TERATOLOGY 42:337-345 (1990)
Growth Failure in Second-Trimester Fetuses With Trisomy 21 JACK FITZSIMMONS, SABINE DROSTE, THOMAS H. SHEPARD, JULIE PASCOE-MASON, AND ALAN FANTEL Department of Obstetrics and Gynecology, University of Wisconsin, Madison, Wisconsin 53715 (J.F.); Departments of Obstetrics and Gynecology (S.D., T.H.S.) and Pediatrics (T.H.S., J.P.-M., A.F.), University of Washington, Seattle, Washington 98195
ABSTRACT
Growth failure in the Down syndrome is common postnatally, but is thought to be less consistent in fetuses and newborns. We describe the growth of individual organs in 53 second-trimester abortuses with trisomy 21 and compare the organ weights to organ weights from 432 spontaneously aborted, but otherwise normal control specimens. Using multiple regression analysis, we found body weight to be the most significant predictor of all organ weights in normal fetuses; therefore, this variable was used to generate the regression lines to which the organ weights of trisomic specimens were compared. All trisomic fetal organs were found to be small, with a n abnormal karyotype being a significant predictor of low organ weight. However, the effect on individual organs was variable, with some organs differing only minimally from the controls. Placental weights were not affected by fetal trisomy. This study demonstrates the presence of well-established, although variably severe, growth retardation in second-trimester fetuses with Down syndrome.
Growth failure is a universal feature of Down syndrome (Barden, '83; Cronk, '78; Kucera and Dolezalova, '72; Thelander and Pryor, '66). Affected individuals show a n increasing deviation from normal growth patterns with increasing age (Barden, '83; Cronk, '78; Thelander and Pryor, '66). However, fetuses and newborns with trisomy 21 show only minimal, if any, evidence of deficient growth (Barden, '83; Cronk, '78; Golbus, '78; Kucera and Dolezalova, '72), raising questions about both the time of onset of growth disturbance as well as factors controlling intrauterine growth. We have previously presented data demonstrating t h a t the growth of fetal limbs is subtly, but consistently diminished as early as the second trimester (FitzSimmons e t al., '89). We have expanded this study to investigate the effects of trisomy 21 on the growth of individual organs. MATERIALS AND METHODS
The Central Laboratory for Human Embryology a t the University of Washington receives fetal specimens from several obstetrical services in Seattle. These specimens are the products of either spontaneous 0 1990 WILEY-LISS, INC
or induced abortion and a r e subjected to complete autopsy. All findings are stored in a computerized data base. A thorough description of the operation of this laboratory has been published previously (Shepard et al., '88). All fetuses with trisomy 21 received in the laboratory t h a t underwent a complete autopsy are included in the present study. These fetuses were identified by means of prenatal diagnosis and terminated electively by prostaglandin induction. The records were reviewed and the following data retrieved: sex, foot length, crownrump length (CRL), body weight, and weights of the heart, lungs, thymus, liver, kidney, adrenals, spleen, and placenta. A control group was selected from the records of the laboratory. These specimens resulted from spontaneous abortions and were subjected to the same thorough postmortem examination as the study group. All fetuses in this group had normal findings on autopsy. In addition, no macerated specimens are included. While chromosomal
Received February 5, 1990; accepted April 24, 1990. This work was supported by NIH grant HD000836.
338
J. FITZSIMMONS ET AL.
.. 0 0
I
I
I
10
20
30
40
50
60
FOOTLENGTH Fig. 1. Values given in mm. Normal controls indicated by dots, trisomy 21 values by open circles. Correlation:y = 35.954 + 3.8710~;RZ = 0.931.
0
10
20
30
40
50
FOOTLENGTH Fig. 2. Foot length is given in mm; body weight is in grams. Normal controls indicated by dots, trisomy 21 values by open circles. Correlation:y = 24.343 > 3.6095~+ 0.3657~';R2 = 0.955.
60
339
FETAL GROWTH FAILURE
1
200
0
400
600
aoo
1000
BODY WEIGHT Fig. 3. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.02925 + 0.007323~;R2 = 0.939.
30
20
10 ,
. .:: f i
...0
. ;.-: 0-2
....:.0, , 0
200
400
600
800
BODY WEIGHT Fig. 4. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = 0.9880 + 0.02464~;R2 = 0.839.
1000
340
J. FITZSIMMONS ET AL.
200
0
400
600
800
1000
BODY WEIGHT Fig. 5. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = 0.02029 + 0.001235~+ 0.000001993~~; R2 = 0.808.
40 -
30 -
20
-
10 -
0
200
400
600
800
BODY WEIGHT Fig. 6. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.2614 + 0.05284~;R2 = 0.916.
1000
341
FETAL GROWTH FAILURE
BODY WEIGHT Fig. 7. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.1159 + 0.009248~;R2 = 0.925. 2
I
.
