A mathematical model for the estimation of human embryonic and fetal age

A mathematical model for the estimation of human embryonic and fetal age

Cell Transplantation,Vol. 5, No. 4, pp. 453464, 1996 Copyright 0 1996Elsevier Science Inc. Printed in the USA. All rights reserved 0963.6897/96$15.00 ...

1MB Sizes 0 Downloads 6 Views

Cell Transplantation,Vol. 5, No. 4, pp. 453464, 1996 Copyright 0 1996Elsevier Science Inc. Printed in the USA. All rights reserved 0963.6897/96$15.00 + .OO ELSEVIER

PI1 SO963-6897(96)00079-6

Original Contribution

A MATHEMATICAL

L.

MODEL FOR THE ESTIMATION FETAL AGE

EVTOUCHENKO,*

L. STUDER,*

C. SPENGER,“’

*Department of Neurosurgery, University of Bern, Inselspital,

University

AND R.W.

Switzerland,

AND

SEILER*

and TDepartment

of Gynecology.

critical for various transplantation approaches as specific time windows exist for successful grafting. For example, behavioral recovery in an animal model of Parkinson’s disease can be achieved by the transplantation of dissociated human fetal mesencephalic tissue between 6 and 8 wk postconception, whereas older tissue gives poor results (6,15,16). Gynecologists traditionally determine embryonic age by calculating the number of weeks after the last menstruation though the duration of menstruation can substantially vary from woman to woman. By simply subtracting 2 wk from the anamnestically estimated menstrual age inaccuracies of up to 2 wk in the determination of postconceptional age have to be expected (5). In addition, embryonic age is determined by ultrasonographic measurements of crown-rump length (CRL), biparietal diameter (BPD), or the diameter of the chorionic sack. In a clinical setting, it has been difficult to achieve a high precision in the determination of embryonic age with bimanual investigation (3) as well as with radiological techniques (8,34), ultrasonographic measurements of BPD (7), or CRL (32), or the determination of kreatinine concentration in the amniotic fluid (14). In the present study, a set of morphological parameters such as greatest length, neck-rump length, proximal and distal arm length, proximal and distal leg length, and foot length of human embryos and fetuses was measured following medically induced suction abortions. The data were correlated to the anamnestic and the ultrasonographically determined age. This allowed to establish a mathematical model for age estimation between 4 and 12 wk postconception. Age determination based on a combination of parameters reduced the effects of single inaccurate measurements, individually different growth pattern among different embryos, and the general statistic variation.

model; Hu-

INTRODUCTION

The use of embryonic and fetal tissue for the treatment of neurodegenerative diseases (20), diabetes mellitus (12), chronic pain (38), and other clinical disorders [for review (13)] evoked a new interest in the description of human embryonic development. Reliable age determination is of utmost importance for developmental studies as well as for clinically oriented research. Optimal donor age is

Department of Neurosurgery, University of Bern, Inselspital, Freiburgstrasse, CH-3010 Bern, Switzerland.

ACCEPTED l/20/96.

‘To whom correspondence

E. DREHER,?

Freiburgstrasse, Bern, CH-3010, of Bern, Bern, CH-3010, Switzerland

0 Abstract -Precise determination of donor age in human embryonic and fetal tissue is crucial for cell transplantation due to the existence of distinct time windows within which successful grafting is possible. This study demonstrates that between 4-12 wk postconception embryonic and fetal age can be estimated based on various morphometric parameters measured on a routine basis in suction abortion material. The greatest length, the neck-rump length, the foot length, and the proximal and distal arm and leg length were correlated with the anamnestic and ultrasonographically estimated age. Multivariate regression analyses showed a linear correlation between age and the logarithmic value of the various morphometric parameters. The best correlation was found for a mathematical model combining the limb parameters (r = 0.904; p < 0.001; n = 37). A prospective follow-up study (n = 40) was carried out to test the validity of the mathematical model. A high correlation was found between the calculated age and the estimated age based on anamnestic data (r = 0.749,~ c 0.001). Outliers due to errors in the anamnestic data were readily identified by comparing anamnestic with calculated age. This method allows determination of embryonic and fetal age within and beyond the age group of the Carnegie classification and may, therefore, be useful for the needs of experimental and clinical cell transplantation. 0 Keywords - Embryonic age; Mathematical man embryo; Neural transplantation.

OF HUMAN EMBRYONIC

should be addressed. C. Spenger, 453

Cell Transplantation 0 Volume 5, Number 4, 1996

4.54 MATERIALS

AND METHODS

Embryonic and fetal tissue was obtained from 185 medically induced suction abortions. The study was approved by the Ethics Committee of the Medical Faculty of the University of Bern. Written informed consent was received from the abortion-seeking woman. Pathological pregnancies, spontaneous abortion, and embryos from women with diabetes, preeclampsia, or drug abuse were excluded from the analysis. The embryonic or fetal age ranged from 4 to 12 wk pc. The age was determined by the date of last menstruation in all cases and by ultrasonographic scanning in 91 of 185 cases. In the cases with ultrasonographic scanning, the medical history was checked for irregular menstruation cycles, special occurrences in the medical history of the family, and the social background of the women.

