British Journal of Anaesthesia 1997; 78: 678683

Pre-anaesthetic assessment of coagulation abnormalities in obstetric

patients: usefulness, timing and clinical implications

L. SIMON, T. M. SANTI, P. SACQUIN AND J. HAMZA

Summary
The usefulness and optimal timing of laboratory
coagulation tests before obstetric extradural
analgesia
are
controversial.
Moreover,
the
significance of mild coagulation abnormalities
during pregnancy remains unclear. We have
assessed the reliability of coagulation tests
performed several weeks before delivery as
predictors of coagulation abnormalities during
labour.
Platelet
count,
plasma
fibrinogen
concentration,
prothrombin
time
(PT)
and
activated partial thromboplastin time (aPTT)
were sampled in 797 women during the ninth
month of pregnancy and checked during labour.
Platelet count was less than 100
109 litre91 for
11 women during labour. Only three had been
detected by the first sample. Platelet count less
than 100
109 litre91 or fibrinogen concentration
less than 2.9 g litre91 during labour were
associated with an increase in the incidence of
postpartum haemorrhage (odds ratio:19.7). We
conclude that a platelet count several weeks
before delivery was not reliable in predicting
thrombocytopenia
during
labour
and
that
women with mild coagulation abnormalities in
early
labour
may
need
special
attention
regarding the risk of postpartum haemorrhage,
(Br. J. Anaesth. 1997; 78: 678683).
Key words
Anaesthesia, obstetric. Anaesthetic techniques, extradural.
Blood, haemostasis. Blood, coagulation.

In a recent survey of 435 French obstetric units,1 most

anaesthetists advocated laboratory coagulation tests
before extradural analgesia for labour. In this study,
coagulation tests were performed during the ninth
month of pregnancy in 74% of obstetric units
surveyed. They included prothrombin time (PT),
activated partial thromboplastin time (aPTT) and
platelet count in almost all cases- Although haemostasis disorders seem to increase the risk of extradural
haematoma,2 the usefulness of laboratory coagulation
screening tests is still debatable.3 However, platelet
count can decrease during pregnancy4 and the
incidence of thrombocytopenia has been shown to
increase during labour.5 The significance of this
thrombocytopenia remains unclear and determination

of a safe lower platelet count limit before performing

extradural analgesia is still controversial. Many
anaesthetists avoid extradural analgesia if platelet
count is less than 100
109 litre91,68 but this limit has
no supporting data.9 Another difficulty in coagulation
assessment is the choice of when to perform it.
Haemostasis disorders can appear late in pregnancy.
Therefore, a medical history, physical examination and
laboratory testing may be inadequate in detecting these
haemostasis disorders if they are performed too
early.10 In addition, no relationship between mild
coagulation abnormalities during pregnancy and
haemorrhagic complications during extradural
analgesia11 or delivery12 have yet been established.
Therefore, we designed a prospective study to
evaluate the accuracy of laboratory coagulation
tests performed during the ninth month of
pregnancy to predict coagulation abnormalities
during labour. We also examined the incidence of
these coagulation abnormalities during labour and
their relationship with post-partum haemorrhage
(PPH).

Patients and methods

This prospective surveillance study was performed
between 1 June 1995 and 31 December 1995.
During this period, we studied all women with a
singleton pregnancy who came into the labour ward
for delivery. As the study was an anonymous survey
of haematological and clinical variables, hospital
Ethics Committee approval was not required.
Inclusion criteria were: pre-anaesthetic clinical
assessment during the ninth month of pregnancy,
physical status ASA I or II, and term 36 and 42
weeks at the time of delivery. We excluded all
women with a medical history known to be associated with haemostasis disorders, such as von
Willebrand disease, diabetes mellitus, idiopathic
thrombocytopenic purpura or lupus erythematosus.
We also excluded women receiving anticoagulant
therapy and parturients admitted for elective
Caesarean section.

