.0001
NS
\0.0001
\0.0001
\0.0001
NS
\0.0001
NS
NS
\0.0001
NS
0.001
0.0006
\0.0001
\0.0001
\0.0001
0.0006

6 (4)
1.0 [0.72.1]
0 (0)

31 (28)
3.2 [2.36.2]
2 (2)

\0.0001
\0.0001
NS

Comparison between MM group and AIP group

Except blood gas, other biological values were measured in plasma

Table 3 Univariate and multivariate estimated odds ratios for each of the five items included in the SPPH score. The results are expressed
as median [first to third quartile]
Data collected on admission

Abnormalities of placental implantation

Prothrombin time \50%a
Heart rate [115 bpm
Fibrinogen \2 g/l
Troponin I detectable
a

Univariate analysis

Multivariate analysis

Odds ratio [95% CI]

p value

Odds ratio [95% CI]

p value

4.83
8.42
2.73
5.02
3.91

0.003
\0.0001
\0.0001
\0.0001
\0.0001

7.05
3.55
2.18
2.75
2.73

0.0007
0.008
0.04
0.005
0.0009

[1.7113.64]
[3.8618.36]
[1.514.95]
[2.98.67]
[2.326.59]

[2.2622.03]
[1.389.17]
[1.034.62]
[1.514.95]
[1.514.95]

Corresponding to an average threshold INR of 1.64 in the seven hospitals

transfusion and a higher rate of catecholamine administration. Coagulation disorders were more pronounced in
the patients who required an AIP: lower platelet counts,
PT, factor V and fibrinogen despite being given more
fresh frozen plasma than the MM patients.
Score predicting the need for an advanced interventional
procedure: the SPPH score. The multivariate analysis
and selection process led to the identification of five independent factors predicting the need for an AIP, all present
on admission, in the 20042005 cohort: abnormalities of
placental implantation, PT \50% (corresponding to an
average International Normalized Ratio, INR, threshold of
1.64, in the seven hospitals), fibrinogen \2 g/l, troponin

detectable and heart rate[115 bpm; uni- and multivariate

odds ratios of the five independent factors are presented in
Table 4. Interestingly, other factors related to bleeding,
such as blood pressure, haemoglobin rate, number of RBCs
transfused and the use catecholamine, were not predictive
of the need for an AIP.
The SPPH score was established and included each of
the five predictive factors with a value of 0 or 1 when
absent or present on admission, respectively, with the
total ranging from 0 to 5. The greater the SPPH score the
greater the percentage of patients requiring an AIP: 11%
for SPPH 0, 39% for SPPH 1 and 75% for SPPH C2 in the
20042005 cohort (Fig. 1b). In addition, the AUC of the
ROC curve of the SPPH score was 0.80 in this cohort.

1821

Table 4 Comparison of patient characteristics between the initial cohort and the multicentre validation cohort. Results are expressed as
median [first to third quartile] or count (%)

Age (years)
Obstetric parameters
First delivery
First pregnancy
Details of delivery
Caesarean section
Instrumental manoeuvre
Cause of haemorrhage
Primary uterine atonia
Genital tract laceration
Abnormalities of placental implantation
Uterine rupture
Surgery before transfer
Hysterectomy
Arterial ligation
Medical management before admission
Sulprostone infusion
Catecholamine
Haemodynamic and biological values on admission
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Heart rate (bpm)
Troponin I (ng/ml)
Haemoglobin (g/dl)
Platelets (103/mm3)
Prothrombin time (%)
Fibrinogen (g/l)a
Need for AIP [n (%)]

Multicentre validation
cohort (n = 237)

31 [2935]

32 [2936]

0.09

109 (46)
85 (35)

106 (46)
88 (38)

0.93
0.50

82 (32)
51 (21)

101 (42)
35 (15)

0.02
0.12

178 (69)
56 (22)
20 (8)
3 (1)

171 (72)
39 (16)
21 (9)
6 (3)

0.49
0.14
0.75
0.32

12 (5)
11 (4)

9 (4)
11 (5)

0.66
1

217 (87)
5 (2)

219 (93)
14 (6)