Abortion Procedure and Setup for Tissue Explantation Following portio dilatation, sterile suction abortions were performed at pressures between 0.3 and 0.6 bar using a 8-13 mm vacuum curette, depending on the anamnestic age of the embryo. The abortion material was collected via a tube with an inner diameter of 12 mm into a sterile specimen collector sock (Abbott E10.5) situated in a plastic canister (Abbott E123) connected to a vacuum pump. Before and after abortion the system was flushed with 50 mL of sterile 0.9% NaCl solution. The abortion material was transferred into a laminar flow hood and the collector sock was placed into a sterile 150 x 20 mm plastic Petri dish (Sarstedt No. 831803) and cut open. Embryonic tissue was carefully recovered and rinsed with Geys balanced salts solution (GBSS) (Gibco, No. 041-04260M). The time needed for recovery of the tissue, measurements and identification of CNS fragments was 40-90 min. For the measurements the fetal tissues were placed into an extra Petri dish filled with GBSS and placed onto the light table of a binocular microscope (Wild/Leitz M 32). A scaled, transparent, and thin paper was placed between the microscope table and the Petri dish such that by viewing the specimen through the microscope the size could be directly read at magnifications 6.5-40x.

Definition of Morphometric Parameters The following morphological structures were quantified as schematically illustrated in Fig. 1. The greatest length (GL) of embryos or fetuses was measured according to O’Rahilly and Mtiller (28) in the naturally curved posture without inclusion of the limbs. The neck-rump

C

PLL -

! FL Fig. 1. Schematic illustration of the morphometric parameters used in the present study. (A) Definitions of greatest length (GL) and neck-rump length (NRL). (B) Definitions of proximal (PAL) and distal (DAL) arm length. The length of the hand is included in the DAL parameter. (C) Definitions of proximal (PLL) and distal (DLL) leg length and foot length (FL).

length (NRL) was defined as the distance between the neck flexure and the most caudal point of the embryo. The proximal arm length (PAL) was defined as the distance between midpoint of the curvature formed by the anlage of the deltoid muscle and the distal ending of the upper arm in a 90” flexion of upper and lower arm. The distal arm length (DAL) was defined as distance between the proximal ending of the lower arm in a 90” flexion of upper and lower arm and the tip of the third finger or the most distal part of the anlage of the hand, when fingers had not been developed. The proximal leg length (PLL) was defined as the distance between midpoint of the curvature formed by the anlage of the gluteus maximus muscle and the distal ending of the upper leg in a 90” flexion of upper and lower leg. The distal leg length (DLL) was defined as the distance between proximal ending of the lower leg in a 90” flexion of upper and lower leg and the distal ending of the lower leg in a 90” flexion of lower leg and foot. The foot length (FL) was defined as the distance between rear end of heel and distal ending of the foot given by the tip of the second toe. The resolution of this method is 0.5 mm.

Age determination

in human embryos

Ultrasonography Ultrasonography was carried out using an ACUSON 128 XP/lO scanner. Embryonic or fetal age was estimated by the morphometric tables of Hansmann (17). For embryos younger than 8 wk p.c. the diameter of the chorionic sack was measured. For older fetuses CRL and BPD were determined and used for age calculation.

Statistical Analysis The parameters measured during explantation were correlated with the anamnestic age and the ultrasonographically estimated age (if available). Data points were fitted by an exponential function of the type y = abx for all morphological parameters using a commercially available statistical package (Systat 5.0, Systat Inc., Evanston, IL). After logarithmic transformation parameters were correlated to the embryonic age in univariate and multivariate linear regression models and the Pearson correlation factors were calculated for all possible combinations of parameters to find an optimal model for age prediction. The mathematical models were based on three different sets of data, namely, on anamnestic and on uncorrected and corrected ultrasonographic data. For comparisons between different models r-values were calculated by formula (1) and compared to the critical value t* of the Student’s distribution at 1 - q’“.975’. la, - a21 (1) t=dm Variables a, and a2 in formula (1) represent the regression coefficients for model 1 and 2, respectively, and r, and r2 the standard errors of the respective regression coefficients.

Comparison with Data of Other Authors A comparative analysis was carried out between our data and those of other authors (4,36). The following parameters derived from these publications could be included in the comparison: Bossy and Katz (4): CRL, FL, PAL, DAL, PLL, DLL (n = 30 for all parameters; inclusion criterion: age < 94 days p.c.); Streeter (36): CRL (n = 38) and FL (n = 59) (inclusion criterion: age < 94 days p.c.>. Prospective Study to Test the Validity of the Mathematical Model for Age Prediction Embryonic or fetal age was determined in 40 additional samples with an anamnestic age of 38 to 77 days p.c. Age was calculated for each sample with the mathematical model established. This calculated age was then correlated to the anamnestic age.