. . .
/
Fig. 8. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.0808 + O.O0144x, R2 = 0.826.
342
J. FITZSIMMONS ET AL.
0
200
400
600
800
1000
BODY WEIGHT Fig. 9. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation: y = 0.09228 0.003797~;R2 = 0.759.
+
400
300
200
100
0 0
200
400
600
800
BODY WEIGHT Fig. 10. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation: y = 57.392 + 0.2653~;R2 = 0.664.
1 ooc
343
FETAL GROWTH FAILURE
TABLE 1 . MultiDle regresswn results for bodv weight Beta €3 t(485) P Foot length ,8375 13.6412 14.499 0.000 Karyotype CRL Sex
-.0495 ,1143 -.0270
-26.5696 .4665 -9.7495
-2.908 1.979 -1.739
0.004 0.046 0.079
analysis was not performed, aneuploidy should be rare in this group (Winter et al., ’88). Data on the same variables as noted above were obtained in these normal controls. Regression curves were plotted for body weight and individual organ weights against foot length and CRL, and for organ weights against total body weight. Values obtained from the trisomy 21 fetuses were then compared to these normal curves. Multiple regression analysis was performed to identify significant predictors of individual organ weights. RESULTS
The control group consisted of 432 fetuses. Since not all measurements were available on all fetuses, individual plots may be based on fewer measurements. Foot length, CRL and body weight were tightly correlated (RB = 0.931 t o 0.955). The study group of 53 fetuses was compared with the normal controls for foot length, CRL, and body weight. As with the controls, all measurements were not available for all fetuses. Foot length and CRL were proportionate and no different from the normal controls (Fig. 1). Multiple regression analysis was used to evaluate the relative importance of fetal sex, CRL, foot length, and karyotype in predicting body weight. All except fetal sex showed significant correlation (Table 1). That is, trisomic fetuses were lighter than normal a t each footlength and at each CRL (data not shown, but available upon request). While the difference between normal and trisomy 21 fetal body weights was highly statistically significant, the actual differences were rather small, with considerable overlap with the normal population (Fig. 2). Weights of each organ were plotted against CRL, foot length, and body wei ht. Correlations for each were quite high (R8 > 0.730),but body weight was consistently the best predictor of organ weight (Table 2). The values for the present controls were quite similar to the normal standards published from this laboratory (Shepard et al., ’88).
Weights for the organs from trisomy 21 fetuses were plotted against the regression curves for the body weights of the normal controls. In each case, the organ weights of fetuses with trisomy 21 were low compared with the normal population. Some organs, notably heart and spleen, appeared closer to the normal mean than others, however (Figs. 3-10). Multiple regression again confirmed that karyotype was a significant predictor of individual organ weights (Table 2). Placental weights were also analyzed. In contrast to the above data, body weight correlated poorly with placental weight, as did karyotype. Foot length was the only independently significant predictor of the weight of the placenta. Placental weights from fetuses with trisomy 21 were not different from the control group. DISCUSSION
Growth retardation is well-documented as a component of Down syndrome in childhood through adult life (Barden, ’83; Cronk, ’78; Kucera and Dolezalova, ’72; Thelander and Pryor, ’66). However, considerably less information is available on the possible fetal manifestations. While there appears to be general agreement that some chromosomal abnormalities are associated with growth retardation (Gabbe et al., ,881, which is generally referred to as “symmetric,”data supporting this impression are scant for the fetus with trisomy 21, and limited to discussion of stature or body weights. Kucera and Dolezalova (‘72) evaluated birth weights of infants with Down syndrome delivered after 27 weeks’ gestation and found diminished growth after 32 weeks. Relatively few newborns under 32 weeks were studied, and only menstrual dates were used to establish gestational age. Golbus (’78)found no difference in a variety of measurements, including organ weights, from 13 second-trimester fetuses with trisomy 21. Unfortunately, the data were presented as a aggregate covering 18-21 weeks’ gestation. Data from the present study suggest that the differences between trisomy 21 and normal fetuses are small, and thus likely to be obscured by using relatively broad ranges of gestational age. Stephens and Shepard (’80) described the physical findings associated with trisomy 21 in the fetus and noted a diminished crownrump length. Additionally, short limblengths have been described both on ultra-