0.045
0.035

110 [95120]
55 [5060]
105 [90120]
0.0 [0.00.2]
9.2 [7.910.3]
118 [80154]
69 [5579]
2.3 [1.63.1]
110 (43)

120 [105135]
70 [6080]
98 [80116]
0.0 [0.00.1]
9.0 [7.510.6]
131 [89167]
67 [5678]
2.7 [1.83.3]
100 (42)

p value

\0.001
\0.001
\0.001
0.01
0.68
\0.001
0.98
0.04
0.93

Measured using the von Clauss assay (chronometric method) in all included centres

Sever Post Partum Hemorrhage (SPPH) Score

Abnormalities of Placental implantation

0 or 1

PT at admission < 50%

0 or 1

HR at admission > 115 bpm

0 or 1

Fibrinogene at admission < 2 gL-1

0 or 1

Troponine I detectable at admission

0 or 1
Total score = 0 to 5

100

Se (%)

60
40

% AIP

80

2004-2005 initial cohort

Multicentric validation cohort

20

Fig. 1 SPPH score. a The five

items included in the score.
b Left, percent of patients
requiring an AIP in relation to
SPPH score; right, ROC curve
of SPPH scores in the 2007
cohort. These analyses led to
the classification of patients
with PPH into three groups:
SPPH score 0, 1 or C2 with
low, intermediate or high risk of
needing an AIP (see our
proposal for SPPH use in the
management algorithm of
PPHFig. 2) (AIP advanced
interventional procedures, PT
prothrombin time, HR heart
rate, ROC receiver operating
curve, AUC area under curve)

Multicentric validation cohort (AUC = 0.83

Lariboisire centre (AUC = 0.83 0.03)
Other centres (AUC = 0.82 0.04)

Initial cohort
(n = 257)

Value of the SPPH score

1 -Spe (%)

0.03)

1822

The multicentre validation cohort

The multicentre validation cohort included 239 patients
admitted to seven French referral centres including Lariboisie`re Hospital during the year 2007. The
characteristics of the validation cohort are presented in
Table 3. The percentage of patients requiring an AIP for
each value of SPPH score and the global performance of
the SPPH score were similar in the multicentre validation
cohort and the initial cohort (Fig. 1b). Furthermore,
AUCs of the SPPH score determined prospectively in the
global cohort, patients from Lariboisie`re centre and
patients from the other centres were 0.83, 0.83 and 0.82,
respectively (Fig. 1b). In addition, no interaction was
found between SPPH score and centre (p = 0.93). Of
note, the SPPH score predicted the risk of needing an AIP
regardless of the severity of the haemodynamics on
admission: the AUC was 0.83 when the SBP was
\120 mmHg and 0.78 when SBP was C120 mmHg.
In line with these results, sensitivity, specificity, and
positive and negative predictive values were 0.91, 0.58,
0.62 and 0.90 for SPPH scores C1, and 0.62, 0.85, 0.76
and 0.76 for SPPH scores C2 in the multicentre validation
cohort.

suggests that the SPPH score might be used as an early

warning system, alerting physicians to the need for an
AIP upon the appearance of additional criteria.
The four other parameters (PT or INR, heart rate,
fibrinogen and troponin) of the SPPH score have already
been reported to be altered in patients with SPPH [8, 23].
Conversely, many other factors usually present in patients
with severe bleeding, such as low blood pressure, haemoglobin concentration, volume of blood transfusion or
catecholamine use, were not found to be independent
predictors of the need for an AIP in our study. Moreover,
thrombocytopenia, previously reported to be related to the
severity of bleeding [24] and also to an increased likelihood of bleeding in the critically ill [25], was not found to
be an independent predictor of the need for an AIP. These
results are very likely related to the fact that the studied
patients had been resuscitated including medical and
surgical treatments before transfer. Furthermore, the
SPPH score performed similarly in patients with different
levels of SBP on admission. Accordingly, PT (or INR),
heart rate, fibrinogen and troponin are the factors most
predictive of persistent bleeding among all measurable
haemodynamic and biological factors in patients with
PPH and possibly, more broadly, in those with haemorrhagic shock of other causes who are resuscitated before
hospital admission. This requires further evaluation.