L. SIMON, MD, T. M. SANTI, MD, P. SACQUIN, MD, J. HAMZA, MD,

Dpartement dAnesthsie-Ranimation, Hpital St Vincent de
Paul 7482, Avenue Denfert, Rochereau, 75014 Paris, France.
Accepted for publication: January 28, 1997.

Coagulation assessment in obstetric anaesthesia

STUDY DESIGN

The first blood sample (BS 1) for laboratory

coagulation tests was obtained in each parturient
during the ninth month of pregnancy, usually just
after pre-anaesthetic clinical assessment. Platelet
count, PT, aPTT and plasma fibrinogen
concentration were determined in each sample.
When the women entered the labour ward for
delivery, a second blood sample (BS 2) was
obtained for the same tests via a 16-gauge i.v.
cannula just before the start of a basal infusion of
Ringers lactate solution. Age, parity, medical
and obstetric history, and gestational age (GA)
were obtained by an anaesthetist, who enquired
again of any previous history of bleeding. He
also noted on the anaesthetic chart all recent
antepartum medical complications such as preeclampsia, sepsis or placenta praevia. Each
abnormal coagulation test result was checked
systematically in a new sample. Moreover,
thrombocytopenia (platelet count 100
109
litre91) was confirmed by examination of the
stained peripheral blood film to exclude platelet
clumping. BS 2 was always available before the
decision to proceed with extradural analgesia. If
indicated, extradural analgesia was provided for
labour in each parturient whose platelet count was
more than 100
109 litre91 at BS 2.
When the women left the labour ward, the route
of delivery (vaginal delivery or Caesarean section)
and the occurrence of PPH were noted. Only
significant haemorrhages were considered. These
were defined as clinically severe bleeding requiring
volume expansion and manual uterine exploration,
associated with either blood transfusion or a
significant decrease (2 g dl91) in haemoglobin
concentration, or both, within 24 h post-partum.
We compared for each parturient the results of
BS 1 and BS 2, and we calculated the changes in
platelet count (%) and fibrinogen concentration
(g litre91) between the two periods. Then we
examined mean (SD) platelet count and fibrinogen
concentration in BS 2. Values less than the lower
threshold of the corresponding 95% confidence
limits (value mean value 91.96 SD) for one or
both variables, or with a PT 70% or an aPTT
40 s (normal 30 s) were classified as coagulation
abnormalities. We compared the incidence of
severe PPH in this group and in the other women.

679
Table 1 Information on 38 women excluded from the study.
Reasons for exclusion are indicated
n
Age (yr) (mean (range))
Gestational age (weeks) mean (SD))
Reasons for exclusion in labour ward
Placenta praevia
Placenta abruptio
Fever 38.5C sepsis syndrome
BS 2 not available
Low molecular weight heparin therapy
Hypertension and/or pre-eclampsia

38
30.9 (2443)
38.2 (2.4)
8
2
6
10
3
9

Figure 1 Variations (in absolute value) in platelet count

between the two samples, BS 1 and BS 2, for each woman.
These variations (Y axis) in function of the BS 1 value for
platelet count (X axis) are shown. Mean variation (9) and 95%
confidence limits (---) for these variations are also indicated.

the results and statistical analyses, and 38 were

excluded. The characteristics of these 38 patients
in the labour ward are shown in table 1. Mean delay
between BS 1 and BS 2 was 17 (9) days (always less
than 30 days).
PLATELET COUNT

Mean values for platelet count in BS 1 and BS 2

STATISTICAL ANALYSIS

Results are expressed as mean (SD). For quantitative data, statistical analysis was performed using
analysis of variance for repeated measures. When
overall differences were detected, individual
comparisons between groups were performed by
Students t test with Bonferronis correction.
Fishers exact test was used to compare nominal
data.

Results
We studied 835 parturients: 797 were included in

Figure 2 Variations (in absolute value) in platelet count

between the two samples, BS 1 and BS 2, for women who had a
platelet count less than 100
109 litre91 (mean value 91.96 SD
at BS 2) in either one or both samples. The shaded area
indicates the area below 100
109 litre91.

