RICHARD J. SHERWOOD,1HAYNES B. ROBINSON, RICHARD S. MEINDL,3 AND RICHARD L. MAY
Abstract Identifying patterns of fetal growth alteration benefits both the clinician and the researcher. Twenty-four measurements in three variable sets (anthropometric measures, organ weights, and long-bone measures from radiographs) were taken on fetuses both with and without pathological conditions that are suspected to result in growth alteration. In addition, radiographs of each case were examined for the presence or absence of ossification centers. Based on least-squares regressions of the normal group, we calculated standardized residuals for the affected group to identify patterns of growth alteration. A large sample of fetuses between 15 and 42 weeks of gestational age with a variety of pathological conditions is described and evaluated for growth alterations. Symmetric and asymmetric growth alteration was detected in a small part of the sample and was predominantly isolated to fetuses in the late third trimester. Although patterns of growth alteration have been suggested as a means for noninvasive diagnoses of syndromes (such as trisomy 21), no consistent patterns are discernible in the current group. The sample provides a unique opportunity to evaluate fetal growth in terms of the interaction between genetic and environmental influences.
KEY WORDS: GROWTH ALTERATION, STANDARDIZED RESIDUALS, FETAL GROWTH
Insults to a fetus appear in a variety of forms (Gruenwald 1970; Woods et al. 1979; Brooke et al. 1984; McFadyen 1989; FitzSimmons et al. 1989), including (1) nutritional and oxygen deprivation (uteroplacental vascular insufficiency), (2) environmental insults (e.g., teratogens and maternal metabolic disorders such as diabetes or chronic infection), and (3) chromosomal, multifactorial, and single-gene disorders. Many of these insults are thought to symmetrically depress growth, producing an overall small infant. However, differential patterns of growth retardation of organs or organ systems may occur, resulting in asymmetric growth retardation. For example, in a study of small-for-date infants, Brooke et al. (1984) noted that forearm growth was differentially affected compared to upper arm growth. This has also been reported for cases of trisomy 21 (FitzSimmons et al. 1989). Barr (1994) noted growth retardation in both arm and leg measures for trisomy 21 individuals. Furthermore, limb proportions are said to differ in such syndromes as Cockayne, Marfan, Coffin-Lowry, and Marden-Walker (Merlob et al. 1984). Honnebier and Swaab (1973) and Kalifa et al. (1989) have also reported anencephalic fetuses that manifest extreme limb disproportions. Skeletal maturity itself also may be affected. As examples, some fetuses with trisomies present retarded onset of ossification, whereas anencephalic fetuses show advanced onset (Pryse-Davies et al. 1974).
Another possible pattern of differential growth involves an organ hierarchy, the most notable of which is termed "brain sparing." In such cases the brain is said to possess priority over the other organs, thus remaining relatively unaffected in fetuses that otherwise show intrauterine growth retardation [McFadyen 1989; Kerr et al. 1973; however, see also Brooke et al. (1984) and Crane and Kopta (1980)]. Ultrasonography now provides a methodology with the potential for early recognition of such symmetric and asymmetric growth disorders (Barr et al. 1994; Chamberlain 1989). Therefore linking patterns of affected growth to specific syndromes may provide a noninvasive method for early detection of developmental alterations [see McBride et al. (1984)]. In addition, identifying the pattern and, more important, the timing of such inequalities not only would aid in assessing pathological conditions but also would further our understanding of human fetal growth. Such identification may also have the additional benefit of providing teratogen surveillance (Kallen 1989).
Although various syndromes have been associated with patterns of retarded growth in childhood, the fetal onset of growth retardation is largely unexplored [e.g., Brooke et al. (1984) and Miller and Hassanein (1971)]. The main focus of the current study is to explore the growth of fetuses with pathological conditions for symmetric and/or asymmetric alterations. If alterations exist, the next step is to identify the responsible processes. After this is done (and only after), the utility for clinical diagnoses will become apparent.
To identify patterns of growth alterations, we test the following null hypotheses: (1) overall growth of affected fetuses does not differ from that of the control group; (2) affected fetuses do not show asymmetric growth; (3) individual pathological conditions are not consistently associated with a pattern of growth alteration. If this last hypothesis cannot be rejected, then noninvasive evaluations of fetuses (by means of such techniques as ultrasonography) will not be a useful diagnostic tool.
Once patterns of growth have been established for affected fetuses, the processes behind those patterns can be examined. Because fetuses in the current sample have been subjected to a number of differential insults, both genetic and environmental, this study provides a unique opportunity to investigate the interaction between these forces.
Materials and Methods
A sample of 522 fetuses ranging from 15 to 43 weeks of gestational age was examined. Requirements for inclusion in the final sample included a detailed pathologist's report and a detailed maternal history. Gestational age for all fetuses was based on maternal reports of last normal menstrual period (LNMP) deemed to be reliable by the investigators. (Instances where multiple conflicting reports of LNMP were given were excluded, as were cases where the maternal report clearly conflicted with clinical evaluation.) Multiple births were excluded from the sample.
Each fetal specimen underwent a complete autopsy performed or overseen by one of us (H.B. Robinson). Principal pathological diagnosis and mechanism of fetal loss (MOE;L) were determined from gestational history, dysmorphic evaluation of the fetus, gross anatomical and histological examination of the fetus and the placenta, x-ray examination of the fetus, and chromosomal analysis.
One hundred sixty-eight fetuses, ranging in age from 15 to 42 weeks of gestational age, met the requirements and were included in the final analysis. Ossification site appearance was examined by means of radiographs, as were measures of long-bone diaphyses. Anthropometric measurements and organ weight data were taken at the time of autopsy.