0 L. EVTOUCHENKO ET AL.

455

RESULTS

Identification of Embryonic and Fetal Structures in the Abortion Material Embryos with an anamnestic age between 4 and 12 weeks pc. were analyzed (n = 185). In a first step completeness of the identified abortion material was checked. The histograms in Fig. 2 illustrate identification rates for various anatomical structures of the embryo. Frequency histograms for measured morphometric parameters are given in Fig. 3. GL and NRL could only be determined at low frequency. However, GL or CRL are historically the most important parameters. Most classical embryological studies included either of these parameters in the analyses [(23,36); for review (29)]. Parameters of upper and lower limbs could be quantified at relatively high frequencies (Fig. 3C-G). For example, the PAL could be measured in 43-75% of all cases in the age range of 6-8 wk p.c. The rate at which the other embryonic and fetal parts were identified in the abortion material and could be measured were 52-75% for the DAL, 33-66% for the PLL, 38-70% for the DLL, and 55-65% for the FL within the same age range. Due to the limited availability of data of one single parameter, age determination had to be based on different parameters either alone or in combination.

Correlation of Morphometric Parameters with Anamnestic Age and Ultrasonographically Determined Age In a first step, anamnestic age and morphometric parameters were correlated (n = 185). In a second step, the age estimated by ultrasonographic scanning was correlated to the morphometric parameters (n = 81). In order to exclude cases in which the age determination was uncertain, 11 embryos with a difference between anamnestic and ultrasonographic age of more than 2 SDS (SD = +7) and two embryos derived from twin pregnancies were excluded. This led to the so called corrected data pool of ultrasonographically estimated ages (n = 68). Interestingly, no statistically significant difference between the anamnestic age and the age determined by ultrasound in the 81 cases was detected despite individual differences of up to 19 days (mean of anamnesticultrasound defined age = -1.0 + 7.3 days; n = 81). No anamnestic data based on the registration of the basal body temperature were available. Determination of GL could only be carried out in 14 cases. In spite of this small number of cases, data points closely follow the fitted exponential curve with little variation in the age interval between 30-78 days p.c. (Fig. 4A). A clustering of data points was shown to occur for NRL in the age range between 45-70 days p.c. (Fig.

Cell Transplantation 0 Volume 5, Number 4, 1996

456 A

No

embryonic

parts

B

Viscerocranlum

ldentlfled

C

Spinal column ldentltled

could be ldentlfled so,

)

1 I

I 7 ,

WY

Y-----l

age [Weaka PO.]

0

E

Thorax ldentlfled

30

Pelvls ldentlfled

30

z

2 20

20

10

10

0

0

agew--b

PP.1

.(la

0

The total number of examlned embtyos and fetuses

m

embryonic parts could bs identiffed

[weeks pa.]

Fig. 2. Frequency histograms of the number of samples and the identification rates for various anatomical structures depending on embryonic and fetal age. Abortions were carried out by a low-pressure aspiration technique. Frequency rates of the following structures are given: (A) No embryonic parts present. (B) Identification rate of viscerocranium. (C) Identification rate of spinal column. (D) Identification rate of thorax. (E) Identification rate of pelvis.

4B). The high variability of the data at a given age level reflects problems in measuring this parameter in a consistent manner. This was due to problems in the identification of the level of the neck flexure as well as to differences in body flexion and in the extent of traumatically induced changes among individual samples. Comparison between graphs for CRL and NRL revealed a threefold increase in CRL for the age between 6 and 10 weeks p.c., but a fivefold increase for the NRL during the same time period. Parameters of the limbs namely PAL, DAL, PLL, DLL, FL could be quantified in the majority of cases (Fig. 4C-G). Because the fitted curve could be approximated with an exponential function of the type y = ab”, a logarithmic transformation of the measured values was carried out. After logarithmic transformation parameters showed a linear correlation with fetal and embryonic age and the Pearson correlation coefficients could be individually calculated and compared among all parameters and all modes of age determination (Table 1). This revealed first, that ultrasonographic age led to higher Pearson coefficients; second, that corrected ultrasonographic values

were generally better correlated than noncorrected values; and third, that curves relying on ultrasound determined age had a slightly increased slope as compared to data based on anamnestic age. By the correction of the ultrasonographic data all outliers could be eliminated. However, there was no statistically significant difference between linear regression models derived from the corrected or uncorrected data pool. The fitted curves for the corrected data of all limb parameters except PLL were between those for anamnestic and those for uncorrected ultrasonographic age (Fig. 4C-G). Age determination based on ultrasonographic data compared to that based on anamnestic data led to relatively lower ages for most parameters (except GL and PLL) when the embryo was younger than 60 days p.c. but to higher ages for embryos older than 60 days p.c. Combination of Morphometric Parameters Combining two logarithmically transformed parameters further improved the correlation between the embryonic or fetal age and the morphometric parameters (Table 2). In an additional step, systematic multivariate

Age determination

in human embryos

L. EVTOUCHENKO

0

ET AL.