Discussion
Limitations of the study
PPH remains the leading cause of death in the peripartum
period in most countries, including France [2, 15]. Death
seems mostly associated with a delay in initiating
appropriate management for haemorrhagic shock (transfusion) and/or delay in appropriate manoeuvres to stop
genital tract bleeding [1621]. Although uterotonic agents
have been used as first-line therapy when bleeding persists [9, 22], no study to date has evaluated which
parameter(s) can indicate the need for more aggressive
management (embolization and/or surgery) when optimal
medical treatment has failed. Our study led to the identification of five independent predictors of the need for
AIP in a large cohort of patients with PPH. We also
describe the SPPH score, which includes five items with a
value of 0 or 1 with a total ranging from 0 to 5. When the
score was C2, there was greater than 70% risk of the need
for an AIP to stop genital tract bleeding.
The presence of abnormalities of placental implantation was strongly associated with an increased risk of the
need for an AIP in our patients with PPH. This does not
mean that an abnormality of placental implantation is, per
se, a risk factor for PPH, but rather that an abnormality of
placental implantation increases sevenfold the risk of
failure of either medical or surgical management
(including sulprostone administration) in patients with
PPH. Our study emphasizes the need to closely monitor
patients with PPH with a placental abnormality, and

An AIP was needed in nearly 40% of patients with PPH

admitted to Lariboisie`re Hospital in the 20042005 initial
cohort. As arterial embolization was available 24 h a day
throughout the year, one could argue that Lariboisie`re
Hospital performed more arterial embolizations than
necessary. However, Table 2 shows that patients benefiting from interventional procedures (mostly uterine
artery embolization) were in a much more unstable haemodynamic condition and with biological signs of active
bleeding, despite a larger number of transfusions and a
higher rate of catecholamine administration, than patients
in the MM group. They therefore needed an AIP to stop
the bleeding. Furthermore, it is interesting to note that the
rate of AIP was similarly near 40% in the multicentre
validation cohort.
Some predictors of AIP could have been missed. First,
some might not have been found to be significant because
of the number of participants. Second, others might not
have been measured in the present study. For instance, the
haemoglobin level before delivery was not available
although it is known to be related to maternal outcome.
Furthermore, parameters included in the SPPH score were
recorded on admission when the patients had already been
resuscitated in their primary centre. Whether this score is
usable at an earlier stage of PPH management remains to
be determined.

1823

Our study used PT \50% as one of the items in the

SPPH score, as PT is the biological marker of altered
coagulopathy used in daily practice, in the context of
bleeding. PT expressed as plain percentage activity might
have less favourable intercentre variability than INR.
However, in our study, the SPPH score used PT as a
dichotomized parameter (0 for PT C50%, 1 for PT\50%)
that would have a better intercentre variability than a
continuous parameter. The INR values corresponding to
the cut-off of PT at 50% was 1.64 with a small standard
deviation (0.07) among the seven centres. Thus the use of
INR would be preferred in future multicentre studies on
coagulopathy and bleeding.
The performance of the SPPH score might be different
among countries as it might be influenced by various
parameters including local health-care resources. The
SPPH score needs to be externally validated in a multinational cohort from countries with different maternal
mortality ratios due to haemorrhage.
Clinical implications
The SPPH score could be an accurate tool to predict the
failure of initial medical and/or surgical management in
patients with SPPH and the need for an AIP to stop
genital tract bleeding, for at least the following three
reasons: (1) the studied patients reflected the diversity of
patients with PPH in France (patients delivered in academic or non-academic, public or private primary
institutions and managed in seven PPH referral centres
located in four different regions in France), (2) the AUC
of the SPPH score was consistently [0.80 in both the
Fig. 2 Proposed management
algorithm including the SPPH
score. (BP blood pressure, HR
heart rate, Hb haemoglobin
level, AIP advanced
interventional procedures,
SPPH SPPH score). Our study
indicates that the SPPH score
should be used while the patient
is being resuscitated. SPPH
scores of 0, 1 and C2
correspond to low (about 10%),
intermediate (about 40%) and
high (75%) risks of needing an
AIP. SPPH scoring could lead
to early detection of persistent
genital tract bleeding and thus
to either immediate initiation of
an interventional procedure or
immediate transfer to a tertiary
centre. Of note, the incidences
indicated on the figure are not
the actual incidence in the
studied centres, but are
estimated numbers based on
literature data [2830]

