Muin J. Khoury

Residential mobility during pregnancy: Implications for environmental teratogenesis

Although most teratogens are suspected to act early in the first trimester of pregnancy, birth defects monitoring programs and etiologic studies usually use residence at birth as a proxy measure for residence in the first trimester in searching for environmental teratogens. Because of the high mobility of the U.S. population, residence misclassification can potentially alter inferences concerning environmental teratogens. To evaluate this potential bias, data from the population-based Maryland Birth Defects Reporting and Information System were analyzed. In 1984, the system ascertained 295 infants with one or more of 12 sentinel defects. Of these cases, 59 (20%) mothers reported they have changed address between the time of conception and the time of birth, and 22 have moved to a different county. The residential mobility rate varied by demographic variables and was highest among white women, in the age group 20-24 years. If residence at birth is used as a screening test for residence at conception, it can be shown that in the presence of an environmental teratogenic exposure, misclassification of exposure increases with increasing mobility rate, and population exposure frequency. Such misclassification tends to weaken associations between residence and birth defects and may lead to missing environmental teratogens. This analysis emphasizes the need to use residence information early in pregnancy rather than exclusively at birth.

Differences in the epidemiology of prematurity and intrauterine growth retardation

Although both preterm (PT) and intrauterine growth retarded (IUGR) infants can have similar birth weights, they are known to show different neonatal and post-neonatal features. Newborns (n = 623) from the Guatemalan longitudinal study of nutritional supplementation during pregnancy were studied. There were 61 PT (less than or equal to 37 wk) and 173 IUGR (less than or equal to 10th percentile) infants. Simultaneous adjustment using the long-linear model showed that calorie and/or protein supplementation during pregnancy lowered the risk of PT (adjusted odd ratios (OR) = 0.52, 95%, CI = 0.40-0.77, and 0.43 CI = 0.36-0.59, respectively), but did not affect the incidence of IUGR. Low maternal head circumference and weight increased the risk of IUGR only (OR = 1.4, CI = 1.02-1.8 and 2.3 CI = 1.8-2.7, respectively). Male fetuses were at higher risk of both PT and IUGR. These data confirm the differential effect of maternal characteristics and nutritional supplementation during pregnancy on both PT and IUGR, and strongly suggest the need to include both gestational age and birth weight as outcome measures in epidemiological studies, thus avoiding the exclusive use of LBW (less than or equal to 2500 g).

Genetic heterogeneity of prematurity and intrauterine growth of retardation: Clues from the Old Order Amish

We studied differences in the role of genetic factors in prematurity and intrauterine growth retardation with the use of data on 312 Amish singleton live children ascertained from Amish records in Lancaster county, Pennsylvania, between 1969 and 1980. Birth and death certificates were obtained on all children, and inbreeding coefficients of child, mother, and father were computed by use of the path method of tracing common ancestors in a unique genealogic registry of Amish ancestors dating back to the 1700s. Multivariate analysis with linear and log linear models showed that a lower mean gestational age and a higher risk of prematurity (less than 37 weeks) and borderline maturity (37 to 38 weeks) were significantly associated with increased maternal inbreeding but not child or paternal inbreeding. On the other hand, a higher risk of intrauterine growth retardation (less than the tenth percentile in birth weight for gestational age) and mild intrauterine growth delay (tenth to twenty-fifth percentile) were associated with increased child inbreeding but not maternal or paternal inbreeding. The analysis suggests the presence of genetic heterogeneity in the etiology of prematurity and intrauterine growth retardation; while prematurity is mostly related to the maternal genotype, intrauterine growth retardation is related to the fetal genotype. The study reemphasizes the need for separating low birth weight into prematurity and intrauterine growth retardation in genetic and epidemiologic studies.

Epidemiological approach to the evaluation of genetic screening in the workplace.

Using several examples of genetic marker and disease associations in the workplace, the authors have applied formulas to estimate the sensitivity, specificity, and positive predictive value (PPV) of screening for these markers. Sensitivity, specificity, and PPV are affected independently by characteristics of the population being screened, ie, the genetic marker frequency, the disease frequency, and the magnitude of the relative risk (R). For a given disease frequency, when the genetic marker is less frequent than the disease, PPV increases with relative risk, although sensitivity remains low. When the genetic marker is more frequent than the disease, PPV remains low while sensitivity increases with R. When marker and disease frequencies are equal, PPV and sensitivity are equal and increase with R. However, when the disease frequency is very low, R must approach 100 before PPV or sensitivity approaches 50%. These relationships may be used effectively in the decision whether to implement a screening program in the workplace.

INBREEDING AND PREREPRODUCTIVE MORTALITY IN THE OLD ORDER AMISH. II. GENEALOGIC EPIDEMIOLOGY OF PREREPRODUCTIVE MORTALITY

The effects of offspring and parental inbreeding on prereproductive mortality (death before age 20 years) in the historical population of the Lancaster County, Pennsylvania, Old Order Amish were investigated using the Amish genealogic registry, which contains information on 42,465 births dating to the time of the pioneer migrants in the 1700s. Inbreeding coefficients for offspring and parents were computed using the path method of tracing common ancestors in the multigenerational pedigrees. In this population, prereproductive mortality declined from about 15% in the late 1800s to about 5% after 1930. Offspring inbreeding was found to be an independent predictor of prereproductive mortality after multivariate adjustment for demographic risk factors for mortality. Moreover, the higher the coefficient, the higher the relative risk of prereproductive death, and the higher the risk of multiple deaths in the same sibship. There was no evidence of declining inbreeding effects over 10 generations of continuous inbreeding, nor of any significant parental inbreeding effects. Because of the high levels of inbreeding, it could be shown that inbreeding accounts for about 40% of all prereproductive deaths in the present population. Genetic load analysis showed an average of about 1.7 lethal equivalents and a mostly mutational load.

MILK DRINKING AND POSSIBLE PROTECTION OF THE RESPIRATORY EPITHELIUM

In the Hopkins investigation, detailed interviews as well as spirometry were obtained on 2539 non-patient adult participants. The interviews included questions regarding smoking habits, family history, socioeconomic status, respiratory symptoms, certain dietary factors and beverage consumption. For the analyses of risk factors, all patients were excluded from the original study population, which consisted not only of several groups of patients along with their relatives, but also neighborhood controls. teachers, and other groups. Thus, only those subjects over 20 years of age who were not ascertained on the basis of their own health status were considered. CB was identified by the report of cough and phlegm production for three or more months per year for two consecutive years. In this population, a significantly lower prevalence of CB was found among study subjects who drank milk. CB was found in 17.7% of 351 individuals who did not drink milk, but in only 11.7% of 1196 adults who drank milk daily (p < 0.01). This significant difference persisted after binary multiple regression adjustment for cigarette smoking, one-second forced expiratory volume and other potentially confounding variables: age, sex, race, socio-economic status, education, alcohol and coffee intake [9]. However, milk drinking was not associated with a reduction in CB prevalence among those who had never been smokers. Of nonsmoking adults who were not milk drinkers, 6.0% had CB, not significantly different from the CB prevalence of daily milk drinkers (6.7%). In contrast,