457

Table 1. Correlation between each morphometric parameter and embryonic age as determined by anamnestic information or by ultrasonographic measurements Parameter Anamnestic Age Uncorrected

Regression Coefficient

Intercept

n

r

p-value

LN GL LN NRL LN PAL LN DAL LN PLL LN DLL LNFL Ultrasound Determined Age LN GL Uncorrected LN NRL LN PLL LN DAL LN PLL LN PLL LNFL Corrected LN GL LN NRL LN PAL LN DAL LN PLL LN DLL LNFL

= = = = = ZZ =

0.036X 0.019x 0.026X 0.028X 0.029x 0.027X 0.021x

+ 1.178 + 2.011 + 0.471 + 0.465 + 0.261 + 0.306 + 0.383

1.5 60 108 110 101 101 99

0.890 0.558 0.599 0.646 0.692 0.648 0.613

10.001
= Z = ZZ = = = ZZ = = = = = =

0.051x 0.025X 0.040x 0.044x 0.038X 0.042X 0.032X 0.051x 0.029x 0.045x 0.048X 0.049x 0.043x 0.036X

+ 0.326 + 1.724 - 0.258 - 0.438 - 0.384 - 0.543 - 0.272 + 0.326 + 1.532 - 0.566 - 0.668 - 0.997 - 0.630 - 0.489

5 26 52 54 48 49 47 5 21 43 45 39 40 38

0.994 0.663 0.827 0.854 0.856 0.829 0.827 0.994 0.746 0.866 0.874 0.882 0.886 0.871


A linear relationship

transformation

of the parameters.

is given after logarithmic

regression analysis with more than two parameters was undertaken to find optimal combinations of parameters for age prediction. Because not all parameters can be measured in all embryos and because effects of randomly occurring imprecise measurements can be reduced by averaging data from various parameters, an increased validity and reliability of age prediction can be expected in multivariate models. The multivariate models were based on the corrected ultrasonographic data only, as these had shown the highest correlation coefficients in univariate models and their residuals matched best criteria of normal distribution. All possible combinations of two, three, four, and five parameters are listed in Table 3. Combinations of limb parameters with GL or NRL are

n = number of samples; r = correlation

coefficient.

omitted due to the small sample number for GL in the ultrasonographic data set (n = 5) and because the combination limb parameter-NRL in bivariate models did not improve correlation coefficients. Combination of three or more limbs parameters displayed very strong correlations (r 0.888 to 0.904). Validation of the Data Data were validated by a scholarly comparison with historical data published by Streeter in 1920 and by Bossy and Katz in 1964 and by carrying out a prospective study. The data of Streeter (36) and of Bossy and Katz (4) were suitable for comparison as similar, though not identical, definitions of the investigated parameters

Table 2. Correlation coefficients r as determined by multiple linear regression of the corrected ultrasonographic age and all possible combinations of two different morphometric parameters

LN LN LN LN LN LN LN

GL NRL PAL DAL PLL DLL FL

LN GL

n

LNNRL

it

LN PAL

n

LNDAL

n

LN PLL

rz

LN DLL

n

LNFL

n

0.984NS 0.977NS 0.978NS 0.994NS 0.999NS 1.000*

0.984NS 0.756$ 0.767? 0.8284 0.832$ 0.803$

4

4 4 4 4 4 4

0.977Ns 0.756$ 0.8883 0.8874 0.8961: 0.884$

4 21

0.978Ns 0.767$ 0.888j 0.8984 0.897$ 0.877$

4 20 42

0.994NS 0.828$ 0.8871 0.898$

4 20 38 38

4 20 38 39 39

0.895$ 0.892$

39 37

0.999NS 0.8324 0.896+ 0.8971: 0.895$ _ 0.897$

1.000* 0.803f 0.8844 0.8771: 0.892$ 0.897$ -

4 20 37 38 37 38

21 20 20 20 20

*p < 0.05; tp < 0.01; & c 0.001: NS = not significant.

42 38 38 37

36 39 38

38

Cell Transplantation 0 Volume 5, Number 4, 1996

458

A

Greatest length measured

.ge [weeks

/

pa.]

Distal arm length measured

age [We&s

G

Neck-rump length measured

M,

w,........,

D

B

*

I

1 I ,

C Proximal

So”

aF [weelo P.-l

E

Proximal leg length measured

PO.1

am t-b

arm length measured

age b-9~

F

pa.1

Distal leg length measured

pa.1

Foot length measured

1

The total number of examined embyos and fetuses ldentlficattonrates of measured parameters of embryonicbe+ and limbs

Fig. 3. Frequency histograms of the number of samples and the identification rates for the various morphometric parameters quantitatively assessed depending on embryonic and fetal age. Frequency rates of the following parameters are given: (A) Greatest length. (B) Neck-rump length. (C) Proximal arm length. (D) Distal arm length. (E) Proximal leg length. (F) Distal leg length. (G) Foot length.