initial Lariboisie`re and the multicentre validation cohorts,

and (3) the high sensitivity ([0.90) of an SPPH score of
C1. For these reasons the use of the SPPH score can be
strongly recommended as a screening tool for persistent
haemorrhage.
Figure 2 illustrates the possible place of the SPPH
score in the management of SPPH. SPPH is initially
diagnosed by genital tract bleeding [1,000 ml associated
with haemodynamic alterations (change in blood pressure,
heart rate and/or haemoglobin level). Patients should
therefore be resuscitated immediately by volume and
blood resuscitation, along with intravenous uterotonic
administration and a recheck of the genital tract if
required. We propose that SPPH scoring be performed in
parallel with the resuscitation (SPPH score 0 corresponds
to a low risk, 1 to an intermediate risk, and C2 to a high
risk of needing an AIP). SPPH scoring could lead to early
detection of persistent genital tract bleeding and thus to
rapid and adequate management.
The development of point-of-care solutions [26, 27] to
assess the three biological parameters included in the
score could facilitate the spread of the use of the SPPH
score.

Conclusion
In summary, abnormalities of placental implantation,
heart rate, fibrinogen, PT (or INR) and troponin are all
independent predictors of failure of the initial medical
and/or surgical management of SPPH. The SPPH score,
including all five parameters, seems to have good

Risk
for AIP
SPPHs 2
SPPHs= 1
SPPHs= 0

High
Intermediate
Low

3/1000

Perform SPPHs to assess

persistance of bleeding
Severe
Post-partum
SeverePost
-partum haemorrhage
Alteration in BP, HR
and / or Hb
Blood loss > 1000 ml

Volume / blood rescucitation

10/1000

Sulprostone
+ Re-Check genital tract

Post-partum haemorrhage
Post-partum

Blood loss 500 -1000 ml

How do diagnose
severityof of bleeding?

Oxytocin
+ Uterin massage
+ Check genital tract

How to stop bleeding?

50/1000

Incidence

1824

performance and could be a reliable screening tool in

identifying patients with PPH at risk of persistent genital
tract bleeding and who will need an interventional haemostatic procedure. Future studies should assess whether
the SPPH score can be used as a monitoring tool in any
labour ward. Undeniably, SPPH monitoring may allow
the early detection of persistent bleeding, resulting in
either immediate initiation of any interventional procedure or immediate transfer of the patient to a tertiary care
centre.
Acknowledgments E. Gayat and A. Mebazaa had full access to all
of the data in the study and take responsibility for the integrity of
the data and the accuracy of the data analysis.
The authors are very grateful to Annie Gouverneur, Alexandrine
Ferrand and Dr. Hosni Khouadja for their exceptional contribution
to this work. They also thank the nurses and the whole teams of
Anaesthesia and Intensive care, Obstetric and Radiology Departments. They are grateful to Dr. Sylvie Joubert from the E.F.S
(Etablissement Francais du Sang) and to Didier Castiel for their
help. The authors also gratefully acknowledge the SAMU (Service
dAide Medicale Urgente) of all five departments of Ile-de-France
who were able to transfer the parturient patients from the primary
institutions to Lariboisie`re Hospital in an average of 30 minutes.
Finally, the authors are grateful to Peter Karpati for editing the
manuscript and to Raphael Porcher for assisting with the statistical
analysis.

Conflict of interest None of the authors have any financial

relationships to disclose concerning this work.

Appendix
The multicentre validation study investigators
Assistance Publique-Hopitaux de Paris and Universite
Paris 7, 6, 11. (1) Hopital Lariboisie`re (Paris): Etienne
Gayat, Alexandre Mebazaa, Marie-Jose`phe Laisne. (2)
Hopital Beaujon (Clichy-La Garenne): Claire Bonneville,
Jean Mantz. (3) Hopital de la Pitie`-Salpetrie`re (Paris):
Armelle Nicolas-Robin, Olivier Langeron. (4) Hopital
Becle`re (Clamart): Frederic Mercier, Sandrine RogerChristoph.
Center Hospitalier Universitaire de Limoges:
Nathalie Nathan-Denizot, Anne Vincelot.
Center Hospitalier Universitaire de Besancon,
Universite de Franche-Comte: Emmanuel Samain,
Frederique Bartholin, Gilles Blasco.
Center Hospitalier Universitaire de Nmes: JeanYves Lefrant, Francoise Casano, Agne`s Cuvillon.