It is common practice to avoid using maternally calculated gestational age in studies such as this one because it may introduce a source of error. A growth indicator such as body weight, brain weight, or biparietal diameter (most often used in ultrasound studies) is often used as a surrogate for age [e.g., Barr et al. (1994) and Cuckle et al. (1989)]. We have chosen gestational age for the following reasons: (1) It is the only independent variable available, (2) identification of event timing is crucial to understanding growth, and (3) use of anthropometric growth indicators (e.g., head circumference) ignores the variation present in these measures and therefore may confound conclusions. Even with the variation inherent in reports of LNMP, the current methodology allows for accurate assessment of relative growth within a given individual.
Three sets of measurements were obtained from each specimen (see Table 1). Principal diagnosis was used to place fetuses into one of several classes, which were then grouped as follows:
1. Nondysmorphic, normal growth expected (n = 88). These fetuses include spontaneous abortions resulting from infections or acute placental-cord compromise. This group serves as a control group for assessing normal growth in the sample.
2. Trisomies (n = 15). Two trisomic abnormalities were encountered with frequency, trisomies 18 and 21; in addition, one case of trisomy 13 is included.
Other chromosomal anomalies (n = 9), that is, infrequent anomalies such as chromosomal breakage, triploid, tetraploid, and any sex chromosome anomaly.
Known and unknown genesis dysmorphic processes (n = 49). This group includes osteogenesis imperfecta, lissencephaly, anencephaly, and Potter's syndrome. It is recognized that this class is large and heterogeneous; because each case is analyzed on an individual basis, the grouping here is merely for convenience.
1. Nondysmorphic, growth retardation expected (n = 9). These fetuses are primarily chronic cases of uteroplacental vascular insufficiency.
A second variable, MOFL, was also identified for each fetus. In almost all cases MOt, and principal diagnosis were identical. There were a few instances in which the two variables differed, but mostly the MOFL was such that it should not play a role in any results. For example, an anencephalic fetus may have infection as its MOFL. Any growth disorders in this case were attributable to diagnosis, not to the MOFL.
Anthropometric Measurements. All linear measurements were made with a ruler to the nearest 0.1 cm. Circumferential measurements were made with a paper measuring tape to the nearest 0.1 cm.
The arm length measurement was taken with the forearm bent to form a 90deg angle with the upper arm. The measurement was taken from the tip of the elbow (olecranon) to the tip of the middle finger. For leg length the leg was flexed at the knee to form a 90deg angle between the lower leg and the thigh. The measurement was taken from the top of the knee (upper margin of the patella) to the sole of the foot positioned at 90deg to the lower leg. Foot length was measured from the heel to the tip of the longest toe. Fetuses were measured for crown-rump length in the supine position with the external auditory canal and the outer canthus lying approximately in a plane at a right angle to the table top and the hips flexed at 90deg to the plane of the back. The measurement was taken from the vertex of the skull to the lowest point of the buttock. Head circumference was measured at the level of the orbital ridges anteriorly and the occiput posteriorly. Abdominal circumference was taken at the level of the umbilicus. Body weight was measured on an Ohaus triple beam balance to the nearest gram.
Organ Weights. All organs were weighed on an Ohaus triple beam balance to the nearest 0.1 g. Brain weight included cerebral hemispheres, cerebellum, pons, and medulla oblongata. The heart was weighed after section of the superior vena cava and the inferior vena cava at their junctions with the right atrium, the pulmonary arteries at their junctions with the left ventricle, and the aorta and pulmonary arteries at their origins at the base of the heart. Thymus, kidneys, and adrenals were dissected free of fat and connective tissue. The kidneys, spleen, and lungs were sectioned at their hilum. The liver was sectioned at the porta hepatis and included the weight of the gall bladder, its contents, and the bile ducts.
Long-Bone and Vertebral Height Measurements. A standard set of xrays was obtained on all specimens using a Picker Vector II x-ray machine. Exposures were on ultrasound film using no filter. Full-body radiographs of specimens less than 16 cm in crown-rump length were exposed at 400 mA for 0.022 s with 34 kV. Exposure time for specimens with crown-rump lengths exceeding 16 cm was 0.030 s. Views of hands and feet were exposed at 300 mA for 0.017 s with 34 kV. Tube height ranged between 91.00 and 102.00 cm. Anteroposterior and lateral full body views and anteroposterior views of hands and feet were obtained on each specimen. Fetuses were placed directly on the plate to minimize enlargement.
Maximum diaphyseal length was measured for the humerus, radius, ulna, femur, tibia, fibula, first metacarpal, and first metatarsal. The maximum height of the first lumbar vertebra was also measured. Measurements were taken to the nearest 0.01 cm. In some cases it was obvious that, because of flexion of limbs, measurements would be affected by parallax. Based on the judgment of the observer, measurements were taken from the side from which the image was judged to be most reliable.
Statistical Analysis. The sample was divided into two age classes: 15-25 weeks and 2543 weeks of gestational age. First, the relationship between each variable and gestational age was assessed to ensure that it was adequately described by a linear equation. For all length measures (foot length, arm length, etc.) the criterion of linearity was met. Quadratic equations were tested for these variables, but they did not make a significant addition to the descriptive power of the model and were abandoned. Body and organ weights were found to have a curvilinear relationship with respect to age. In these cases a logarithmic transformation was used to achieve linearity. Leastsquares regressions were calculated for each variable on gestational age for both age classes in the control group (group 1). The model for each variable was where Y is the dependent variable, x is the independent variable (gestational age), b is the slope, a is the y intercept, and Res is the vertical distance from the fitted regression line, sign considered. Standardized residuals based on Eq. (1) were then calculated using the standard error of the regression for each affected fetus. Figure 3, for example, is a representation of the standardized residual scores. This method has been shown to be effective for comparing a large assortment of variables in equivalent terms for both interand intra-individual growth (Sherwood, Robinson et al. 1992) by allowing for identification of normal (near zero line) and altered (away from zero line) growth and for direction of growth alteration (advanced or retarded). Furthermore, true detection of asymmetric growth is ascertainable with this technique.