were used, data of the same age group were presented and the extraction of the raw data from the original publication was feasible, allowing further computer-based statistical analyses. An overlay of the data points derived from the original publications with those given in the present study is shown on Fig. 5. Data of the different publications were computer fitted by exponential models. The data extracted from Streeter’s publication displayed a large variance for a given age, whereas the variance in Bossy and Katz’s data was minimal. The variance of our data points was somewhat between that of the two historical studies. Age prediction relying on Streeter’s data for CRL or FL would lead to significantly lower ages (in average about minus 15 days) than age

prediction based on data presented in this study. Better correspondence of our data was found for CRL and FL with the study of Bossy and Katz leading to differences of O-8 days for CRL and O-4 days for FL within the age of interest. Substantial differences at given ages between our data and those of Bossy and Katz could be detected for PAL and DLL and to a lesser extent for PLL (Fig. 5D-F). The parallel course of the two curves suggests that a systematic difference exists and may reflect the fact that the definitions of PAL, DLL, and PLL used by Bossy and Katz were not exactly identical to those used here for practical reasons. In order to confirm the predictive power of our mathematical model, a prospective study measuring different

Age

.

.

w-

459

.’

4/-.

.

?? ;

40.

..

.

/

.

..

20.

0

0’ 10

203040506070

greatest

C

L. EVTOUCHENKO ET AL.

.

SO.

*

determination in human embryos 0

Ial

0

10

length [mm]

20

a0

40

so

50

neck-rump length [mm]

D

r

loo

1

w60.

40

20

Oi 5

0

10

15

20

.. .c...*

25

5

proximal leg length [mm]

E

100

i

I.

..

.

10

15

20

distal leg length [mm]

.

.

I

.

.

. .

.

tern

a: :

*b’(: *. ??

.

??

a

0

I 0

0 ’ 10

5

15

20

25

5

0

proximal arm length [mm]

G

10

15

20

25

distal arm length [mm]

100 1

20

0

-

0

-

??0

-

-

- - -

anamnestic

age

uncorrected

ultrasonographic

corrected

ultrasonographic

age age

I 0

5

ID

15

foot length [mm]

Fig. 4. Comparison of the three methods used for embryonic and fetal age between corrected and uncorrected ultrasonographic data and between ultrasonographic measurements for the various morphometric parameters. The following parameters are given: (A) greatest length. (B) Neck-rump Proximal arm length. (F) Distal arm length. (G) Foot length.

determination. No significant differences could be detected data based on anamnestic information and that based on Data are fitted by distance weighted least square method. length. (C) Proximal leg length. (D) Distal leg length. (E)

460

Cell Transplantation 0 Volume 5, Number 4, 1996 Table 3. Correlation coefficients I as determined by multiple linear regression of the corrected ultrasonographic age and all possible combinations of three, four, and five limb parameters Parameter LN LN LN LN LN LN LN LN LN

PAL/LN DAL/LN PLL PAL/LN DAL/LN DLL PAL/LN DAL/LN FL PAL/LN PLL/LN DLL PAL/LN PLL/LN FL PAL/LN DLL/LN FL DAL/LN PLL/LN DLL DAL/LN PLL/LN FL PLL/LN DLL/LN FL

LN LN LN LN

PAL/LN PAL/LN PAL/LN DAL/LN

DAL/LN PLL/LN DAL/LN PLL/LN PLL/LN DLL/LN PLL/LN DLL/LN

DLL FL FL FL

LN PAL/LN DAL/LN PLL/LN DLL/LN FL

parameters in 40 additional embryos and fetuses was performed. Based on the linear correlation between the logarithmic values of the limb and the corrected ultrasonographic ages, as shown in Table 1, a general formula combining these five parameters was produced [formula (2)] where LN is the logarithm to the base e of the measured values (VAL) in [mm], a is the intercept, and b the regression coefficient of the curve and n is the number of parameters measured. n LN(VAL,) - aj ci=l

b.I

Age[days (p.c.>] =

LN(DAL) + 0.668 +

0.045 LN(PLL) + 0.997 0.049

0.048

+

LN(DLL) + 0.630 +

0.043

+

LN(FL) + 0.489 0.036

38 38 37 38 37 37 38 37 37

0.899 0.900 0.888 0.899 0.894 0.901 0.904 0.899 0.900


38 37 37 37

0.904 0.899 0.901 0.904


37

0.904


p-Value

In the 40 additional cases of the prospective study as many parameters as possible were measured and the age calculated using formula (2). Thereafter, the calculated age of the different embryos and fetuses was correlated to the anamnestic age (Fig. 6). Data points were fitted by a linear regression model. Although a substantial variance was detected (Pearson coefficient of correlation = 0.75), the regression coefficient that was close to 1 and the intercept at -3.8 days point to a high overall correspondence between the calculated and the anamnestic age. No data about the ultrasonographic age were available for these 40 embryos and fetuses. DISCUSSION

=

LN (PAL) + 0.566

r

(2)

n

For example if the parameters PAL, DAL, PLL, DLL, or FL can be measured the formula will be as follows: Age[days(p.c.)]

n

(3)

5 If one or several parameters cannot be measured, the respective component will be removed from the formula and n reduced accordingly.

