Human Reproduction Vol.21, No.5 pp. 1291–1294, 2006
doi:10.1093/humrep/dei456
Advance Access publication January 12, 2006.
Histological classification of chorionic villous vascularization
in early pregnancy
R.A.Hakvoort1,2, B.A.M.Lisman2, K.Boer2, O.P.Bleker2, K.van Groningen3,
M.van Wely2 and N.Exalto1,2,4
1
Department of Obstetrics and Gynaecology, Spaarne Ziekenhuis, Hoofddorp, 2Division of Obstetrics and Gynaecology, Academic
Medical Centre, University of Amsterdam, Amsterdam and 3Clinical Pathology, Spaarne Ziekenhuis, Hoofddorp, The Netherlands
4
To whom correspondence should be addressed at: Spaarne Ziekenhuis Hoofddorp, PO Box 770, 2130 AT Hoofddorp, The Netherlands.
E-mail: [email protected]
BACKGROUND: The objective of the study was to assess the reproducibility of a new classification for early pregnancy chorionic villous vascularization (Grade: I, normal; IIA, mild hypoplasia; IIB, severe hypoplasia and III, avascular) for routine microscopic examination in daily clinical practice. METHODS: In this observational study, four
observers scored first trimester chorionic villous vascularization. Scoring was performed in microscopic slides of chorionic tissue obtained by D&C in 30 patients with early pregnancy loss due to empty sac (n = 10), fetal death (n = 10)
and termination of pregnancy (n = 10) using the new classification. Ultrasonographic measurement of trophoblastic
thickness (TT) at the implantation site was available in all patients and in a reference group of 100 ongoing singleton
pregnancies. The vascularization score could therefore be related to the TT. RESULTS: The new classification
resulted in a good-to-excellent agreement in histological scoring (0.73–0.90) between investigators (kappa 0.64–0.86).
TT was not related to either vascularization or pregnancy outcome and only partly to hydropic degeneration.
CONCLUSION: The vascularization scoring system is a simple, valid and effective method for assessment of chorionic villous vascularization. It is helpful in understanding the underlying cause of pregnancy loss, as the classification
can distinguish between normal and abnormal embryonic development. We did not find either a relation between TT
and pregnancy outcome or between TT and vascularization.
Key words: chorionic villous vascularization/early pregnancy/trophoblastic thickness
Introduction
First trimester miscarriages are mainly due to chromosomal or
structural abnormalities of the conceptus, resulting in an empty
sac or fetal death. Morphological examination and documentation of abnormal development is difficult. However, it is
important in determining why pregnancy loss occurred (Blanch
et al., 1998; Quenby et al., 2002; Dawood et al., 2004).
Although ultrasound is the first step in morphologic documentation, it is not always conclusive with regard to the
underlying cause, especially in case of fetal death. Routine
cytogenetic investigation of miscarriage specimens is
expensive and not always successful. Several studies have
been performed to find an association between histological
features and chromosomal abnormalities in miscarriages
(Minguillon et al., 1989; Roberts et al., 2000). A clear correlation between these histomorphologic features and abnormal karyotype was only found in triploids (Fox, 1993; van
Lijnschoten et al., 1994a). Villous morphology, relying
for an important part on hypovascularization of peripheral
villi, unfortunately turned out to be an inaccurate indicator
of chromosomal abnormalities (Minguillon et al., 1989;
van Lijnschoten et al., 1994b; Genest et al., 1995; Jauniaux
and Hustin, 1998).
Histomorphologic studies on first trimester chorionic villous
vascularization at our department, using CD34 immunohistochemistry, have formed the basis for our understanding of normal
and abnormal placental development during organogenesis.
Vasculogenesis starts from haemangioblastic cell cords, which
are the precursors of both the capillary endothelium and haematopoietic stem cells. These basic structures can easily be
found in trophoblastic tissue of empty sacs and complete
hydatidiform mole pregnancies (Lisman et al., 2004, 2005).
Vasculogenesis of normal chorionic villi is characterized by
maturation of these haemangioblastic cell cords to luminized
vessels and margination to peripherally located vessels at the
end of organogenesis, forming the vasculosyncytial membrane
(te Velde et al., 1997). A defective development of the vasculosyncytial membrane is seen in pregnancies complicated by
fetal death and even more pronounced in empty sacs (Meegdes
et al., 1988; Exalto and te Velde, 1994; Lisman et al., 2004).
Hypovascularization turned out to be a marker for abnormal
fetal development, without being influenced by prolonged
© The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
1291
R.A.Hakvoort et al.
intrauterine retention after fetal death. (Meegdes et al., 1988).
Vascular changes in first trimester chorionic villi are the result
of defective development due to abnormal vasculogenesis
rather than to postmortem changes (Lisman et al., 2004).