Radiographs were also examined to determine whether ossification centers were present, and a Guttman scale model was used to determine the sequence of appearance [see Meindl and Lovejoy (1985)]. Within the age range sampled nine centers appeared: ischium, pubis, calcaneus, talus, first lateral sacral mass, distal femoral epiphysis, proximal tibial epiphysis, proximal humerus epiphysis, and cuboid.
To frame ossification in terms of age, we chose six of the nine centers that adequately sample the period between 15 and 42 weeks (ischium, pubis, calcaneus, talus, lateral sacral mass one, and cuboid). The total number of ossification centers (TNOC) (Pryse-Davies et al. 1974; Newell-Morris and Tarrant 1978) was calculated for each individual based on these six centers. Mean, standard deviation, and range of ages for each TNOC value in the control group provide the means for comparison (Figure 1).
Results
Ossification Sequences. The Guttman analysis revealed the modal ossification sequence to be stable. The sequence of appearance of ossification centers is as follows: ischium (first), calcaneus, pubis, talus, first lateral sacral mass, distal femur, cuboid, proximal tibia, and proximal humerus (last). Pryse-Davies et al. (1974) described a similar sequence; however, they lacked the pelvic centers.
Deviations from the modal sequence in this study were rare and occurred between centers that appear in temporal proximity [Garn and Rohmann (1959, 1968) found frequent variation in hand and foot centers that also show temporal proximity.] The following deviations were noted in the sample: Nine cases display the ischium and pubis as their only centers, whereas eight cases have the ischium and calcaneus. In other words, the pubis and calcaneus appear to have an equal likelihood of being the second center of ossification. Similarly, the lateral sacral mass exchanged position with the cuboid once and with the distal femur twice [Pryse-Davies et al. (1974) also noted this deviation with the cuboid and distal femur.] In the current study the variation in sequence showed no relationship to diagnosis, MOFL, or sex.
Residual Analysis. Regression models describing the control sample are listed in Table 1. Distributions of maternal age, maternal race, and fetal age for all groups are given in Figure 2.
Because reports have indicated that maternal age or gravidity plays a role in fetal size, it was necessary to test the independence of the variables from these factors. One-way analysis of variance was used to test each variable by gravidity (number of previous pregnancies) and maternal age (divided into 5-year age classes). No relationship was found with either maternal variable.
In this study we define slight deviations as those that fall within one standard deviation of the expected value; moderate deviations are between one and two standard deviations; and those identified as sizable exceed two standard deviations. Positive residual values indicate a measurement larger than expected (advanced growth), and negative values indicate a measurement smaller than expected (retarded growth).
2. Trisomies. There are 11 cases of trisomy 21, 3 cases of trisomy 18, and 1 case of trisomy 13 (see Figure 3). The growth of this group is not different from the control group during the second trimester, excluding one fetus (16 weeks, trisomy 21), which showed sizable advancement on many measures except organ weights that fell within two standard deviations of expected. This case of advancement is especially curious because the mother reported smoking two packs of cigarettes a day, a condition normally associated with growth depression (McFadyen 1989; Yerushalmy 1971; Rubin et al. 1986). Two third-trimester cases (25 weeks and 31 weeks, both trisomy 21) showed sizable retardation (over 2.5 standard deviations below expected in nearly all measures).
The three cases of trisomy 18 showed reduction in all variables, including ossification scores; however, only slight retardation was seen in the early cases. Droste et al. (1990) and Golbus (1978) reported similar findings. The older specimen (36 weeks) was slightly more retarded with several measures falling more than one standard deviation below those of the controls.
The trisomy 13 individual (27 weeks) showed slight to moderate advancement on all measures except foot length (- 0.42) (organ weights were not available). Many measures for this fetus were close to normal and none showed deviation greater than 1.50 (crown-rump length).
Of the trisomy cases showing growth retardation the long bones were most affected, whereas crown-rump length, head circumference, and brain weight were less affected. In specimens that were small for most measures brain weight remained close to or above expected values. It is interesting to note that for a 31-week-old fetus that displayed growth retardation across all other measures, the heart and lungs were larger than expected (standardized residuals 2.95 and 1.24, respectively) and considerably larger given the status of the remaining measures.
Three specimens fell outside the observed range of ossification for control specimens. The first (16 weeks, trisomy 21) had one ossification center that would be scored as somewhat advanced. The paucity of young individuals in our control sample may offer an explanation for this. Two cases (31 weeks, trisomy 21; 36 weeks, trisomy 18) showed only four ossification centers at an age when five centers would be expected.
Other Chromosomal Anomalies. The composite group of fetuses with other chromosomal anomalies contained five specimens with Turner's syndrome (16 weeks, 18.5 weeks, 19.5 weeks, 20 weeks, 22.5 weeks), two specimens with Klinefelter's syndrome (19 weeks, 33 weeks), one specimen with a chromosomal breakage (22 weeks), and one case of triploidy (19 weeks). Most specimens, with the exception of one fetus with Turner's syndrome (22.5 weeks), showed uniform advancement on most measures. [It should be noted that two specimens with Turner's syndrome (18.5 weeks, 22.5 weeks) had only linear autopsy measures (x-ray and organ weights were unavailable).]