680

British Journal of Anaesthesia

926 (20) % for these 11 thrombocytopenic
patients whereas it was 93.7 (14) % in the overall
population.
PT, aPTT AND PLASMA FIBRINOGEN
CONCENTRATION

Figure 3 Variations (in absolute value) in plasma fibrinogen

concentration between the two samples, BS 1 and BS 2, for
women who had a plasma fibrinogen concentration less than
2.9 g litre91 (mean value 91.96 SD at BS 2) in either one or both
samples. The shaded area indicates the area below 2.9 g litre91.

were,
respectively,
216
(54)
109
litre91
9
(extremes 88538) and 206 (54)
10 litre91
(extremes 48415). From BS 1 to BS 2, platelet
count was variable and unpredictable in all
women (fig. 1).
The lower threshold of the 95% confidence limit
for platelet count was 100
109 litre91 (BS 2 mean
value 91.96 SD:206 9(1.96
54):100
109
litre91). Four women were thrombocytopenic
(100
109 litre91) at BS 1. For one patient,
platelet count increased to 132
109 litre91 at
BS 2, whereas for the others, platelet count
at BS 2 remained less than 100
109 litre91.
Thrombocytopenia (100
109 litre91) appeared
in eight other women at BS 2. Therefore, a total of
11 women had platelet counts less than 100
109
litre91 at BS 2. All of these parturients had platelet
counts less than 160
109 litre91 at BS 1 (fig. 2).
Mean variation between the two samples was

There were no abnormal values for PT or aPTT in

either BS 1 or BS 2. Mean values for fibrinogen
concentrations in BS 1 and BS 2 were, respectively,
4.6 (0.8) g litre91 (extremes 2.07.6) and 4.8 (1.0)
g litre91 (extremes 2.19.0). The lower threshold of
the 95% confidence limit for plasma fibrinogen
concentration was 2.9 g litre91 (BS 2 mean value
91.96
SD). Thirteen parturients had fibrinogen
concentrations less than 2.9 g litre91 at BS 1.
Twenty-six parturients had fibrinogen concentrations less than 2.9 g litre91 at BS 2. For these
women, the variation in fibrinogen concentration
from BS 1 to BS 2 was 91.23 (0.88) g litre91,
whereas it was ;0.18 (0.84) g litre91 in the overall
population. As observed with platelet count, BS 1
was not reliable in detecting women with fibrinogen
concentrations less than 2.9 g litre91 at BS 2
(fig. 3).
COAGULATION DISORDERS IN EARLY LABOUR AND
PPH

The mode of delivery was vaginal for 677 women

and a Caesarean section was performed in the
other 120. Among the 677 who had a vaginal
delivery, 46 had significant PPH (table 2) and
all occurred in patients asymptomatic for
haemostasis disorders before delivery. They were
all considered as clinically significant haemorrhages
requiring volume expansion (Ringers solution,
blood transfusion, or both) and manual uterine
exploration. In two patients, hysterectomy was
necessary to stop massive haemorrhage. Mean

Table 2 Number of patients and main clinical characteristics of women who had Caesarean section or vaginal
delivery with or without post-partum haemorrhage (PPH). *P0.05 vs value in women who had vaginal delivery
with no PPH (Students t test)
Vaginal delivery
Caesarean sections
n
Age (yr) (mean (range))
Gestational age (weeks)
Haematology values in labour ward (BS 2)
Platelet count (109 litre91)
Fibrinogen level (g litre91)

120
32.7 (2145)
39.3 (1.5)
(mean (SD))
206 (61)
5.0 (0.9)

PPH
46
30.9 (2141)
39.8 (1.2)
185 (59)*
4.3 (1.3)*

No PPH
631
31.5 (1946)
39.5 (1.3)
208 (53)
4.9 (1.0)