For more than 100 years researchers have been trying to quantify growth of human embryos and fetuses. The classic, widely accepted approach divides embryonic development between 0 and 56 days p.c. into 23 Carnegie stages according to distinct features of embryonic development (27,29). The practical use of this staging system for age determination in fetal transplantation research, however, is limited due to the fact that first, age range of interest in experimental neural transplantation includes ages clearly above 56 days p.c.; second, material collected after suction abortions is heavily damaged; and third, an easy, but highly efficient routine procedure is warranted in order to minimize explantation time. Therefore, we chose an alternative approach, namely the standardized quantification of anatomical structures, which can be readily identified between 6-10 weeks p.c., the age of interest in neural transplantation. The parameters investigated include the body size parameters GL and NRL and the limb parameters PAL, DAL, PLL, DLL,

Age determination in human embryos 0 L.

EVTOUCHENKO ET AL.

foot length [mm]

crown-rump length or greatest length respectively [mm] __

_

.

present study

--

0

Bossy and Katz ‘64

. _

*

Streeter’20

distal arm length [mm] __

E

.

present study

0

Bossy and Katz ‘64

461

. 0

present study Bossy and Katz ‘64 Z_ Streeter ‘20

proximal arm length [mm] --

I

present study

0

Bossy and Katz ‘64

loor

distal leg length [mm] _-

w present study 0

Bossy and Katz ‘64

‘it

,

0

5

,

,

,

g_; *

10 15 20 proximal leg length [mm] __

??

0

present study Bossy and Katz ‘64

Comparison of morphometric data (corrected ultrasonographic age) with that derived from the studies by Streeter (36) (A, B) and by Bossy and Katz (4) (A-F). The estimated Carnegie stage, based on the relationship between embryonic age and Carnegie stage by O’Rahilly and Miiller (29), is illustrated as a second y-axis. The following parameters are given: (A) greatest length. (B) Foot length. (C) Distal arm length. (D) Proximal arm length. (E) Distal leg length. (F) Proximal leg length.

-lad LIaupnol pazypnpuejs 103 palys II~M SF alnpaDold luasald aql ~eq~ lsa8%ns aM ‘snqJ *amyn3 aql UTuo!s!DaId a~ asEa.wy laqUn3 01 pue anbruyal syl augaJ layun3 01 MOI@ lcsu~ Iood ~lep %qst?alXQ AI!pvals B ‘.IaAaMoH .Iapow I~~~lsuraqler_u aql30 Bupsal aAy3adsold I? Aq pue sarpnis layea 30 ivyi qlt~ alvp mo %uFduro:, dq paDuap -[Aa SB Iapour Iecgvtuaqlsur e Bursn as!DaJd PUB aIq!s -sod s! uoy?uyJaiap a% l&y1 ~oqs Apnis sgl30 s1Insa.I ayL ..yn3yvd uy uoywnqdsuwl Ieluawyladxa uy masaId suopypuo3 aqi Kul.rap!suo3 w2p 30 ias tiau e awaua%l 01 paluv~m set l! suoseal asayi JO~ .papnIDxa aq lou -UK) suogwalp2 az!s aIqelap!suo3 pue ‘(OE‘6‘Z) SUOTlEXI -qnd sno!.wA u! LIIyslaAonuoD passncwp uaaq wq slala -tue.wd q@uaI uo uoy2xy 30 aDuanau1 ~1~Iwu.1030boI u! pax!3 11”alaM q3~r.1~ ‘uoyaIIo3 ay8au.y) aql30 saIduws