We designed a histological classification for evaluation of
chorionic villous vascularization in routine clinical practice.
This offered an opportunity to study vascularization against the
background of overall trophoblastic development determined
by ultrasonographic appearance (Bajo et al., 2000).
The aim of our study was to assess the reproducibility and
inter-observer agreement of our histological classification for
chorionic villous vascularization on the basis of the number
and localization of vessels.
• Grade I: normal. Vessels with nucleated blood cells are present in
almost every (at least nine of 10) villus, have a very clear appearance
and are located centrally as well as peripherally (in contact with the
trophoblastic layer). In some villi, the number of vessels are even
numerous (>5).
• Grade IIA: mild hypoplasia. Vessels with nucleated blood cells
are not present in all villi, less numerous and predominantly located
centrally.
• Grade IIB: severe hypoplasia. Villi are predominantly avascular;
however, in a single villus, a vessel is present with one or more nucleated blood cells.
• Grade III: avascular. All villi are avascular, although sporadically
a very small vessel, with or without a nucleated blood cell may be
present.
Materials and methods
Hydropic degeneration is registered additionally if present in a considerable amount of the villi.
An observational study was carried out at our department during recent
years combining ultrasound and histological investigation in early pregnancy. The study design was approved by the institutional medical ethical committee. All women were included after informed consent.
Ultrasound scanning of trophoblastic thickness
Trophoblastic thickness (TT) was measured at the implantation site in
all patients undergoing routine vaginal first trimester ultrasound scanning using ultrasound equipment type Aloka (SSD 5500) or Hitachi
(EUB 525) with a 7.5 MHz vaginal probe. The largest part of the
echogenic layer representing the developing trophoblast was measured perpendicular to the chorionic sac. All measurements were performed by three experienced ultrasonographers.
A total of 100 consecutive ultrasonographical TT measurements in
ongoing singleton pregnancies served as the reference group (REF).
A total of 10 consecutive patients were selected for each of the following three subgroups requiring a dilatation and curettage: Group I,
empty sac (ES); Group II, fetal death (FD) and Group III, termination
of pregnancy (TOP). The TT measurements in all the three different
subgroups of early pregnancy loss were compared with the reference
group.
In order to minimize maternal and other influences on chorionic villous vascularization, especially in case of fetal death, we excluded
patients with any known maternal disease including thrombophilia or
pre-eclampsia during previous pregnancies, pregnancies established
by assisted reproductive techniques and multiple pregnancies.
Gestational age (GA) was derived from crown rump length (CRL)
measurements in TOP and FD in the viable state or based on the last
menstrual period in the ES group.
Histology
In all cases, sections of the chorionic villi, obtained by dilatation and
curettage (D&C), were routinely stained with haematoxylin/eosin
(HE) for histological investigation. Observations were carried out by
observers with varying experience to test the expected reproducibility.
Therefore, an experienced gynaecologist, two residents and a pathologist were asked to investigate and classify the slides in one run without
knowing any clinical information. The slides were coded at random
(1–30) by a secretary. In the three subgroups, chorionic villous vascularization was assessed in three slides per specimen according to our
newly developed histological classification as described below.
Classification of vascularization
• Grade 0: unknown. There are insufficient number of villi available for evaluation.
1292
Statistics
Differences between groups in patient characteristics were tested for
significance using analysis of variance.
Reproducibility was expressed by the observed rates of interobserver agreement and by kappa statistics. Kappa ranges between 0
(if agreement is accomplished totally by chance) and 1 (if agreement
is complete). We used the directives of reproducibility as described by
Landis and Koch in 1977: lower than 0, bad; 0.00–0.20, poor; 0.21–
0.40, moderate; 0.41–0.60, reasonable; 0.61–0.80, good and 0.81–
1.00, excellent.
Results
Age did not differ significantly between the groups (Table I).
GA was significantly higher in the ES subgroup as compared
with the FD and TOP subgroup where the GA can not be corrected using a CRL measurement.
Vascularization was normal in nine out of 10 cases of the
TOP subgroup. Due to incomplete sampling, one specimen
showed no chorionic villi in all three slides. There was no
necessity to remove these from the study because this was foreseen in the classification (Grade 0). One specimen showed
hydropically degenerated villi.
Vascularization in ES showed a narrow range of either
severe hypoplasia or avascularity. Three specimens showed
hydropic degeneration, whereas the FD group showed more
varied vascularization ranging from avascular to mild hypoplasia
(Figure 1).