Table 3 Sensitivity specificity, positive and negative predictive values of fibrinogen concentration 2.9 g litre91, at
either BS 1 or BS 2, predicting occurrence of PPH

Fibrinogen 2.9 g litre91 at

BS 1
BS 2

Sensitivity
(%)

Specificity
(%)

Positive
predictive value (%)

Negative
predictive value (%)

4.3
19.6

98.4
97.8

16.6
39.1

93.4
94.3

Coagulation assessment in obstetric anaesthesia

681

Figure 4 Incidence of post-partum haemorrhage (PPH) in women with platelet counts or fibrinogen concentrations
at BS 2 less than 100
109 litre91 and 2.9 g litre91, respectively. The incidence was significantly different compared
with the rate observed in women with both platelet count and fibrinogen concentrations greater than these thresholds
(**P0.01; Fischers exact test). Rates are indicated in percent (Y axis) and absolute values.

decrease in haemoglobin concentration within

24 h post-partum was 4.0 (1.5) g dl91 for these
46 patients.
Platelet count and fibrinogen concentration at
BS 2 were significantly lower in women with
PPH than in those who did not bleed (table 2). If
one or both of these variables were below the
corresponding mean value 91.96 SD threshold,
the incidence of PPH was increased significantly
compared with the other women (fig. 4).
Sensitivity, specificity, positive and negative
predictive values for both BS 1 and BS 2 for
low fibrinogen predicting subsequent PPH are
presented in table 3.

Discussion
In our study, the percentage of women with
platelet count values less than 100
109 litre91 and
less than 150
109 litre91 increased, respectively,
from 0.5 to 1.4% and from 7.3 to 12.4%
between BS 1 and BS 2. These results are in
agreement with those of Rolbin and colleagues who
measured platelet count in 1621 healthy women.9
Only seven (0.4%) and 104 (6.4%) had platelet
counts less than 100 and 150
109 litre91,
respectively. However, these results were obtained
either during pregnancy or in the post-partum
period. All patients were asymptomatic, but some
had significant thrombocytopenia (lower platelet
count 21
109 litre91). Burrows and Kelton found
in 1357 healthy women an incidence of
8.3%
asymptomatic
mild
thrombocytopenia
(97150
109 litre91) during labour.5
Gestational thrombocytopenia is quite common
but its significance is still unclear. The decrease in
platelet count could be related to haemodilution
and acceleration of platelet turnover with
increased production of platelets from the bone
marrow and increased trapping or destruction at
the placental level. This quantitative decrease in
platelet count could be counterbalanced by a
marked increase in platelet reactivity during

pregnancy.13 On the other hand, some authors

have described decreased platelet activation
during pregnancy14 15 and suggested the presence
of a plasma factor that selectively inhibits
prostaglandin-dependent activation.15 Therefore, it
remains difficult in clinical practice to define
precisely the clotting capacities of a woman with a
mild gestational thrombocytopenia.
The risk of spinal haematoma has been estimated
to be approximately 1 in 150 000 after extradural
analgesia.2 Even if the risk of this complication is
increased in patients with a low platelet count, it is
difficult to define the safe lower threshold for
platelet count. The haematological definition of
thrombocytopenia is a platelet count less than
150
109 litre91. The lower limit of 100
109 litre91
for platelet count, often judged as safe to perform
extradural analgesia,68 has no supporting data. In
the absence of any other coagulation disorder,
some authors would prefer to lower this limit to
80
109 litre91.16 Nevertheless, some authors do
not systematically observe the platelet count before
extradural analgesia.7 8 Rolbin and colleagues
estimated that in their institution approximately
5000 thrombocytopenic parturients have benefitted
from extradural analgesia over a 30-yr period
without any spinal haematomas.9 The same team
has reported the case of uncomplicated extradural
analgesia in an asymptomatic parturient whose
platelet count was 2
109 litre91 immediately after
delivery.7
Whatever the significance of these low platelet
counts and the acceptable limit, one remaining
problem is the choice of the appropriate time to
perform a platelet count. Indeed, some authors
found a decrease in mean platelet count whereas
others reported no change in platelet count during
pregnancy.17 Many anaesthetists consider it sensible
to obtain the results of coagulation tests several
weeks before extradural analgesia.1 However, our
study showed that the result of this nine month
platelet count was useless in predicting platelet
count less than 100
109 litre91 during labour.