The degree of advancement for the younger specimens is minimal with average scores of about 0.40 standard deviation above the expected and maximums rarely exceeding 1 standard deviation. The oldest Klinefelter's syndrome specimen displayed a pattern of moderate advancement. Measures were not all within this range, however; some fell just above expected (metacarpal x-ray, metatarsal x-ray, foot) and others fell slightly below expected (lumbar height, leg, adrenal weight). Unusually, the crown-rump length showed moderate retardation (1.26 standard deviations below expected).
The case of Turner's syndrome that showed growth retardation presents moderate reduction on most measures. The case of triploidy showed moderate advancement on most measures, although some measures (abdominal circumference and liver weight) were close to expected values.
All cases were within expected values for TNOC (the case of triploidy and one Turner's syndrome case did not have ossification data available).
Genesis Dysmorphic Processes. There were 49 specimens in the group with genesis dysmorphic processes, including 11 anencephalic cases, 5 cases of neural tube defects (spina bifida or myelomenigocele), 2 cases with omphalocele, 9 cases of hydrops (it is understood that a condition such as hydrops has numerous causes; however, the placement in this or the subsequent group is based on the complete evaluation, and this term is used for identification), 1 case of partial trisomy 4, 4 cases of osteogenesis imperfecta, 1 case of lissencephaly, and 1 case of class B maternal diabetes (Table 2). Because of multiple fractures in cases of osteogenesis imperfecta, long-bone and appendicular measurements were not available.
The two cases of nonimmune hydrops showed a mosaic of scores. The youngest showed scores that were advanced. Organ weights showed the lowest deviation from expected with only heart weight exceeding one standard deviation above expected. Anthropometric measures showed a range from moderate to advanced growth, and most of the x-ray measures showed advanced growth. The older case had most of its scores ranging from slightly to sizably retarded with only abdominal circumference and body weight approximating their expected values (0.25 and 0.11, respectively). These scores may be more the result of the MOFL, chronic placental-cord compromise, than the primary diagnosis of hydrops.
The partial trisomy 4 and myelomenigoceles subjects had scores with slight growth advancement. Their maturity, as measured by TNOC, was within the mean for the control group.
Of the cases of osteogenesis imperfecta the youngest showed advancement in body weight and all organ weights (none greater than 2 standard deviations). The remaining three cases displayed moderately retarded body weights, and their organ weights showed a mixture of advancement and retardation. Regarding the organ weights two consistencies emerged: The thymus was mildly retarded in the three older specimens, whereas the lung weights were severely retarded.
The lissencephalic subject had moderate to significant retardation in all anthropometric and x-ray measures, except lumbar height, which was mildly advanced. Thymus and kidney weights were mildly advanced, and other organ weights were mildly to moderately retarded. Lung weight was severely depressed ( - 4.03).
The case of maternal diabetes showed moderate advancement on most measures with others showing significant advancement. Given the relationship of insulin to bone growth (Canalis 1983) this is not unexpected and has been reported in other cases (Russell 1973; McFadyen 1989).
It should be noted that the five oldest specimens in this group had significantly depressed lung weights with scores ranging from 3.88 to 6.35 standard deviations below the expected values. This is in accord with the findings of Naeye (1965), who suggested that the lungs were one of the more susceptible organs to growth suppression.
The cases of anencephaly showed a uniform pattern of normal growth for limb measures (for obvious reasons measures such as crown-rump length and body weight were not considered for this group). Most measures fell within 1 standard deviation of expected and only occasionally fell between 1 and 2 standard deviations of expected.
Organ weights were recorded for eight of the anencephalic specimens. The younger cases tended to show moderate retardation of most measures. Adrenal weights showed a consistent pattern of retardation, often reaching values below - 2.00 standard deviations. Some measures such as thymus and spleen weight showed normal or retarded growth in the second trimester but advanced growth in later months.
There were three cases of spina bifida, all showing advancement. Two cases showed mild advancement, and one case (20 weeks) showed significant advances on most measures. Despite advancement in limb length, one specimen (22 weeks) showed slight retardation in adrenal, heart, and lung weights. There are two cases of Potter syndrome (renal agenesis; 32 weeks and 34.5 weeks). Both showed retardation of arm length (moderate in the younger, severe in the older). Also, in both cases the head circumference was advanced (mildly and moderately, respectively). Organ weights were advanced in both cases; however, both liver and heart weights in the younger case were normal. X-ray measures showed no pattern with most measures being retarded in the younger case and advanced in the older.
The single case of Meckel-Gruber syndrome provides some interest. The x-ray measures and anthropometrics indicate mild to severe retardation, whereas body weight is slightly above normal. Among the organ weights the spleen and the heart were moderately advanced, but the adrenals and thymus were retarded (mildly and moderately, respectively); the lungs were severely retarded (more than 6 standard deviations below expected).
Several cases are outside the expected range for TNOC scores. The young case of nonimmune hydrops showed a score that was well advanced. A 21.5week-old amniotic band case showed a lack of ossification centers at a time when at least one center would be expected. The following cases display early attainment of the first ossification center: hydrops with hydronephrosis (15.5 weeks), omphalocele (16 weeks), pentalogy of Cantrell (17 weeks), congenital leukemia (17.5 weeks), spina bifida (17.75 weeks). Finally, the case with Meckel-Gruber syndrome (36 weeks) displayed only four centers when five would be expected.