344
J. FITZSIMMONS ET AL.
TABLE 2 . Abbreviated multiple regression results indicating independent variables significantly predicting
Heart Lungs Thymus Liver Kidneys Adrenals Spleen Placenta
n for each group Controls Trisomy 21 241 33 247 33 224 31 239 33 242 33 240 34 225 32 185 44
individual organ weights p values for each independent variable Body weight Karyotype Crown-rump length Foot length
Growth Failure in Second-Trimester Fetuses With Trisomy 21 JACK FITZSIMMONS, SABINE DROSTE, THOMAS H. SHEPARD, JULIE PASCOE-MASON, AND ALAN FANTEL Department of Obstetrics and Gynecology, University of Wisconsin, Madison, Wisconsin 53715 (J.F.); Departments of Obstetrics and Gynecology (S.D., T.H.S.) and Pediatrics (T.H.S., J.P.-M., A.F.), University of Washington, Seattle, Washington 98195
ABSTRACT
Growth failure in the Down syndrome is common postnatally, but is thought to be less consistent in fetuses and newborns. We describe the growth of individual organs in 53 second-trimester abortuses with trisomy 21 and compare the organ weights to organ weights from 432 spontaneously aborted, but otherwise normal control specimens. Using multiple regression analysis, we found body weight to be the most significant predictor of all organ weights in normal fetuses; therefore, this variable was used to generate the regression lines to which the organ weights of trisomic specimens were compared. All trisomic fetal organs were found to be small, with a n abnormal karyotype being a significant predictor of low organ weight. However, the effect on individual organs was variable, with some organs differing only minimally from the controls. Placental weights were not affected by fetal trisomy. This study demonstrates the presence of well-established, although variably severe, growth retardation in second-trimester fetuses with Down syndrome.
Growth failure is a universal feature of Down syndrome (Barden, '83; Cronk, '78; Kucera and Dolezalova, '72; Thelander and Pryor, '66). Affected individuals show a n increasing deviation from normal growth patterns with increasing age (Barden, '83; Cronk, '78; Thelander and Pryor, '66). However, fetuses and newborns with trisomy 21 show only minimal, if any, evidence of deficient growth (Barden, '83; Cronk, '78; Golbus, '78; Kucera and Dolezalova, '72), raising questions about both the time of onset of growth disturbance as well as factors controlling intrauterine growth. We have previously presented data demonstrating t h a t the growth of fetal limbs is subtly, but consistently diminished as early as the second trimester (FitzSimmons e t al., '89). We have expanded this study to investigate the effects of trisomy 21 on the growth of individual organs. MATERIALS AND METHODS
The Central Laboratory for Human Embryology a t the University of Washington receives fetal specimens from several obstetrical services in Seattle. These specimens are the products of either spontaneous 0 1990 WILEY-LISS, INC
or induced abortion and a r e subjected to complete autopsy. All findings are stored in a computerized data base. A thorough description of the operation of this laboratory has been published previously (Shepard et al., '88). All fetuses with trisomy 21 received in the laboratory t h a t underwent a complete autopsy are included in the present study. These fetuses were identified by means of prenatal diagnosis and terminated electively by prostaglandin induction. The records were reviewed and the following data retrieved: sex, foot length, crownrump length (CRL), body weight, and weights of the heart, lungs, thymus, liver, kidney, adrenals, spleen, and placenta. A control group was selected from the records of the laboratory. These specimens resulted from spontaneous abortions and were subjected to the same thorough postmortem examination as the study group. All fetuses in this group had normal findings on autopsy. In addition, no macerated specimens are included. While chromosomal
Received February 5, 1990; accepted April 24, 1990. This work was supported by NIH grant HD000836.
338
J. FITZSIMMONS ET AL.
.. 0 0
I
I
I
10
20
30
40
50
60
FOOTLENGTH Fig. 1. Values given in mm. Normal controls indicated by dots, trisomy 21 values by open circles. Correlation:y = 35.954 + 3.8710~;RZ = 0.931.
0
10
20
30
40
50
FOOTLENGTH Fig. 2. Foot length is given in mm; body weight is in grams. Normal controls indicated by dots, trisomy 21 values by open circles. Correlation:y = 24.343 > 3.6095~+ 0.3657~';R2 = 0.955.
60
339
FETAL GROWTH FAILURE
1
200
0
400
600
aoo
1000
BODY WEIGHT Fig. 3. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.02925 + 0.007323~;R2 = 0.939.
30
20
10 ,
. .:: f i
...0
. ;.-: 0-2
....:.0, , 0
200
400
600
800
BODY WEIGHT Fig. 4. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = 0.9880 + 0.02464~;R2 = 0.839.
1000
340
J. FITZSIMMONS ET AL.
200
0
400
600
800
1000
BODY WEIGHT Fig. 5. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = 0.02029 + 0.001235~+ 0.000001993~~; R2 = 0.808.
40 -
30 -
20
-
10 -
0
200
400
600
800
BODY WEIGHT Fig. 6. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.2614 + 0.05284~;R2 = 0.916.
1000
341
FETAL GROWTH FAILURE
BODY WEIGHT Fig. 7. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.1159 + 0.009248~;R2 = 0.925. 2
I
.