682
The nature of the other laboratory tests
which are useful before extradural analgesia and
their lower safe value are still debatable.
Coagulation tests such as PT and aPTT
measurements have not been studied extensively
in obstetric patients but they seem to be less useful
for early detection of haemostasis disorders
during labour.18 Indeed, these tests were always
normal in our study. Therefore, we believe that
these tests should no longer be performed in
clinically normal parturients before extradural
analgesia.
Fibrinogen concentration increases physiologically during pregnancy.19 20 A safe lower limit
for plasma fibrinogen concentration has not been
evaluated, while the safe laboratory lower limit
defined for the general population may not be
appropriate for pregnant women. Nevertheless, we
had to define a lower limit of normality for both of
these variables. Therefore, all values for either
platelet count or fibrinogen concentration at BS 2
which were below the lower threshold of the
corresponding 95% confidence limit were
considered as abnormal.
Our study has shown that significant PPH were
more frequent when platelet count or fibrinogen
concentration, or both, were low in early labour.
In previous studies, this association has not
been recognized as a significant factor associated
with PPH.12 The significance of the relationship
between these mild coagulation abnormalities and
the occurrence of PPH is not clear. During
pregnancy, biological changes occur such as an
increase in many clotting factor concentrations and
placental synthesis of plasminogen activator
inhibitor. Early in pregnancy, intravascular fibrin
deposition can be found in the uteroplacental
circulation.19
This
reflects
local
chronic
intravascular coagulation which induces local
fibrinolysis.21 22 These biological changes are
associated with mechanical factors that limit the
risk of bleeding, such as structural changes in the
spiral arteries and myometrial contractions.6 19
These physiological changes must prepare the
pregnant woman for delivery and placental separation and the inherent risk of haemorrhage. In case
of failure of the mechanical protective mechanisms,
there is an increased need for haemostatic
components such as fibrinogen and platelets. We
may hypothesize that in such situations there is
inadequate compensation which is reflected in a
low platelet count and a low fibrinogen plasma
concentration. Combs, Murphy and Laros showed
that the main clinical variables related to the risk of
PPH were prolonged labour and factors associated
with uterine atony.12 If impaired uterine
contractility occurred in patients with a low platelet
count
and
fibrinogen
concentration,
local
coagulation factors may not ensure good haemostasis and therefore, at the time of delivery, an
increased incidence of PPH could occur. This
hypothesis of a relationship between obstetric
factors and mild coagulation abnormalities may
explain the occurrence of PPH, but this will require
further studies.

British Journal of Anaesthesia

In clinical practice, detection of coagulation
abnormalities,
and
especially
detection
of
thrombocytopenia, may be important in the
management of the obstetric patient. Nevertheless, unhelpful coagulation tests should be
avoided.
Our
results
demonstrated
that
coagulation tests performed several weeks before
labour should be abandoned and that PT and
aPTT are unnecessary in clinically normal
parturients.