uo Pasvq sQM (LZ) JaII!lIN Pue hlW.I
-ma uI pau%!sse dIlua.ww lvyl uwy la&nod sLap j7-I aq 01 wvaddv a% pauywalap dIIwgdw8ouoswlIn aqi lvql paq!_wap wavourisod pamwaw wql qiy LIIw!qdt& -ouossqIn pamwaw sohrqwa 30 1x3 aql %ul.wduroD dIIgvx,o pun uuutua dq dpnls e ‘aIdwvxa JOA .snla3 JO ohrqura aqi 30 a% Ir?al aql lnoqe luatuaiels Iaug v MOI -1~ wp 3~qdeBouoswlIn JOU 3gsautusuE laql!aN ‘a3ua -la3al30 lu!od e 30 XXI aql SF sa%qs IquauxdoIaAap ih.y -leIa*Io3 sa!pnls Iw$oIohquIa 30 uIaIqold Iwaua2 v ‘(s) anbluqDa$ uoyoqe pap@ punoswIn UE %uruIlojlad dq paAayqN aq UBD1uauraAo~dw~ laqvng ‘Icy -almu aql30 dl@alu! aql aAo_Idw! ~I]ue~@?rs UE?D(pZ) anb!uqDal uogw~dw alnssa.Id-MOI ‘.IaAaMoH ‘sast23 aql30 (s) go6 ueq] a.IouI us .nw3o 01 papadxa s! Ir?!.Ialer_u30 a% -w~p Ivywlsqns ‘suoyoqr! uo!lDns augnol UI .pal!nbaJ s! JoleQsaAuy aql 30 aDuayadxa a%.wI B pue pagluapi aq 01 ipmgyp a.n2 uuu I I-8 ueq1 JaIIEurs 733 B 30 sohrq -uIa ‘(62) .IaIInAI puk? dIIye&o 01 %u~p~o~v wawdoIaA -ap 30 aBeis aql uo puadap AI%o.w samptuls Iela3 pw cyuohqwa 30 uoygguap! In3ssaDDns 30 huanbald w2p ~~~lauroqdrour UIOJJ paAnap a.n2 ieqi sIapotu Itgiwuaqleur uo Bu+IaJ aIqIsea3 s! ai% Iala pue D!uohrqura 30 uoyDrpaJd alveJn33r!ue Ivyi ‘p.uql pur! fsasw aql 30 lsoux UT pawopad ICIaugnoJ aq uw sa.uwn_us asaql 30 IwaAas 30 uogwgguenb lay3 ‘puoDas fkuanb -ay q&q t? qi!M uoploqe uopDns amssa3d-moI .rai3e paypuapr aq uv3 samptuls Ieia3 put2 cyuohrqtua snopA lt?qi ‘islg SMoqs Lpnis luasald au ~uoy?u~uuaiap a% u! d%Itw~c aAo.Idw! 01 pasn aJaM slalawwed 30 SUO~EU -FqtuoD ‘LIuo slalaruend aIk?uFs %u!sn 30 pvalsu1 ‘m pup

9661 ‘p laqmnN‘s au.mloA??uoymydsue~~

[[aa

Z9P

Age determination

in human embryos

formed age determination in experimental and clinical transplantation of human embryonic and fetal tissue. Acknowledgments - This research was supported by Swiss National Science Foundation (Grant No. 31-36243.92), by the Swiss federal office of education and science (Grant BBW 93.0349) and the Swiss Parkinson Society. We would like to thank Dr. A. Dityatev (University of Bern) for his helpful advise concerning the statistical analysis of our data.

REFERENCES

1. Barr, M.; Blackburn, W.R.; Reede Cooley, N. Human fetal somatic and visceral morphometrics. Teratology 49:487496; 1994. 2. Bartelmez, G.W.; Dekaban, AS. The early development of the human brain. Contr. Embryol. Carnegie Inst. 37: 13-32; 1962. 3. Beazley, J.M.: Underhill, R.A. Fallacy of the fundal height. Br. Med. J. 4:404-406; 1970. 4. Bossy, J.; Katz, J.M. Croissance segmentaire d’une serie de 100 foetus humains. Ann. Desarrollo 12:181-213; 1964. 5. Brundin, P. Grafting human embryonic brain tissue. In: Dunnett, S.B.; Bjorklund, A., eds. Neural transplantation. A practical approach. New York: IRL Press; 1992:139160. 6. Brundin, P.; Nilsson, O.G.; Strecker, R.E.; Lindvall, 0.; Astedt, B.; Bjorklund, A. Behavioral effects of human fetal dopamine neurons grafted in a rat model of Parkinson’s disease. Exp. Brain Res. 65:235-240; 1986. 7. Campbell, S. The prediction of fetal maturity by ultrasonic measurement of the biparietal diameter. J. Obstet. Gynaecol. Br. Commonwealth 76(7):603-609; 1969. 8. Dee, P.M.; Parkin, J.M.; Simpson, W. A study of some radiological criteria used in assessing the gestational age of the human infant. Clin. Radiol. 47:125-131; 1966. 9. Drumm, J.E.; O’Rahilly, R. The assessment of prenatal age from the crown-rump length determined ultrasonically. Am. J. Anat. 148(4):555-560; 1977. 10. Drumm, J.E. The effect of the maternal environment on the fetal crown-rump length. Recent advances in ultrasound diagnosis. International Congress Series No. 436. Kurjak, A., ed. Amsterdam: Excerpta Media; 1977:155-159. 11. Dylevsky, I. Growth of the human embryonic hand. Acta Universitatis Carolinae. Medica monographia CXIV; 1986. 12. Federlin, K.; Bretzel, R.G.; Hering, B.J. Recent achievements in experimental and clinical islet transplantation. Diabetes Med. 8(1):5-12; 1991. 13. Fine, A. Human fetal tissue research: Practice, prospects, and policy. Cell Transplant. 3:113-145; 1994. 14. Foulds, J.W.; Pennock, C.A. Amniotic fluid creatinine: An unreliable index of fetal maturity. J. Obstet Gynaecol. Br. Commonwealth 79(10):911-915; 1972. 15. Freeman, T.B.; Kordower, J.R. Human cadaver embryonic substantia nigra grafts: Effects of ontogeny, pre-operative graft preparation and tissue storage. In: Lindvall, 0.;

0 L.

EVTOUCHENKO

ET AL.

