Background
Exceeding the Institute of Medicine guidelines for pregnancy weight gain increases childhood and adolescent obesity. However, it is unknown if these effects extend to midlife.
Objective
We sought to determine if exceeding the Institute of Medicine guidelines for pregnancy weight gain increases risk of overweight/obesity in daughters 40 years later.
Study Design
This cohort study is based on adult offspring in the Child Health and Development Studies and the Collaborative Perinatal Project pregnancy cohorts originally enrolled in the 1960s. In 2005 through 2008, 1035 daughters in their 40s were recruited to the Early Determinants of Mammographic Density study. We classified maternal pregnancy weight gain as greater than vs less than or equal to the 2009 clinical guidelines. We used logistic regression to compare the odds ratios of daughters being overweight/obese (body mass index [BMI] ≥25) at a mean age of 44 years between mothers who did not gain or gained more than pregnancy weight gain guidelines, accounting for maternal prepregnant BMI, and daughter body size at birth and childhood. We also examined potential family related confounding through a comparison of sisters using generalized estimating equations, clustered on sibling units and adjusted for maternal age and race.
Results
Mothers who exceeded guidelines for weight gain in pregnancy were more likely to have daughters who were overweight/obese in their 40s (odds ratio [OR], 3.4; 95% confidence interval {CI}, 2.0–5.7). This magnitude of association translates to a relative risk (RR) increase of 50% (RR = 1.5; 95% CI, 1.3-1.6). The association was of the same magnitude when examining only the siblings whose mother exceeded guidelines in 1 pregnancy and did not exceed the guidelines in the other pregnancy. The association was stronger with increasing maternal prepregnancy BMI ( P trend < .001). Compared to mothers with BMI <25 who did not exceed guidelines, the relative risks (RR) for having an overweight/obese adult daughter were 1.3 (95% CI, 1.1-1.7), 1.7 (95% CI, 1.4-2.1) and 1.8 (95% CI, 1.5-2.1), respectively, if mothers exceeded guidelines and their prepregnancy BMI was <25, overweight (BMI 25-<30), or obese (BMI >30). This pattern held irrespective of daughters’ weight status at birth, at age 4 years, or at age 20 years.
Conclusion
Our findings support that obesity prevention before pregnancy and strategies to maintain weight gain during pregnancy within the IOM guidelines might reduce the risk of being overweight in midlife for the offspring.
Introduction
The prevalence of overweight (body mass index [BMI] ≥25-<30 kg/m 2 ) and obesity (BMI ≥30 kg/m 2 ) in US women of childbearing age has increased greatly over the last 30 years. The US Institute of Medicine (IOM) has targeted pregnancy weight gain as a possible modifiable risk factor for reducing obesity in future generations. Earlier versions of the IOM guidelines, established in 1970 and first revised in 1990, were originally designed to prevent small-for-gestational-age children and to make sure that mothers gained enough weight during pregnancy. Their current (2009) guidelines recognize the increasing rates of overweight/obesity in women of childbearing age and provide prepregnancy BMI-specific guidelines. For singleton pregnancies, the IOM recommends that mothers with a BMI of 18.5-<25 kg/m 2 should gain no more than 16 kg. Underweight (BMI <18.5 kg/m 2 ), overweight, and obese mothers should gain no more than 18, 11.5, and 9 kg, respectively.
Observational studies demonstrate that mothers who gain in excess of the current recommended guidelines are more likely to have offspring with higher birthweights and higher childhood BMI than mothers who gain within the recommended ranges. A recent metaanalysis of 7 studies found that pregnancy weight gain outside the ranges recommended by IOM led to a higher risk for childhood overweight/obesity between the ages of 2 and 20 years (combined adjusted odds ratio [OR], 1.43; 95% confidence interval [CI], 1.24–1.65). It is currently unknown, however, whether the association between excessive pregnancy weight gain and offspring’s risk of being overweight/obese extends to midlife. We hypothesized that exceeding guidelines is associated with daughters’ risk of overweight/obesity and that this association is independent of shared environmental or behavioral factors between the mother and daughter that may persist into adulthood. We tested this hypothesis, using both cohort and sibling-based analyses, by applying current IOM pregnancy weight gain guidelines to a cohort of women born in the 1960s, when pregnant women were discouraged from gaining much weight at all (4-6 kg).