Table I. Patient age and gestational age at the time of measurement of
trophoblastic thickness
Mean age (years)
(SD)
GA based on
CRL (days) (SD)
GA based on
LMP (days) (SD)
TOP (n = 10)
FD (n = 10)
ES (n = 10)
REF (n = 10)
33.2 (4.7)
32.0 (3.1)
33.0 (4.6)
33.8 (4.8)
53.4 (11.8)
52.9 (7.8)
52.9 (11.9)
72.8 (9.0)
60.3 (12.0)
75.4 (12.5)a
60.4 (11.8)
CRL, crown rump length; ES, empty sac; FD, fetal death; GA, gestational age;
LMP, last menstrual period; REF, ongoing singleton pregnancy; TOP, termination
of pregnancy.
a
Only gestational age in the ES group differed significantly from all other
groups.
Classification of chorionic villous vascularization
25
trophoblastic thickness (mm)
Vascularization score
I
IIA
IIB
20
15
10
5
III
0
30
40
50
60
70
80
90
gestational age (days)
0
TOP
FD
ES
Early pregnancy loss group
Figure 1. Vascularization score of investigator A in subgroups of
early pregnancy loss.
Table II. Reproducibility between observers A, B, C and D, expressed as the
observed and expected proportion of agreement and kappa values, for the
histological vascularization score in 30 specimens
Observer pair Observed proportion
of agreement
Expected proportion
of agreement
Kappa
A versus B
A versus C
A versus D
B versus C
B versus D
C versus D
0.25
0.25
0.25
0.25
0.25
0.25
0.77 (good)
0.64 (good)
0.86 (excellent)
0.64 (good)
0.73 (good)
0.64 (good)
0.83
0.73
0.90
0.73
0.80
0.73
Observed proportions of agreement between the four
observers, concerning the scoring of the 30 specimens, varied
from 0.73 to 0.90, with kappa values ranging from 0.64 to 0.86
(Table II).
As expected, in the TOP group, TT was similar to the reference group. Thickness in the FD and TOP group did not differ
from the reference group, with one exception in the FD group
where greater thickness was accompanied by hydropic degeneration. In ES, GA could not be corrected using CRL. Therefore, values shown are positioned at the GA reached at the
moment of measurement. TT in the ES and FD group in general was below 1 cm, with one exception in the case of hydropically degenerated tissue in one specimen of the FD group
(Figure 2).
We did not find a relation between TT and vascularization.
This was only partly due to some cases with hydropic degeneration occurring in the ES (n = 3), FD (n = 1) and even the TOP
(n = 1) group.
Discussion
There is evidence that vascularization is diminished in empty
sacs and fetal death (Meegdes et al., 1988; Exalto and te Velde,
▲= empty sac
○ = fetal death
+ = termination of pregnancy
■ = ongoing singleton pregnancy
Figure 2. Trophoblastic thickness in ongoing singleton pregnancies
(REF) and subgroups of early pregnancy failure. ▲, empty sac;
❍, fetal death; +, termination of pregnancy; ■, ongoing singleton
pregnancy.
1994; Lisman et al., 2004). These findings were confirmed in
our study using routine HE staining. We indeed observed a
trend towards severe hypovascularity and avascularity in
empty sacs. In the cases of fetal death, hypoplasia of the vascular system is also present, although less severe. In our study on
the development of the vasculosyncytial membrane, we also
observed significantly more centrally located angiogenetic
cords in the group of fetal death, with the highest number in the
group of empty sac, as compared to normal pregnancies
(Lisman et al., 2004). It is therefore very unlikely that angiogenetic cells are influenced by retention. From these findings,
we concluded that vascular changes in first trimester chorionic
villi are the result of defective development due to abnormal
vasculogenesis rather than due to postmortem changes.
In literature, only limited data are available regarding interobserver variation in studies on early placental morphology.
Although the observed variation for chorionic villous vascularization appeared to be moderate to reasonable, the results of
the different studies are not comparable because of differences
in research design (van Lijnschoten et al., 1993; Nelen et al.,
2000; Lisman et al., 2004).
Our new classification appeared to be useful in the assessment of vascularization, was easy to learn for non-experienced
observers and was readily accepted by experienced observers.
The review judgements of four different observers using our
classification resulted in a good-to-excellent rate of agreement.
Rates of agreement such as obtained in our study are considered to be realistic and acceptable for implementation in daily
clinical practice. Equally important for implementation is that
the classification does not require additional staining and is
therefore relatively cheap. From our previous study, it was
concluded that initiation of placental vasculogenesis is a basic
feature in all types of pregnancy (haemangiogenetic cords) but
is subsequently modulated (from cords to vessels) directly or
1293
R.A.Hakvoort et al.
indirectly by embryonic signalling (Lisman et al., 2004, 2005).
We therefore are convinced that the classification is helpful in
distinguishing between embryopathic and maternal causes.
This distinction may also be useful in recurrent miscarriage
evaluation, simply by using microscopic slides from miscarriages in the past (Dawood et al., 2004).