1. Nondysmorphic, Growth Retardation Expected. There are nine specimens in this group of nondysmorphic cases with expected growth retardation. The youngest five showed slight to moderate retardation. One (21 weeks) showed moderate retardation and was associated with maternal diabetes; another case (23 weeks) showed slight retardation. The third case (25 weeks) showed measures that fell on both sides of the expected values with only one measure exceeding 1 standard deviation in either direction (adrenal weight was 1.44 standard deviations below expected). A 27-week-old fetus showed slight to moderate retardation, and a 28-week-old fetus was normal.
The remaining four specimens displayed severe retardation on all measures except brain weight. They were all 35 weeks old (2 males, 2 females). One male and one female specimen displayed normal brain weights. The other two individuals showed consistent retardation of all measures, but the brain was relatively less affected.
Ossification scores showed evidence of severe retardation. Four cases fell within expected ranges, whereas the remaining five all showed retardation. The first was a 28-week-old fetus with only one center. At this age it would be expected to display three or four centers. The four oldest cases would all be expected to display five centers of ossification. Three of these displayed four and one displayed only three.
Discussion
Ontogenetic Trajectory. Alberch et al. (1979) characterized growth by four parameters governing the ontogenetic trajectory of the individual. Three of these are of particular interest to the current study: onset age of growth, offset signal for growth, and growth rate. For those interested in understanding the growth of fetuses with pathological conditions, especially those syndromes linked with patterns of asymmetric growth, identifying the affected parameters is a necessary starting point.
Onset age of growth is not useful for variables that show a linear relationship with age because growth can be traced only back to conception. However, for measures such as organ weights that have a curvilinear relationship with age it is possible to identify onset as that point at which growth is accelerated. This point occurs roughly at the 25th week for fetuses in this study. Because organs appear to be most susceptible (and differentially so) to fetal insults, identification of acceleration points in normal and affected fetuses is useful for understanding the mechanisms for differential growth.
The offset signal for growth would be indicated by a similar point at which growth is decelerated or terminated and would be readily identifiable for any variable. The most obvious examples of offset signal are epiphyseal fusion of long bones and sutural fusion of the cranium. Premature fusion of either system often results in easily recognizable deformities. Cases of intrauterine asymmetric growth (e.g., brain sparing) may be situations where an offset signal is produced by one organ (in this case the brain) affecting other organ systems.
Onset and offset of growth should occur over a small period of time and therefore should be associated with deflection points in growth curves. From the data presented it appears that most of the observed growth alteration has no notable deflection point and is probably best described as a shift in growth rate rather than a shift in onset or offset signals. Although growth rate is fairly well characterized for normal fetuses, growth rate of affected fetuses is not.
Symmetric and Asymmetric Growth. The use of standardized residuals allows the characterization of both overall fetal growth and the growth of individual organs and organ systems. Residuals that fall within two standard deviations of the expected values are considered to be within the bounds of normal growth. Values exceeding two standard deviations in either direction indicate significantly advanced or retarded growth. This allows the stated hypotheses to be tested.
The first hypothesis to be tested, namely, that overall growth of affected fetuses does not differ from that of the control group, can be rejected. In comparison with the control group it is apparent that the experimental group contains cases of affected growth. Most dramatically, this occurs as severe retardation of growth late in gestational life, being consistent only in those specimens 34 weeks or older. These are cases of symmetric growth retardation; residual scores for any measurement of these individuals frequently approach three standard deviations below normal and occasionally approach six standard deviations below normal. The best evidence for this is found in the nondysmorphic group with expected growth retardation. This is not surprising because any disorder that hinders nutrient intake and/or waste removal (whether fetal or postnatal) will affect the growth process.
It is important to note that before 34 weeks of gestational age specimens present no consistent abnormal pattern. Residual scores primarily fall within two standard deviations of expected values and can be positive or negative. Gruenwald (1970) indicated that, even if factors that would normally affect the growth of a fetus are present in the second trimester, their effects will not be seen until the third trimester. This is likely to be the case for the younger specimens in the nondysmorphic group with expected growth retardation.
Here, although the diagnosis implies a chronic problem, the disorder probably had not had time to severely affect the individuals in the younger part of this group. Therefore, by examining the second- and third-trimester fetuses separately, the hypothesis of normal growth can be rejected only for the thirdtrimester group.
Even though most of the affected fetuses do not show severely affected growth, cases were examined to see whether patterns of asymmetric growth could be detected among the variables, thus testing the second hypothesis. Sherwood, Robinson et al. ( 1992) showed that a measure of dispersal is useful in assessing growth asymmetry (Figure 4). Therefore we calculated an asymmetry index by taking the standard deviation of the variable set (anthropometrics, x-rays, organ weights) for each individual. [Sherwood, Robinson et al. (1992) used range as the measure of dispersal, but we now realize that standard deviation is a better unbiased statistic and have chosen this for the asymmetry index. This statistic was also used by Garn et al. (1985, 1987) and was termed the pattern variability index. It is important to note that use of standard deviation would not have changed the conclusions of our earlier paper.] It is interesting to note that the slope of the asymmetry index by age for all three variables sets in the control group approximates 0, indicating that the level of asymmetry does not significantly change with age in this group. The y intercept does differ between the three sets, slightly showing that x-ray measures show the least degree of asymmetry, organ weights the most, with anthropometric measures intermediate between the two. This asymmetry index has the added value of being relatively insensitive to errors in age assignment.
Again, it is possible to identify individual cases of asymmetric growth and a trend across groups for the affected fetuses. The overall trend is that asymmetric growth is more likely to be identified with organ weights and to occur primarily late in gestation. X-ray measures show little sign of asymmetry, often falling below the expected values (i.e., more symmetric growth than the control group). Anthropometric measures are variable with some cases showing strong indications of asymmetry (although many do not). Again, these are seen most often late in gestation.