Materials and Methods
Study design and population
This study is based on adult daughters who were born in the 1960s to women enrolled in the Child Health and Development Studies (CHDS) and the Collaborative Perinatal Project (CPP) pregnancy cohorts. In 2005 through 2008, 40-50 years after birth, daughters (N = 1134) participated in a new follow-up, Early Determinants of Mammographic Density (EDMD). Pregnant women who received prenatal care at the Kaiser Family Health Plan in Oakland, CA, were eligible for the CHDS, and were recruited in California between 1959 and 1967. Pregnant women, who received prenatal care at one of the teaching hospitals of Harvard Medical School or Brown University, were approached to take part in the New England sites of the CPP (also known as the New England Family Study), which was conducted between 1959 and 1966. We used a subset of the CHDS and CPP based on adult follow-up eligibility (see details of EDMD ). Both studies were designed to study prenatal and familial factors influencing child growth and development. Out of 1134 daughters, 2% were missing data on BMI at midlife and 7% were missing data on maternal pregnancy BMI, leaving a total of 1035 daughters with complete data. In a subsample of pregnancies in the same mother, we also employed a sibling study design, consisting of 247 sibling pairs, to control for family-level confounding. The institutional review boards at Columbia University Medical Center, Kaiser Permanente, Brigham and Women’s Hospital, and Brown University approved the EDMD study. Informed consent was obtained from the mothers in the original cohorts and daughters in the follow-up study.
Data collection
Baseline maternal and childhood data
Maternal interviews collected information on prepregnancy BMI, smoking during pregnancy, and maternal education at registration. Weight gain during pregnancy was abstracted from the records of prenatal visits and was determined as the difference between the last predelivery weight and the first recorded weight. A trained interviewer inquired after the first day of the last menstrual period at the prenatal registration interview to calculate gestational age, which is the time period between the last menstrual period and the day of delivery.
We categorized pregnancy weight gain into a binary variable according to the current IOM clinical guidelines: mothers either did not gain more weight than the upper limit of the BMI-specific guidelines–currently established as 16 kg, 11.5 kg, and 9 kg for women whose prepregnancy BMI is 18.5-<25, 25<30, and >30, respectively–or mothers exceeded the upper limit of the guidelines.
We assessed growth in height and weight measurements at 4 years of age because children were measured at different time points in the CPP and CHDS and these were the common time points between the 2 cohorts. In the CHDS, serial growth measurements were abstracted from medical records. In the CPP, trained clinical staff measured childhood height and weight at 4 years. Because the actual dates of the clinic visit differed for each individual and did not correspond to exactly 4 years, we performed interpolations of height and weight measurements using individual cubic interpolation splines and then calculated BMI percentiles according to the Centers for Disease Control and Prevention growth chart reference percentiles.
Adult daughter interviews
When daughters were in their 40s, we conducted a 45-minute computer-assisted telephone interview, gathering data on personal health, medication, sociodemographic factors, and anthropometric measures. We calculated BMI at age 20 years and current BMI from self-reported height and weight. A subgroup of the daughters participating in the EDMD study also participated in the Early Determinants of Adult Health study where current adult height and weight was measured clinically. Interobserver agreement between self-reported and clinically measured BMI was high (n = 190; kappa = 0.81). The mean and SD of BMI in the subset (27.7 [SD 7.0] kg/m 2 ] was no different from the BMI in the overall cohort (27.2 [SD 6.5] kg/m 2 ) suggesting minimal differential measurement error between the subset and the overall cohort due to body size.
Statistical analysis
We assessed the association between maternal pregnancy weight gain and daughter overweight/obese status (BMI ≥25 kg/m 2 ) at midlife (mean age 44 years, range 39-49 years). We also defined daughter body size categories at 3 previous times in the life course as stratifying variables: high birthweight (>4000 g), overweight/obese (BMI >85th percentile) at age 4 years, and overweight/obese (BMI ≥25 kg/m 2 ) at age 20 years. We adjusted for race/ethnicity, education, age at delivery, smoking during pregnancy, and gestational age, and compared models with and without these potential confounders or intermediary variables. We retained covariates in the models if they affected the key exposure associations (exceeded guidelines) by >10% (none of the variables met criterion), were an independent predictor of the outcome (race/ethnicity), or were based on a priori hypotheses (maternal age).