References
7. Prasertcharoensuk W, Swadpanich U,
1. World Health Organization (2009)
Lumbiganon P (2000) Accuracy of the
WHO guidelines for the management of
blood loss estimation in the third stage
postpartum haemorrhage and retained
of labor. Int J Gynaecol Obstet
placenta. WHO Press, Geneva. ISBN
71:6970
978-92-4-159851-4
8. Karpati PC, Rossignol M, Pirot M,
2. Khan KS, Wojdyla D, Say L,
Cholley B, Vicaut E, Henry P,
Gulmezoglu AM, Van Look PF (2006)
Kevorkian JP, Schurando P, Peynet J,
WHO analysis of causes of maternal
Jacob D, Payen D, Mebazaa A (2004)
death: a systematic review. Lancet
High incidence of myocardial ischemia
367:10661074
during postpartum hemorrhage.
3. Potts M, Hemmerling A (2006) The
Anesthesiology 100:3036
worldwide burden of postpartum
9. Gulmezoglu AM, Villar J, Ngoc NT,
haemorrhage: policy development
Piaggio G, Carroli G, Adetoro L,
where inaction is lethal. Int J Gynaecol
Abdel-Aleem H, Cheng L, Hofmeyr G,
Obstet 94(Suppl 2):S116S121
Lumbiganon P, Unger C, Prendiville W,
4. Knight M; UKOSS (2007) Peripartum
Pinol A, Elbourne D, El-Refaey H,
hysterectomy in the UK: management
Schulz K (2001) WHO multicentre
and outcomes of the associated
randomised trial of misoprostol in the
haemorrhage. BJOG 114:13801387
management of the third stage of
5. Subtil D, Somme A, Ardiet E, Depretlabour. Lancet 358:689695
Mosser S (2004) Postpartum
10. Winter C, Macfarlane A, Deneuxhemorrhage: frequency, consequences
Tharaux C, Zhang WH, Alexander S,
in terms of health status, and risk
Brocklehurst P, Bouvier-Colle MH,
factors before delivery. J Gynecol
Prendiville W, Cararach V, van
Obstet Biol Reprod (Paris)
Roosmalen J, Berbik I, Klein M, Ayres33:4S94S16
de-Campos D, Erkkola R, Chiechi LM,
6. Goffinet F, Mercier F, Teyssier V,
Langhoff-Roos J, Stray-Pedersen B,
Pierre F, Dreyfus M, Mignon A,
Troeger C (2007) Variations in policies
Carbonne B, Levy G (2005) Postpartum
for management of the third stage of
haemorrhage: recommendations for
labour and the immediate management
clinical practice by the CNGOF
of postpartum haemorrhage in Europe.
(December 2004). Gynecol Obstet
BJOG 114:845854
Fertil 33:268274

11. American Society of Anesthesiologists

Task Force on Obstetrical Anesthesia
(1999) Practice guidelines for
obstetrical anesthesia: a report by the
American Society of Anesthesiologists
Task Force on Obstetrical Anesthesia.
Anesthesiology 90:600611
12. American Society of Anesthesiologists
Task Force on Obstetric Anesthesia
(2007) Practice guidelines for obstetric
anesthesia: an updated report by the
American Society of Anesthesiologists
Task Force on Obstetric Anesthesia.
Anesthesiology 106:843863
13. Ahonen J, Stefanovic V, Lassila R
(2010) Management of post-partum
haemorrhage. Acta Anaesthesiol Scand
54:11641178
14. Driessen M, Bouvier-Colle MH,
Dupont C, Khoshnood B, Rudigoz RC,
Deneux-Tharaux C (2011) Postpartum
hemorrhage resulting from uterine
atony after vaginal delivery: factors
associated with severity. Obstet
Gynecol 117:2131
15. Audureau E, Deneux-Tharaux C,
Lefevre P, Brucato S, Morello R,
Dreyfus M, Bouvier-Colle MH (2009)
Practices for prevention, diagnosis and
management of postpartum
haemorrhage: impact of a regional
multifaceted intervention. BJOG
116:13251333