Reports of specific patterns of asymmetric growth deserve special attention. The most common report concerns the concept of brain sparing. Experiments on rats and monkeys have shown that in cases of uteroplacental insufficiency asymmetric fetal growth retardation occurs. In these cases the brain displays minimal retardation relative to the liver (Wigglesworth 1964; Hill et al. 1971; Crane and Kopta 1980). The concept of brain sparing postulates that an organ hierarchy exists. The hierarchy is such that under conditions of deprivation available resources will be redirected to those organs with priority with the brain holding high (if not the highest) priority.
With regard to humans, brain sparing has been discussed in absolute and relative terms. Absolute brain sparing results in a brain weight that does not vary from the standard, whereas relative brain sparing allows for variance from the norm but maintains that the brain will be less affected than other organs. Crane and Kopta (1980) charged that absolute brain sparing does not occur and that evidence for relative sparing exists. Davies (1981) reported that in small-for-date babies there is a period of accelerated growth postnatally (often termed "catch-up") in which head circumference (and brain size by inference) shows higher acceleration than other measures. Again, the brain is thought to have a higher priority.
The question of why the brain is the focus of sparing is an interesting one. It can be argued that once the neural systems governing life-sustaining functions exist, the brain is no more important than any other organ. In other words, cognitive function is of little value if the liver, kidneys, etc. are not working properly. Given this, it is interesting to note that, although not definitive, the data presented here do show evidence of brain sparing. Some of these cases could be classified as absolute sparing (e.g., the nondysmorphic group with expected growth retardation). One reason for this may be the particular sensitivity of some organs to environmental factors. For example, the lungs are often affected to extreme levels, but this is probably due to their sensitivity to physical factors such as compression, hydrothorax, or encroaching organs (such as in cases of hiatal hernia).
An alternative explanation for the identification of brain sparing may be found by examining the ontogenetic trajectory of organs. There are two aspects of ontogenetic trajectory that are particularly important: onset of growth and growth rate. As noted, organ weights have a curvilinear relationship with age, and it is possible to identify an inflection point where growth accelerates in each case. The inflection point seen in brain weights is earlier than that found in other organs. It is possible that fetal growth is altered after the inflection point in the brain weight curve but before the inflection point of other organs. Even if organ growth is equally affected after the point of insult, the brain will have enjoyed a period of accelerated growth while the other organs have not. This will give the appearance that the brain has been spared (in this case relatively so), whereas in actuality it may be as affected. As for growth rate, the postinflection rate of brain growth is much higher than that of other organs. In this case a small change in growth rate may have little effect on brain size but a severe impact on other organs. The combination of these two factors can easily result in a situation in which the brain appears to be spared.
From this discussion it is not possible to ascertain whether the brain holds some priority over the other organs. It is possible to argue that the concept of brain sparing needs to be modified. The original scenario posits reduced susceptibility of the brain to insults. It can be argued from the evidence presented here that what is perceived as brain sparing is an artifact of two aspects of ontogenetic trajectory (onset point and growth rate). Alternatively, it could be argued that the brain does show reduced susceptibility and that the mechanism for this reduction has been an altered ontogenetic trajectory in the form of earlier onset of growth and an increase in rate of growth. It is important to note that neither scenario requires that the brain present an offset signal to other organs or that the growth rate of other organs be drastically altered relative to that of the brain.
Claims for another pattern of asymmetric growth have also been discussed. Several researchers (Barr 1994; Benacerraf et al. 1991; Cuckle et al. 1989; FitzSimmons et al. 1989; Nyberg et al. 1993; Rotmensch et al. 1992) have claimed that a pattern of limb reduction is diagnostic for trisomic fetuses. Most of these studies compare limb measures to biparietal diameter (determined ultrasonographically). As in the current study, Barr (1994) used standardized residuals in the analysis. However, age is not a factor and all variables are initially standardized by brain or body weights. Our analysis allows the following comments on this and similar studies. As mentioned, anthropometric measures (Figure 3a) show most of the trisomy cases clustering around expected values. This is especially true for cases in the second trimester. There are four cases in the third trimester, three of which show depressed growth. This is in agreement with the trends seen by Kucera and Dolezalovd (1972) and Golbus (1978). There is an indication from these three cases that limb (most notably arm length) measures may be more affected than other measures.
There is, however, an important point to make. The practice of standardizing variables by another variable (body or brain weight, head circumference) must assume that these measures are normal for these fetuses. If a case were to present normal limb lengths but large body weight, the conclusion would still be that limb lengths are small when, in actuality, body weight is the anomaly. Figure 3a shows clear cases where this occurs. Figure 3c clearly shows that long-bone measures do not vary from the expected in most cases (contrary to the aforementioned studies). Ratio data, such as those presented in the studies mentioned, may prove useful as a diagnostic tool but are misleading in descriptions of growth. Asymmetry indexes for the trisomy cases confirm that these cases do not differ from the normal group.
Kalifa et al. (1989) also made a claim of asymmetric growth, noting that in 11 of 14 cases of anencephaly viewed radiographically there was an excessive length of the upper limbs. So dramatic is this finding that Kalifa et al. (1989) stated "the long limbs make the fetus look like an 'ape' " (p. 112). Unfortunately, Kalifa and colleagues provided no methodology or actual metrics for this claim. The current study does not support these results, showing little deviation from normal in limb measures of anencephalics.
The final hypothesis states that individual pathological conditions are not consistently associated with a pattern of growth alteration. It is a great hope that early noninvasive diagnostic techniques can be developed. In nondysmorphic situations the only way this could be done is through identification of differential growth patterns. Despite the identification of both symmetric and asymmetric growth disturbances in the current sample, we were unable to identify consistent pattern differences associated with a given syndrome.