We use generalized logistic regression to regress daughter overweight/obese status at midlife on maternal weight gain status (exceeded vs did not exceed) both before and after stratifying by prepregnancy BMI. The number of mother/daughter pairs with complete data for prepregnancy BMI, pregnancy weight gain, and BMI was 885. We also ran relative risk regressions for the final models to place in context the magnitude of the odds ratios in terms of relative risks. After the analysis of all daughters, we limited the analysis to daughters who were not high birthweight, or overweight/obese at age 4 or 20 years. We also stratified by parity of the daughter. In sensitivity analyses, we also applied the cut-offs from previous guidelines established before and after 1970.
We further examined potential family-related confounding through a comparison of daughter siblings. Sibling pairs were concordant for pregnancy weight gain if the mother did not exceed or exceeded guidelines in both pregnancies; sibling pairs were discordant if the mother only exceeded guidelines in 1 pregnancy. We used general estimating equations among the discordant siblings adjusted for maternal age and race, and accounted for the shared correlation between siblings, to calculate the OR and 95% CI of being overweight/obese if a mother exceeded the guidelines and was overweight or obese.
Materials and Methods
Study design and population
This study is based on adult daughters who were born in the 1960s to women enrolled in the Child Health and Development Studies (CHDS) and the Collaborative Perinatal Project (CPP) pregnancy cohorts. In 2005 through 2008, 40-50 years after birth, daughters (N = 1134) participated in a new follow-up, Early Determinants of Mammographic Density (EDMD). Pregnant women who received prenatal care at the Kaiser Family Health Plan in Oakland, CA, were eligible for the CHDS, and were recruited in California between 1959 and 1967. Pregnant women, who received prenatal care at one of the teaching hospitals of Harvard Medical School or Brown University, were approached to take part in the New England sites of the CPP (also known as the New England Family Study), which was conducted between 1959 and 1966. We used a subset of the CHDS and CPP based on adult follow-up eligibility (see details of EDMD ). Both studies were designed to study prenatal and familial factors influencing child growth and development. Out of 1134 daughters, 2% were missing data on BMI at midlife and 7% were missing data on maternal pregnancy BMI, leaving a total of 1035 daughters with complete data. In a subsample of pregnancies in the same mother, we also employed a sibling study design, consisting of 247 sibling pairs, to control for family-level confounding. The institutional review boards at Columbia University Medical Center, Kaiser Permanente, Brigham and Women’s Hospital, and Brown University approved the EDMD study. Informed consent was obtained from the mothers in the original cohorts and daughters in the follow-up study.
Data collection
Baseline maternal and childhood data
Maternal interviews collected information on prepregnancy BMI, smoking during pregnancy, and maternal education at registration. Weight gain during pregnancy was abstracted from the records of prenatal visits and was determined as the difference between the last predelivery weight and the first recorded weight. A trained interviewer inquired after the first day of the last menstrual period at the prenatal registration interview to calculate gestational age, which is the time period between the last menstrual period and the day of delivery.
We categorized pregnancy weight gain into a binary variable according to the current IOM clinical guidelines: mothers either did not gain more weight than the upper limit of the BMI-specific guidelines–currently established as 16 kg, 11.5 kg, and 9 kg for women whose prepregnancy BMI is 18.5-<25, 25<30, and >30, respectively–or mothers exceeded the upper limit of the guidelines.
We assessed growth in height and weight measurements at 4 years of age because children were measured at different time points in the CPP and CHDS and these were the common time points between the 2 cohorts. In the CHDS, serial growth measurements were abstracted from medical records. In the CPP, trained clinical staff measured childhood height and weight at 4 years. Because the actual dates of the clinic visit differed for each individual and did not correspond to exactly 4 years, we performed interpolations of height and weight measurements using individual cubic interpolation splines and then calculated BMI percentiles according to the Centers for Disease Control and Prevention growth chart reference percentiles.