References
1. Souron V, Simon L, Sacquin P, Santi TM, Delaporte
Cerceau S, Hamza J, Saint Maurice C. Coagulation
assessment before epidural analgesia in obstetric patients:
A French national survey. British Journal of Anaesthesia
1996; 76: A321.
2. Vandermeulen EP, Van Aken H, Vermylen J.
Anticoagulants and spinalepidural anesthesia Anesthesia
and Analgesia 1994; 79: 11651177.
3. Watel A, Jude B, Caron C, Vandeputte H, Gaeremynck E,
Cosson A. Heurs et malheurs du temps de cphaline active
dans
le
bilan
propratoire.
Annales
Franaises
dAnesthesiologie et de Ranimation 1986; 5: 3539.
4. Pitkin RM, Witte DL. Platelet and leukocyte counts in
pregnancy. Journal of the American Medical Association 1979;
242: 26962698.
5. Burrows RF, Kelton JG. Incidentally detected thrombocytopenia in healthy mothers and their infants. New England
Journal of Medicine 1988; 319: 142145.
6. Douglas J. Coagulation abnormalities and obstetric anaesthesia. Canadian Journal of Anaesthesia 1991; 38: R1721.
7. Hew-Wing P, Rolbin SH, Hew E, Amato D. Epidural
anaesthesia and thrombocytopenia. Anaesthesia 1989; 44:
775777.
8. Bromage PR. Neurologic complications of regional
anesthesia for obstetrics. In: Shnider SM, Levinson G, eds.
Anesthesia for Obstetrics, 3rd Edn, Baltimore: Williams and
Wilkins, 1993; 433454.
9. Rolbin SH, Abbot D, Musclow E, Papsin F, Lie LM,
Freedman J. Epidural anesthesia in pregnant patients with
low platelet counts. Obstetrics and Gynecology 1988; 71:
918920.
10. How HY, Bergmann F, Koshy M, Chediak J, Presperin C,
Gall SA. Quantitative and qualitative abnormalities during
pregnancy. American Journal of Obstetrics and Gynecology
1991; 164: 9298.
11. Hamza J, El Haddouri M, Langeron O. Low platelet count
before epidural is frequent in healthy parturients with post
partum hemorrhage. Anesthesiology 1993; 77: A1022.
12. Combs CA, Murphy EL, Laros RK. Factors associated with
postpartum hemorrhage with vaginal birth. Obstetrics and
Gynecology 1991; 77: 6976.
13. Morrison R, Crawford J, Mac Pherson M, Heptinstall S.
Platelet behaviour in normal pregnancy, pregnancy
complicated by essential hypertension and pregnancyinduced hypertension. Thrombosis and Haemostasis 1985; 54:
607611.
14. OBrien WF, Saba HI, Knuppel RA, Scerbo JC, Cohen
GR. Alterations in platelet concentration and aggregation in
normal pregnancy and preeclampsia. American Journal of
Obstetrics and Gynecology 1986; 155: 486490.
15. Nicolini U, Guarneri D, Gianotti GA, Campagnoli C,
Crosignani PG, Gatti L. Maternal and fetal platelet
activation in normal pregnancy. Obstetrics and Gynecology
1994; 83: 6569.
16. Letsky EA. Hemostasis and epidural anesthesia.
International Journal of Obstetric Anesthesia 1991; 1: 5154.
17. Tygart SG, Mac Royan DK, Spinnato JA. Longitudinal
study of platelet indices during normal pregnancy.
American Journal of Obstetrics and Gynecology 1986; 154:
883887.
18. Barker P, Callander CC. Coagulation screening before
epidural analgesia in pre-eclampsia. Anaesthesia 1991; 46:
6467.

Coagulation assessment in obstetric anaesthesia

19. Letsky E. Haemostasis in pregnancy. In: Morgan BM, ed.
Foundations of Obstetric Anaesthesia. London: Farrand Press,
1987; 189207.
20. Cheek TG, Gutsche BB. Maternal physiologic alterations
during pregnancy. In: Shnider SM, Levinson G, eds.
Anesthesia for Obstetrics, 3rd Edn. Baltimore: Williams and
Wilkins, 1993; 317.

683
21. Gerbasi FR, Bottoms S, Farag A, Mammen EF. Changes in
hemostasis activity during delivery and the immediate postpartum period. American Journal of Obstetrics and Gynecology
1990; 162: 11581163.
22. Fletcher AP, Alkjaersig NK, Burstein R. The influence of
pregnancy upon blood coagulation and plasma fibrinolytic
function. American Journal of Obstetrics and Gynecology
1979; 134: 743751.