The Collaborative Perinatal Project enrolled women (55,000 pregnancies) at the first prenatal visit at 12 U.S. sites (1959-1965).25
Women were followed prospectively through delivery, and their offspring were followed to age 7 or 8. We refer to the pregnant women enrolled in the study as “mothers” and their offspring as “daughters.” A sample of daughters from two sites was recontacted in 1987-1991 when they were ages 22-32 years. Study staff interviewed 627 daughters (the subjects for the present study) who reported having had at least one pregnancy lasting through at least 20 weeks’ gestation. The design and methods of this follow-up study have been described previously. 26
This study was ruled exempt from Institutional Review Board review by the NIH Office of Human Subjects Research.
Interviewers recorded a health history of mothers at the first prenatal visit, including the age (in years) at menarche, number and outcome of all prior pregnancies, and pre-pregnancy weight. Maternal parity was defined as the number of non-aborted pregnancies of 20+ weeks’ gestation that occurred before the conception of the daughter in the present study. Parity categories were defined as 0, 1-2, and 3+ prior pregnancies. Maternal age at menarche was categorized by quartiles (< 12, 12, 13, > 13 years).
Maternal pre-pregnancy BMI (weight [kg] / height [m2
]) was calculated based on height measured at the first visit and self-reported pre-pregnancy weight. Categories of underweight or normal weight (BMI < 25), overweight (BMI = 25.0-29.9) and obese (BMI ≥ 30) were constructed. A continuous socioeconomic index, developed by the Census Bureau, was applied that incorporated education level and occupation of the head of household and family incomes.27,28
Birthweight and gestational age of the daughters were measured at birth, and daughters participated in a visit around the age of 7 years.26
At the later visit, height was measured to the nearest 0.5 centimeters using a standardized backboard, and weight was recorded in pounds to the nearest 0.25 pounds or grams to the nearest 100 grams using scales calibrated semiannually.26
For this analysis, daughters’ childhood BMI-for-age z-scores were calculated based on the Centers for Disease Control and Prevention growth charts for daughters who were aged 6 years, 9 months to 8 years, 3 months at the age 7 visit.29
We excluded from our mediator analysis involving BMI at age 7, those daughters who missed the age 7 study visit, were missing anthropometrics, or attended the visit when they were older than 8 years, 3 months (n=103; 17.3%).
Daughters’ age at menarche is the outcome of interest for this study. As part of the follow-up study of daughters during adulthood, interviewers asked, “how old were you when you had your first period?” and “do you remember what grade you were in?”. We relied on the stated age for 617 daughters. Four daughters provided a grade but no age, and we converted seventh grade to 12 years old and eighth grade to 13 years old. Six daughters were missing both age and grade information. Thus, 621 daughters (99 percent) had menarcheal age reported. Daughters were also asked to identify their race. When missing, race recorded during childhood was used. Analytic categories of race were established as Black and all others.
There were 597 of 627 interviewed daughters (95 percent) with complete data for age at menarche and their mother’s pre-pregnancy BMI. These were included in the main analysis and a mediator analysis involving SGA status. The 494 daughters with a calculable BMI at age 7 were included in a mediator analysis involving childhood BMI.
We used SAS version 9.1 for all statistical analyses.30
Associations between covariates and the exposure and outcome were evaluated using the mean score statistic. We used polytomous logistic regression to examine the relationship between daughters’ age at menarche and maternal pre-pregnancy BMI. The categories of maternal obesity and overweight were compared with a reference of normal/underweight. Socioeconomic index, maternal parity, maternal age at menarche, and daughter’s race were identified as potential confounders a priori and were included in the adjusted models. No covariates were found to be notable effect modifiers. Because of the sampling procedures, a variable for study site was included in models. We conducted a mediator analysis to examine to what extent BMI-for-age Z-score at age 7 or being born SGA might mediate the association between maternal BMI and age at menarche.31
To test for mediation by BMI-for-age Z-score at age 7, we built 3 regression equations: BMI-for-age Z-score on maternal BMI, menarcheal age on maternal BMI, and menarcheal age on both maternal BMI and BMI-for-age Z-score. BMI-for-age Z-score would be a mediator if all of the following hold in the expected direction: maternal BMI is associated with BMI-for-age Z-score, maternal BMI is associated with daughter’s menarcheal age, and BMI-for-age Z-score is associated with daughters’ menarcheal age (in the third equation), plus the association between maternal BMI and menarcheal age is attenuated when BMI-for-age Z-score is included in the model. We repeated this to examine SGA as a potential mediator.