Adult daughter interviews
When daughters were in their 40s, we conducted a 45-minute computer-assisted telephone interview, gathering data on personal health, medication, sociodemographic factors, and anthropometric measures. We calculated BMI at age 20 years and current BMI from self-reported height and weight. A subgroup of the daughters participating in the EDMD study also participated in the Early Determinants of Adult Health study where current adult height and weight was measured clinically. Interobserver agreement between self-reported and clinically measured BMI was high (n = 190; kappa = 0.81). The mean and SD of BMI in the subset (27.7 [SD 7.0] kg/m 2 ] was no different from the BMI in the overall cohort (27.2 [SD 6.5] kg/m 2 ) suggesting minimal differential measurement error between the subset and the overall cohort due to body size.
Statistical analysis
We assessed the association between maternal pregnancy weight gain and daughter overweight/obese status (BMI ≥25 kg/m 2 ) at midlife (mean age 44 years, range 39-49 years). We also defined daughter body size categories at 3 previous times in the life course as stratifying variables: high birthweight (>4000 g), overweight/obese (BMI >85th percentile) at age 4 years, and overweight/obese (BMI ≥25 kg/m 2 ) at age 20 years. We adjusted for race/ethnicity, education, age at delivery, smoking during pregnancy, and gestational age, and compared models with and without these potential confounders or intermediary variables. We retained covariates in the models if they affected the key exposure associations (exceeded guidelines) by >10% (none of the variables met criterion), were an independent predictor of the outcome (race/ethnicity), or were based on a priori hypotheses (maternal age).
We use generalized logistic regression to regress daughter overweight/obese status at midlife on maternal weight gain status (exceeded vs did not exceed) both before and after stratifying by prepregnancy BMI. The number of mother/daughter pairs with complete data for prepregnancy BMI, pregnancy weight gain, and BMI was 885. We also ran relative risk regressions for the final models to place in context the magnitude of the odds ratios in terms of relative risks. After the analysis of all daughters, we limited the analysis to daughters who were not high birthweight, or overweight/obese at age 4 or 20 years. We also stratified by parity of the daughter. In sensitivity analyses, we also applied the cut-offs from previous guidelines established before and after 1970.
We further examined potential family-related confounding through a comparison of daughter siblings. Sibling pairs were concordant for pregnancy weight gain if the mother did not exceed or exceeded guidelines in both pregnancies; sibling pairs were discordant if the mother only exceeded guidelines in 1 pregnancy. We used general estimating equations among the discordant siblings adjusted for maternal age and race, and accounted for the shared correlation between siblings, to calculate the OR and 95% CI of being overweight/obese if a mother exceeded the guidelines and was overweight or obese.
Results
Table 1 shows descriptive characteristics of mothers and daughters included in the analysis and compares these characteristics by maternal weight gain status according to the current guidelines. Maternal age at delivery, education, smoking habits, and race/ethnicity were similar between mothers who did and did not exceed guidelines. Compared to mothers with a BMI <25 kg/m 2 , who on average gained 9.8 kg during pregnancy, overweight and obese mothers gained less weight (8.8 kg and 6.6 kg, respectively; data not shown). However, because BMI-specific guidelines are lower for overweight and obese women, mothers who exceeded guidelines had higher mean prepregnancy BMI compared to mothers who did not. Mothers who exceeded guidelines also had longer gestations and daughters with higher birthweights, birth lengths, and BMIs at ages 4 and 20 years. These characteristics of the cohort are similar to national data of women during the 1960s.
| Maternal characteristics | Did not exceed IOM guidelines, n = 942 | Exceeded IOM guidelines, n = 103 | P value | ||
|---|---|---|---|---|---|
| Mean or % | SD | Mean or % | SD | ||
| Prepregnancy BMI, kg/m 2 | 26.1 | ±3.5 | 26.9 | ±4.3 | <.0001 |
| <25 | 81% | 34% | |||
| 25–<30 | 15% | 44% | |||
| ≥30 | 4% | 22% | |||
| Age at delivery, y | 26.1 | ±5.7 | 25.7 | ±5.6 | .48 |
| Education (more than high school) | 35% | 27% | .22 | ||
| Smoking during pregnancy (yes vs no) | 41% | 33% | .18 | ||
| Race/ethnicity | |||||
| Non-Hispanic white | 83% | 79% | .52 | ||
| Non-Hispanic black | 11% | 17% | .31 | ||
| Hispanic, Asian, Native American, other | 6% | 4% | .70 | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