This is in contrast to the common knowledge that various syndromes (e.g., Down syndrome; FitzSimmons et al. 1989) present identifiable adult morphologies that are the result of growth alterations [see Garn et al. (1987)]. It is clear from the current study that fetal growth is not responsible for these differences and may result from the timing of offset signals such as epiphyseal fusion.
Conclusion
We have found that in our sample of fetuses with pathological conditions growth alterations do occur and some of these may be classified as asymmetric growth. However, the two gestational age groups are notably different. By the end of the second trimester of gestation no syndrome presents any identifiable growth disorder based on the pattern of deviation. Although individual fetuses show growth advancement or retardation, it is clear that at this stage it is not possible to make any inferences regarding syndromes from a noninvasive examination of anthropometric variables. Although sample size is certainly a factor in this conclusion, it is important to note the differences between this and other studies. This study is unique in looking at a large number of variables in fetuses with known gestational age. This allows us to identify the specific growth differences and their direction (advanced or retarded). This is not possible with ratio data.
In addition, timing of growth changes becomes evident; most notably, a trend does start to emerge in the third trimester. Within variable sets calculation of the asymmetry index can provide an indication of the degree of asymmetric growth. Organ weights tend to show greater asymmetry earlier in gestation than either linear variable set. The anthropometric variable set also shows increased asymmetry in several cases but rarely approaches or exceeds the values of the organ weights (see Figure 1) (Sherwood, May et al. 1992; Sherwood, Robinson et al. 1992). Asymmetric growth of organs is shown to be a result of differential ontogenetic trajectory, which may or may not be explained by an organ hierarchy.
Ossification data show an interesting pattern. A modal sequence of ossification center appearance is present and is unaffected by any of the studied pathological conditions. Timing of appearance, however, is affected in several cases, most notably in cases of uteroplacental insufficiency.
Evaluation of growth alteration in rare or common syndromes is often based on anecdotal data without an appropriate comparative sample. We have presented here and elsewhere (Sherwood, May et al. 1992; Sherwood, Robinson et al. 1992) a means to investigate both linear growth and a measure of maturity (progression of ossification centers) in fetuses. We believe that a detailed systematic program of fetal analysis such as this one will prove beneficial to both clinicians and researchers.
Acknowledgments We wish to thank Akron Children's Hospital, in particular, the staff of the Pathology Department, for their help and cooperation. Norman Taslitz was instrumental to the inception of this study and we thank him for his enthusiastic and continued support. Henry Harpending, Victoria Emch, Leslea Hlusko, and several reviewers read early drafts of this paper, and we thank them for their time and comments. R.J. Sherwood would like to personally thank Alan Walker, Jeffrey Kurland, and Dana Duren for helpful discussions and support.
Literature Cited
Alberch, P., S.J. Gould, G.F. Oster et al. 1979. Size and shape in ontogeny and phylogeny. Paleobiology 5:296317.
Barr, M., Jr. 1994. Growth profiles of human autosomal trisomies at midgestation. Teratology 50:395-398.
Barr, M., Jr., W.R. Blackburn, and N.R. Cooley. 1994. Human fetal somatic and visceral morphometrics. Teratology 49:487496.
Benacerraf, B.R., D. Neuberg, and F.D. Frigoletto. 1991. Humeral shortening in secondtrimester fetuses with Down syndrome. Obstet. Gynecol. 77:223-227. Brooke, O.G., C. Wood, and F. Butters. 1984. The body proportions of small-for-dates infants. Early Hum. Dev. 10:85-94.
Canalis, E. 1983. The hormonal and local regulation of bone formation. Endocr. Rev. 4:62-77. Chamberlain, G. 1989. The epidemiology and etiology of intrauterine growth retardation. In Fetal Growth Retardation: Diagnosis and Treatment, A. Kurjak and J. Beazley, eds. Boca Raton, FL: CRC Press, 1-16.
Crane, J.P., and M.E. Kopta. 1980. Comparative newborn anthropometric data in symmetric versus asymmetric intrauterine growth retardation. Am. J. Obstet. Gynecol. 138:518522.
Cuckle, H., N. Wald, J. Quinn et al. 1989. Ultrasound fetal femur length measurement in the screening for Down's syndrome. Br. J. Obstet. Gynecol. 96:1373-1378. Davies, D.P. 1981. Growth of "small-for-dates" babies. Early Hum. Dev. 5:95-105. Davies, D.P., and D. Beverley. 1979. Changes in body proportions over the first year of life: Comparisons between "light-for-dates" and "appropriate-for-dates" term infants. Early Hum. Dev. 3:263-265.
Droste, S., J. FitzSimmons, J. Pascoe-Mason et al. 1990. Growth of linear parameters in trisomy 18 fetuses. Am. J. Obstet. Gynecol. 163:158-161.
FitzSimmons, J., S. Droste, T.H. Shepard et al. 1989. Long-bone growth in fetuses with Down syndrome. Am. J. Obstet. Gynecol. 161:1174-1177.
Garn, S.M., and C.G. Rohmann. 1959. "Communalities" of the ossification centers of the hand
and wrist. Am. J. Phys. Anthropol. 17:319-323.
Gam, S.M., and C.G. Rohmann. 1968. "Communalities" in the ossification timing of the growing foot. Am. J. Phys. Anthropol. 24:45-50.
Gam, S.M., M. LaVelle, and B.H. Smith. 1985. Quantification of dysmorphogenesis: Pattern variability index, sz. Am. J. Roentgenol. 144:365-369. Gam, S.M., W.R. Leonard, and A.K. Poznanski. 1987. Applications of the pattern variability index (sz) to the quantification of dysmorphogenesis in the hand. Am. J. Med. Genet. 27:143-152.