Ultrasonographic measurement of the TT at the implantation
site has been mentioned as a sensitive predictor of pregnancy
outcome (Bajo et al., 2000). However, we did not find a clear
relation between TT and pregnancy outcome. The only
observed trend was the critical level of 1.0 cm TT. All cases
of early pregnancy failure (except one case with hydropic
degeneration) occurred below this level. Therefore, the use of
TT measurements does not seem to be a valuable tool in predicting pregnancy outcome nor in counselling the pregnant
patient. Neither did we find a relation between TT and vascularization. This was only partly due to some cases with
hydropic degeneration.
The classification can be used in further studies testing the
clinical validity in early, late and recurrent pregnancy loss.
References
Bajo J Moreno-Calvo FJ, Martinez-Cortés L, Haya FJ and Rayward J (2000) Is
trophoblastic thickness at the embryonic implantation site a new sign of negative evolution in first trimester pregnancy? Hum Reprod 15,1629–1631.
Blanch G, Quenby S, Ballantyne E, Holland K, Gosden CM and Nielson JP
(1998) Embryonic abnormalities at medical termination of pregnancy with
mifepristone and misoprostol during the first trimester: observational study.
Br Med J 7146,1712–1713.
Dawood F, Farquharson R and Quenby S (2004) Recurrent miscarriage. Curr
Obstet Gynaecol 14,247–253.
Exalto N and Te Velde J (1994) Vascular system development in early human
chorionic villi. Proceedings 10th Annual Meeting of ESRHE, Brussels. Hum
Reprod 9,10.
Fox H (1993) Histological classification of tissue from spontaneous abortions:
a valueless exercise? Histopathology 22,599–600.
1294
Genest D, Roberts D, Boyd T and Bieber F (1995) Fetoplacental histology as a
predictor of karyotype: a controlled study of spontaneous first trimester
abortions. Hum Pathol 26,201–209.
Jauniaux E and Hustin J (1998) Chromosomally abnormal early ongoing pregnancies. correlation of ultrasound and placental histological findings. Hum
Pathol 29,1195–1199.
Landis RJ and Koch GG (1977) The measurement of observer agreement for
categorical data. Biometrics 33,159–174.
van Lijnschoten G, Arends JW, Leffers P, de la Fuente AA, van der Looij
HJAM and Geraerdts JPM (1993) The value of histomorphological features
of chorionic villi in early spontaneous abortion for the prediction of karyotype. Histopathology 22,25–29.
van Lijnschoten G, Arends JW, Thunnissen BJM et al. (1994a) A morphometric approach to the relation of karyotype, gestational age and histological
features in early spontaneous abortions. Placenta 15,189–200.
van Lijnschoten G, Arends JW and Geraerdts JPM (1994b) Comparison of histological features in early spontaneous and induced trisomic abortions. Placenta 15,765–773.
Lisman BAM, Boer K, Bleker OPV, Wely MV, Groningen K and Exalto N
(2004) Abnormal development of the vasculosyncytial membrane in early
pregnancy failure. Fertil Steril 82 (3),654–660.
Lisman BAM, Boer K, Bleker OP, van Wely M and Exalto N (2005) Vasculogenesis in complete and partial hydatidiform mole pregnancies studied with
CD 34 immunohistochemistry. Hum Reprod 20 (8),2334–2339.
Meegdes BHL, Ingenhoes R, Peeters LL and Exalto N (1988) Early pregnancy
wastage: relation between chorionic vascularization and embryonic development. Fertil Steril 49,216–220.
Minguillon C, Eiben B, Bähr-Porsch, Vogel M and Hansmann I (1989) The
predictive value of chorionic villus histology for identifying chromosomally
normal and abnormal spontaneous abortions. Hum Genet 82,373–376.
Nelen WLDM, Bulten J, Steegers EAP, Blom HJ, Hanselaar AGJM and Eskes
TKAB (2000) Maternal homocysteine and chorionic vascularization in
recurrent early pregnancy loss. Hum Reprod 15,954–960.
Quenby S, Farquharson R, Vince G and Aplin J (2002) Recurrent miscarriage:
a defect of Nature’s quality control? Hum Reprod 17,534–538.
Roberts L, Sebire D, Fowler D and Nicolaides H (2000) Histomorphological
features of chorionic villi at 10–14 weeks of gestation in trisomic and chromosomally normal pregnancies. Placenta 21,678–683.
te Velde EA, Exalto N, Hesseling P and van der Linden HC (1997) First trimester development of human chorionic villous vascularization studied with
CD34 immunohistochemistry. Hum Reprod 7,1577–1581.
Submitted on August 15, 2005; resubmitted on October 31, 2005; accepted on
November 21, 2005