463

Bjorklund, A.; Widner, H., eds. Intracerebral transplantation in movement disorders. New York: Elsevier Science Publishers; 1991:163-169. 16. Freeman, T.B.; Spence, M.S.; Boss, B.D.; Spector, D.H.; Strecker, R.E.; Olanow, C.W.; Kordower, J.H. Development of dopaminergic neurons in the human substantia nigra. Exp. Neurol. 113(3):344-353; 1991. 17. Hansmann, M.; Hackeloer, B.-J.; Staudach, A. Ultraschalldiagnostik in Geburtshilfe und Gynakologie: Lehrbuch und Atlas. Berlin: Springer; 1985:37-57. 18. Hem, W.M. Correlation of fetal age and measurements between 10 and 26 weeks of gestation. Obstet. Gynecol. 63(1):26-32; 1984. 19. Jirasek, J.E. Atlas of human prenatal morphogenesis. ton: Martinus Nijuhoff Publishers; 1983.

Bos-

20. Lindvall, 0.; Odin, P. Clinical application of cell transplantation and neurotrophic factors in CNS disorders. Curr. Opin. Neurobiol. 4:752-757; 1994. 21. Mercer, B.M.; Sklar, S.; Shariatmadar, A.; Gillieson, M.S.; D Alton, M.E. Fetal foot length as a predictor of gestational age. Am. J. Obstet. Gynecol. 156(2):350-355; 1987. 22. Mhaskar, R.; Agarwal, N.; Takkar, D.; Buckshee, K.; Anandalakshmi; Deorari, A. Fetal foot length-A new parameter for assessment of gestational age. Int. J. Obstet. Gynaecol. 29( 1):35-38; 1989. 23. Moore, G.W.; Hutchins, G.M.; O’Rahilly, R. The estimated age of staged human embryos and early fetuses. Am. J. Obstet. Gynecol. 139(5):500-506; 1981. aspiration 24. Nauert, G.M.; Freeman, T.B. Low-pressure abortion for obtaining embryonic and early gestational fetal tissue for research purposes. Cell Transplant. 3(2): 147151; 1994. 25. Nishimura, H.; Takano, K.; Tanimura, T.; Yasuda, M. Normal and abnormal development of human embryos: First report of the analysis of 1,213 intact embryos. Teratology 1(3):281-290; 1968. 26. Nishimura, H.; Tanimura, T.; Semba, R.; Uwabe, C. Normal development of early human embryos: Observation of 90 specimens at Carnegie stages 7 to 13. Teratology 10( 1): l-5; 1974. 27. O’Rahilly, R. Developmental stages in human embryos, including a survey of the Carnegie collection. Carnegie Institution of Washington 631; 1973. 28. O’Rahilly, R.: Mtiller, F. Embryonic length and cerebral landmarks in staged human embryos. Anat. Rec. 209(2): 265-27 1; 1984. 29. O’Rahilly, R.; Mtiller F. Developmental stages in human embryos. Including a revision of Streeter’s “Horizons” and a survey of the Carnegie Collection. Carnegie Institution of Washington 637: l-306; 1987. 30. Patten, B.M.; Philpott, R. The shrinkage of embryos in the processes preparatory to sectioning. Anat. Rec. 20:393413; 1921. 31. Pedersen, J.F. Fetal crown-rump length measurement by ultrasound in normal pregnancy. Br. J. Obstet. Gynaecol. 89( 11):926-930; 1982.

464

Cell Transplantation 0 Volume 5, Number 4, 1996

32. Robinson, H.P.; Fleming, J.E.E. A critical evaluation of sonar “crown-rump length” measurements. Br. J. Obstet. Gynaecol. 82:702-710; 1975. 33. Scammon, R.E.; Calkins, L.A. The development and growth of the external dimensions of the human body in the fetal period. Minneapolis: University of Minnesota Press; 1929. 34. Scott, K.E.; Usher, R. Epiphyseal development in fetal malnutrition syndrome. N. Engl. J. Med. 270822-824; 1964. 35. Skidmore, F.D. An analysis of the age and size of 483 human embryos. Teratology 15(1):97-102; 1976.

36. Streeter, G. Weight, sitting height, head size, foot length, and menstrual age of the human embryo. Contributions to Embryology No. 55. Carnegie Institution of Washington; 1920:143-170. 37. Warda, A.H.; Deter, R.L.; Rossavik, I.K.; Carpenter, R.J.; Hadlock, F.P. Fetal femur length: A critical reevaluation of the relationship to menstrual age. Obstet. Gynecol. 66(l): 69-75; 1985. 38. Winnie, A.P.; Pappas, G.D.; Das Gupta, T.K.; Wang, H.; Ortega, J.D.; Sagen, J. Subarachnoid adrenal medullary transplants for terminal cancer pain. A report of preliminary studies. Anesthesiology 79(4):644-653; 1993.