1825

16. Gissler M, Deneux-Tharaux C,

Alexander S, Berg CJ, Bouvier-Colle
MH, Harper M, Nannini A, Breart G,
Buekens P (2007) Pregnancy-related
deaths in four regions of Europe and the
United States in 19992000:
characterisation of unreported deaths.
Eur J Obstet Gynecol Reprod Biol
133:179185
17. Deneux-Tharaux C, Carmona E,
Bouvier-Colle MH, Breart G (2006)
Postpartum maternal mortality and
cesarean delivery. Obstet Gynecol
108:541548
18. Bouvier-Colle MH, Pequignot F, Jougla
E (2001) Maternal mortality in France:
frequency, trends and causes. J Gynecol
Obstet Biol Reprod (Paris) 30:768775
19. Bouvier-Colle MH, Salanave B, Ancel
PY, Varnoux N, Fernandez H,
Papiernik E, Breart G, Benhamou D,
Boutroy P, Caillier I, Dumoulin M,
Fournet P, Elhassani M, Puech F,
Poutot C (1996) Obstetric patients
treated in intensive care units and
maternal mortality. Regional Teams for
the Survey. Eur J Obstet Gynecol
Reprod Biol 65:121125
20. Bouvier-Colle MH, Varnoux N (2001)
Maternal mortality and severe
morbidity in 3 French regions: results of
MOMS, a European multicenter
investigation. J Gynecol Obstet Biol
Reprod (Paris) 30:S5S9

21. Bouvier-Colle MH, Varnoux N, Costes

P, Hatton F (1991) Maternal mortality
in France. Frequency and reasons for its
under-estimation in the statistics of
medical cause of death. Group of
experts on maternal mortality.
J Gynecol Obstet Biol Reprod (Paris)
20:885891
22. Weeks A (2006) Oral misoprostol for
postpartum haemorrhage. Lancet
368:2123
23. Charbit B, Mandelbrot L, Samain E,
Baron G, Haddaoui B, Keita H, Sibony
O, Mahieu-Caputo D, Hurtaud-Roux
MF, Huisse MG, Denninger MH, de
Prost D (2007) The decrease of
fibrinogen is an early predictor of the
severity of postpartum hemorrhage.
J Thromb Haemost 5:266273
24. Padmanabhan A, Schwartz J, Spitalnik
SL (2009) Transfusion therapy in
postpartum hemorrhage. Semin
Perinatol 33:124127
25. Vanderschueren S, De Weerdt A,
Malbrain M, Vankersschaever D, Frans
E, Wilmer A, Bobbaers H (2000)
Thrombocytopenia and prognosis in
intensive care. Crit Care Med
28:18711876
26. Van Cott EM (2009) Point-of-care
testing in coagulation. Clin Lab Med
29:543553

27. Lewandrowski K (2009) Point-of-care

testing for cardiac markers in acute
coronary syndromes and heart failure.
Clin Lab Med 29:561571
28. Knight M, Callaghan WM, Berg C,
Alexander S, Bouvier-Colle MH, Ford
JB, Joseph KS, Lewis G, Liston RM,
Roberts CL, Oats J, Walker J (2009)
Trends in postpartum hemorrhage in
high resource countries: a review and
recommendations from the
International Postpartum Hemorrhage
Collaborative Group. BMC Pregnancy
Childbirth 9:55
29. Lu MC, Fridman M, Korst LM,
Gregory KD, Reyes C, Hobel CJ,
Chavez GF (2005) Variations in the
incidence of postpartum hemorrhage
across hospitals in California. Matern
Child Health J 9:297306
30. Waterstone M, Bewley S, Wolfe C
(2001) Incidence and predictors of
severe obstetric morbidity: case-control
study. BMJ 322:10891093