Log-linear allometry of fetal craniofacial growth in Down's syndrome.

Trisomy 21 develops as a result of nondisjunction of two homologous chromosomes during either the first or second meiotic division. One of the more important consequences of these genetic alterations is the predictable, although variable disturbance in the architecture of the craniofacial region [1]. Postnatal craniofacial morphology has been extensively studied in Down's syndrome (DS). However, little information is available on human prenatal development of the head and face in such patients. The time at which changes in craniofacial phenotype first emerge in Down's syndrome fetuses and at which physical growth begins to diverge from normal is unknown. To explore these questions, we compared prenatal craniofacial growth in 50 Down's syndrome fetuses with that of 555 fetuses judged to be "typical for body weight and age" using the method of log-linear allometry [2].

Log-linear allometry of normal fetal craniofacial growth.

Normative data on human craniofacial growth during the fetal period and important to provide a broader perspective on normal morphogenesis as well as to serve as reference for analyzing craniofacial syndromes in which growth has gone awry. Over a 19-year period, the Teratology Unit at the University of Michigan Medical Center has collected data on 2,568 legally donated fetuses that have undergone necropsy examination at various gestational ages. From previous analyses, 609 of the total fetal population (25%) were designated as typical for age or body weight on the basis of normal morphology, absence of maceration, and general growth symmetry. Of the 609 fetuses reviewed, 54 were excluded secondary to incomplete data. The remaining 555 constitute the basis of this study. Seven craniofacial measurements were recorded, including head circumference (HC), brain weight, inner canthal and outer canthal distances, and distances from nasion to menton, outer canthus to tragus and auditory meatus to vertex. Statistical analysis was carried out using the single-factor allometric model of Sewall Wright. Size was estimated as the first unstandardized principal component of the logarithms of lengths and of cube roots of weights, and then allometry was expressed in the regressions of each log variable on size. Significant allometry was found as were significant differences in errors about the allometric relation, but no evidence for more than a single factor or of "nonlinearity" in the regression curves was noted. Although there were differences of specific allometric coefficients between the various measurements (i.e., the slope of the curve for IC was significantly smaller than the slope of the curve generated for HC), these specific growth rates remain in relatively strict proportion to one another from early in gestation (body weight, 54.2 gm) to later in gestation (body weight, 1,000 gm).