Gain, S.M., A.K. Poznanski, and J.M. Nagy. 1971. The operational meaning of maturity criteria. Am. J. Phys. Anthropol. 35:319-326.
Golbus, M.S. 1978. Development in the first half of gestation of genetically abnormal human fetuses. Teratology 18:333-336.
Gruenwald, P. 1970. Fetal malnutrition. In Fetal Growth and Development, H.A. Waisman and G.R. Kerr, eds. New York: McGraw-Hill, 235-239.
Hill, D.E., R.E. Myers, A.B. Holt et al. 1971. Fetal growth retardation produced by experimental placental insufficiency in the rhesus monkey. Biol. Neonate 18:379-394. Honnebier, W.J., and D.F. Swaab. 1973. The influence of anencephaly upon intrauterine growth of fetus and placenta and upon gestation length. Br. J. Obstet. Gynecol. 80:577-588. Kalifa, G., J.P. Barbet, F. Labbe et al. 1989. Value of systematic postmortem radiographic
examinations of fetuses: 400 cases. Pediatr. Radiol. 19:111-113. Killen, B. 1989. Population surveillance of congenital malformations possibilities and limitations. Acta Pediatr. Scand. 78:657-663.
Kerr, G.R. 1972. Nutritional requirements of subhuman primates. Physiol. Rev. 52:415-459. Kerr, G.R., H.A. Waisman, J.A. Allen et al. 1973. Malnutrition studies in Macaca mulatta. II.
The effect on organ size and skeletal growth. Am. J. Clin. Nutr. 26:620-630. Kucera, J., and V. Dolezalova. 1972. Prenatal development of malformed fetuses at 2842 weeks of gestational age (anencephalus, hydrocephalus, Down's syndrome, cleft lip and palate, and hypospadias). I. Weight gains. Biol. Neonate 20:253-261. McBride, M.L., J. Baillie, and B.J. Poland. 1984. Growth parameters in normal fetuses. Teratology 29:185-191.
McFadyen, I.R. 1989. The pathophysiology of intrauterine growth retardation. In Fetal Growth Retardation: Diagnosis and Treatment, A. Kurjak and J.M. Beazley, eds. Boca Raton, FL: CRC Press, 29-51.
Meindl, R.S., and C.O. Lovejoy. 1985. Ecotcranial suture closure: A revised method for the determination of skeletal age at death based on the lateral-anterior sutures. Am. J. Phys. Anthropol. 68:57-66.
Merlob, P., Y. Sivan, and H. Salomon. 1984. Anthropometric measurements of the newborn infant (27 to 41 gestational weeks). In Birth Defects Original Article Series, v. 20, CMV, Pathogenesis, and Prevention of Human Infection, S.A. Plotkin et al., eds. New York: Liss, 1-52.
Miller, H.C., and K. Hassanein. 1971. Diagnosis of impaired fetal growth in newborn infants. Pediatrics 48:511-522.
Naeye, R.L. 1965. Malnutrition: Probable cause of fetal growth retardation. Arch. Pathol. 79:284291.
Newell-Morris, L., and L.H. Tarrant. 1978. Ossification in the hand and foot of the macaque (Macaca nemstrina). I. General features. Am. J. Phys. Anthropol. 48:441-454. Nyberg, D.A., R.G. Resta, D.A. Luthy et al. 1993. Humerus and femur length shortening in the detection of Down's syndrome. Am. J. Obstet. Gynecol. 168:534-538. Pryse-Davies, J., J.H. Smitham, and K.A. Napier. 1974. Factors influencing development of secondary ossification centers in the fetus and newborn. Arch. Dis. Child. 49:425-431. Rotmensch, S., J. Luo, M. Liberati et al. 1992. Fetal humeral length to detect Down syndrome.
Am. J. Obstet. Gynecol. 166:13301334.
Rubin, D.H., P.A. Krasilniroff, J.M. Leventhal et al. 1986. Effect of passive smoking on birth weight. Lancet 2:415417.
Russell, J.G.B. 1973. Radiology in Obstetrics and Antenatal Paediatrics. London, England: Butterworth.
Sherwood, R.J., R.L. May, R.S. Meindl et al. 1992. Growth alteration in the pathological human fetus. Am. J. Phys. Anthropol. 14(suppl.):150.
Sherwood, R.J., H.B. Robinson, R.L. May et al. 1992. Standardized residuals as a means for detection of growth alteration in the pathologic human fetus. Teratology 46:419-427. Wigglesworth, J.S. 1964. Experimental growth retardation in the fetal rat. J. Pathol. Bacteriol. 88:1-13.
Woods, D.L., A.F. Malan, and H.H. Heese. 1979. Patterns of retarded fetal growth. Early Hum. Dev. 3:257-262.
Yerushalmy, J. 1971. The relationship of parent's cigarette smoking to outcome of pregnancy: Implications as to the problem of inferring causation from observed causes. Am. J. Epidemiol. 93:443-456.
' Department of Anthropology, Pennsylvania State University, University Park, PA 16802. Current address: Department of Anthropology, University of Wisconsin, Madison, WI 53706.
2 Department of Pathology, Toledo Hospital, Toledo, OH 43606.
3 Department of Anthropology and Biological Anthropology Program, Division of Biomedical Sciences, Kent State University, Kent, OH 44242.
4 Biological Anthropology Program, Division of Biomedical Sciences, Kent State University, Kent, OH 44242.
Copyright Wayne State University Press Dec 1997
Provided by ProQuest Information and Learning Company. All rights Reserved
Contact
Home
A
Angioedema