LEAD is of the more extensively studied reproductive toxicants. (1) Lead readily crosses the placenta; therefore, it may cause fetal and maternal adverse outcomes, including low birth weight, preterm delivery, and congenital anomalies. (2-4) Lead concentrations have markedly declined in the last 2 decades because there have been concerted public health interventions; however, the potential for adverse maternal and fetal outcomes at lower blood lead concentrations remains unknown. (1,3,4) The identification of these risks requires multiple measurements of lead during pregnancy--a time when blood lead levels change. (5) A study of first trimester elective abortion tissues suggested that lead does not accumulate in the embryo/fetus during the first trimester; lead begins to cross the placenta during mid-pregnancy in animal models. (4) In the current study, we examined the relationship between adverse maternal events or newborn outcomes and low circulating blood lead concentrations in a nonselected population of women.
Materials and Method
The population comprised 705 enrollees who presented for prenatal care at 3 clinics that serve the Medicaid population in Camden, New Jersey. Ethnically, the study population, aged 12-34 yr, was 42% African American, 19% Caucasian, and 38% Hispanic, thus reflecting the racial and ethnic mix of the community. Exclusion criteria included a history of serious nonobstetric conditions (e.g., chronic hypertension, diabetes type I or II, drug or alcohol abuse, lupus, malignancies, seizure disorders). All participants provided written informed consent consistent with the policies of the Institutional Review Boards at the University of Medicine and Dentistry of New Jersey and the University of Michigan.
Data and blood samples were collected during each trimester of pregnancy, at delivery, and at the first postpartum visit. In addition to physical assessment and phlebotomy, each participant was interviewed at these times to determine her sociodemographic characteristics and medical history.
Blood lead concentrations. Whole blood was collected by venipuncture into trace-metal-free, heparinized vacutainer[R] tubes (Becton Dickinson [Rutherford, New Jersey]). Samples were mixed on a blood rocker for 2 min and aliquoted into prerinsed cryovials for storage at-70[degrees]C until analysis. Lead concentrations were determined in accordance with the method of Miller et al. (6) at 283.3 nm, and a model Z3030 Zeeman background-corrected atomic absorption spectrophotometer (Perkin-Elmer [Norwalk, Connecticut]) was used. Each analytical run was accompanied by the analysis of "Reference Material for Whole Blood" lots #039 and #041 (Wadsworth Laboratory, State of New York Department of Health). These reference materials were analyzed at the beginning and end of each analytical run, which typically included 27 blood samples. The mean and standard error (SE) for certified lead concentrations were 7.1 [+ or -] 0.5 [micro]g/dl and 3.6 [+ or -] 0.2 [micro]g/dl for lots #39 and #041, respectively. Analysis of these reference materials during analytical runs for this and other studies resulted in mean lead concentrations of 6.98 [+ or -] 0.59 [micro]g/dl and 3.60 [+ or -] 0.21 [micro]g/dl at the beginning and 7.00 [+ or -] 0.57 [micro]g/dl and 3.56 [+ or -] 0.29 [micro]g/dl at the end of each run (n = 66-104). We assessed instrument drift, which was less than 1.5%, by determining the within-run mean for each control and the departure from the mean for each analytical result. Our laboratory participates in the Centers for Disease Control and Prevention's (CDC's) "Blood Lead Laboratory Reference System" program and obtains results within [+ or -] 2% of the CDC's target values.
Outcomes. Diagnoses of hypertension in pregnancy (HIP), preeclampsia, or toxemia were abstracted from medical records and were based on blood pressure cutoff points of 140 (systole) and 90 (diastole). Preterm delivery was defined as the occurrence of birth prior to 37 wk gestation, as determined by ultrasound. Small birth weight for gestational-age neonates (SGA) were distinguished from infants of normal size in accordance with the methods of Brenner, (7) including adjustment for ethnicity, infant gender, and parity. Low birth weight (LBW) was defined as a newborn weight of less than 2,500 gm. Apgar scores, including values assessed at 1 min and 5 min, were dichotomized as 1-8 vs. 9-10, respectively.
Covariates. Usual dietary calcium intakes (mg/d) and energy intakes were averaged from multiple 24-hr food recalls administered during each trimester. Data were analyzed as the lowest quartile of dietary intake vs. the 3 higher quartiles of intake. Pregravid body mass index, calculated as weight (kg) divided by height ([m.sup.2]), was used as a continuous variable or was trichotomized as (1) < 19.8 kg/[m.sup.2] (underweight), (2) 19.8-26 kg/[m.sup.2] (normal weight), or (3) > 26 kg/[m.sup.2] (overweight). (8) Weight gain during pregnancy was the difference between self-reported pregravid weight and weight at delivery. Parity was a dichotomous variable that described whether women were nulliparous or parous. Growth during pregnancy was determined by measurement of the difference in knee height at (1) the time of entry into medical care and (2) the first postpartum visit. (9) Women who were 19 yr of age or younger and who had 1 or more mm of growth were classified as growers. Participants were classified as (1) smokers if they smoked at the time medical care was first administered through to the first postpartum visit and (2) as pot users if they reported smoking marijuana at least once each week during pregnancy.
Data analysis. All statistical procedures in this analysis were performed with SAS (ver. 6.12) (SAS Institute [Cary, North Carolina]). Univariate statistics were calculated for continuous variables and frequencies were calculated for categorical variables. Variables were evaluated for the normality of their distributions or for outliers, and log transformations were undertaken for lead concentration and weight change. (Unless noted otherwise, we used the method of Whitaker and Lind (10) to adjust blood lead concentrations for plasma volume.)
Data were analyzed cross-sectionally (i.e., within-trimester and at delivery). We used analysis of covariance to compare the mean blood lead concentrations and SEs for groups at each examination contact. We used longitudinal regression analyses with mixed models to investigate the pattern of repeated maternal lead measurements over time and to relate those patterns to the selected maternal and infant outcomes, or to maternal behaviors and characteristics. The explanatory variable "weeks since last menstrual period (LMP)" represented time in all longitudinal models. "Time since LMP" was modeled as a random effect, and the intercept (referring to the time immediately following conception) was modeled as a random intercept. Variables identified as consistently important in the cross-sectional analyses were entered into the longitudinal analyses as main effects. Biologically plausible interaction terms were added to the model, and those interaction terms with a significance level of p [less than or equal to] .10 were candidates for the final model. In the final model, nonsignificant main effects were eliminated if the variable was not a component of a significant interaction term. We have retained the [beta] coefficients and SEs in the longitudinal models in their transformed state because the SE measure involved too much complexity.
Results
White women were older, smoked more, and were less likely to have dietary calcium intakes in the lowest quartile of distribution than black and Hispanic women (Table 1). Maternal blood lead concentrations, at each trimester and at delivery, are shown in Table 2. The average blood lead concentration in pregnancy was 1.2 [micro]g/dl (SE = [+ or -] 0.03), which is comparable with data reported in recent national studies? Seventeen women had blood lead concentrations that exceeded 4 [micro]g/dl, and of these, 9 (1.2%) women had provided repeat samples, but the blood lead concentrations nevertheless exceeded 4 [micro]g/dl.