. . .
/
Fig. 8. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation:y = -0.0808 + O.O0144x, R2 = 0.826.
342
J. FITZSIMMONS ET AL.
0
200
400
600
800
1000
BODY WEIGHT Fig. 9. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation: y = 0.09228 0.003797~;R2 = 0.759.
+
400
300
200
100
0 0
200
400
600
800
BODY WEIGHT Fig. 10. Values are given in grams. Normal controls indicated by dots; trisomy 21 values by open circles. Correlation: y = 57.392 + 0.2653~;R2 = 0.664.
1 ooc
343
FETAL GROWTH FAILURE
TABLE 1 . MultiDle regresswn results for bodv weight Beta €3 t(485) P Foot length ,8375 13.6412 14.499 0.000 Karyotype CRL Sex
-.0495 ,1143 -.0270
-26.5696 .4665 -9.7495
-2.908 1.979 -1.739
0.004 0.046 0.079
analysis was not performed, aneuploidy should be rare in this group (Winter et al., ’88). Data on the same variables as noted above were obtained in these normal controls. Regression curves were plotted for body weight and individual organ weights against foot length and CRL, and for organ weights against total body weight. Values obtained from the trisomy 21 fetuses were then compared to these normal curves. Multiple regression analysis was performed to identify significant predictors of individual organ weights. RESULTS
The control group consisted of 432 fetuses. Since not all measurements were available on all fetuses, individual plots may be based on fewer measurements. Foot length, CRL and body weight were tightly correlated (RB = 0.931 t o 0.955). The study group of 53 fetuses was compared with the normal controls for foot length, CRL, and body weight. As with the controls, all measurements were not available for all fetuses. Foot length and CRL were proportionate and no different from the normal controls (Fig. 1). Multiple regression analysis was used to evaluate the relative importance of fetal sex, CRL, foot length, and karyotype in predicting body weight. All except fetal sex showed significant correlation (Table 1). That is, trisomic fetuses were lighter than normal a t each footlength and at each CRL (data not shown, but available upon request). While the difference between normal and trisomy 21 fetal body weights was highly statistically significant, the actual differences were rather small, with considerable overlap with the normal population (Fig. 2). Weights of each organ were plotted against CRL, foot length, and body wei ht. Correlations for each were quite high (R8 > 0.730),but body weight was consistently the best predictor of organ weight (Table 2). The values for the present controls were quite similar to the normal standards published from this laboratory (Shepard et al., ’88).
Weights for the organs from trisomy 21 fetuses were plotted against the regression curves for the body weights of the normal controls. In each case, the organ weights of fetuses with trisomy 21 were low compared with the normal population. Some organs, notably heart and spleen, appeared closer to the normal mean than others, however (Figs. 3-10). Multiple regression again confirmed that karyotype was a significant predictor of individual organ weights (Table 2). Placental weights were also analyzed. In contrast to the above data, body weight correlated poorly with placental weight, as did karyotype. Foot length was the only independently significant predictor of the weight of the placenta. Placental weights from fetuses with trisomy 21 were not different from the control group. DISCUSSION
Growth retardation is well-documented as a component of Down syndrome in childhood through adult life (Barden, ’83; Cronk, ’78; Kucera and Dolezalova, ’72; Thelander and Pryor, ’66). However, considerably less information is available on the possible fetal manifestations. While there appears to be general agreement that some chromosomal abnormalities are associated with growth retardation (Gabbe et al., ,881, which is generally referred to as “symmetric,”data supporting this impression are scant for the fetus with trisomy 21, and limited to discussion of stature or body weights. Kucera and Dolezalova (‘72) evaluated birth weights of infants with Down syndrome delivered after 27 weeks’ gestation and found diminished growth after 32 weeks. Relatively few newborns under 32 weeks were studied, and only menstrual dates were used to establish gestational age. Golbus (’78)found no difference in a variety of measurements, including organ weights, from 13 second-trimester fetuses with trisomy 21. Unfortunately, the data were presented as a aggregate covering 18-21 weeks’ gestation. Data from the present study suggest that the differences between trisomy 21 and normal fetuses are small, and thus likely to be obscured by using relatively broad ranges of gestational age. Stephens and Shepard (’80) described the physical findings associated with trisomy 21 in the fetus and noted a diminished crownrump length. Additionally, short limblengths have been described both on ultra-
344
J. FITZSIMMONS ET AL.
TABLE 2 . Abbreviated multiple regression results indicating independent variables significantly predicting
Heart Lungs Thymus Liver Kidneys Adrenals Spleen Placenta
n for each group Controls Trisomy 21 241 33 247 33 224 31 239 33 242 33 240 34 225 32 185 44
individual organ weights p values for each independent variable Body weight Karyotype Crown-rump length Foot length
