Materials and Methods
We used data from the PRAMS, an ongoing, state-representative, population-based surveillance system of the Centers for Disease Control and Prevention (CDC) and state health departments. PRAMS collects information in participating states about maternal behaviors and experiences before, during, and after pregnancies resulting in live infants. In each participating site, PRAMS uses birth certificates to draw a stratified sample of live births, and oversamples certain high-risk populations. Self-administered questionnaires are mailed to the mothers’ homes, with telephone follow-up for nonresponders. Data from maternal questionnaires are linked to the data from the child’s birth certificate. Data are weighted to account for sample design, nonresponse, and noncoverage. More detail on PRAMS methodology is available at http://www.cdc.gov/prams/methodology.htm .
We used 2000 through 2009 data from states that met the established PRAMS response rate threshold of ≥70% from 2000 through 2006, or ≥65% from 2007 through 2009. Year of infant birth, 2000 through 2009, was categorized into 2-year increments for this analysis (eg, 2000 through 2001, 2002 through 2003) to maximize the number of states eligible for inclusion in this analysis. Fourteen states met the response rate threshold criteria for at least 1 year in each 2-year increment from 2000 through 2009 (Alaska, Arkansas, Colorado, Hawaii, Maine, Maryland, Michigan, Nebraska, New York, Ohio, Oklahoma, Utah, Washington, and West Virginia). We included women who had a singleton live birth and were ≥18 years of age. We limited the analysis to women with full-term infants (37-41 weeks and 6 days’ gestation) (n = 147,706) and conducted sensitivity analyses among women delivering infants at 39-40 weeks to limit confounding associated with pregnancy duration. Respondents were excluded if they had missing data on weight gain (5.9%) or prepregnancy BMI (4.8%), extreme values for BMI (<12 or >75 kg/m 2 ) (n = 48) or missing data on ≥1 covariates (9.4%). In total, 15.8% of respondents (n = 23,358) were excluded, resulting in a final sample size of 124,348 women. Mean infant age at time of PRAMS survey completion for women in this analysis was 112.6 days (SE 0.21). Compared to women included in the full analytic sample, women excluded due to missing data or extreme values were younger, less educated, less likely to gain above IOM GWG recommendations, less likely to smoke during pregnancy, less likely to report nausea during pregnancy, more likely to be a racial and ethnic minority, more likely to be Medicaid insured at delivery, more likely to have ≥1 previous births, and more likely to have gestational or preexisting diabetes (χ 2 P < .05 for all).
We used birth certificate data to categorize maternal race-ethnicity as: non-Hispanic white, non-Hispanic black, Hispanic, Alaska Native, American Indian, Asian/Pacific Islander, and other (women reporting mixed race or any race-ethnicity other than those described above). Using birth certificate data, we categorized self-reported age (18-19; 20-24; 25-29; 30-34; ≥35 years), education (less than high school; high school; greater than high school), parity (no previous birth; ≥1 previous births), gestational or preexisting hypertension (yes/no), and gestational or preexisting diabetes (yes/no). PRAMS questionnaires provided self-reported data on Medicaid coverage at delivery (yes/no), prenatal smoking (smoker throughout pregnancy; quit smoking before third trimester; nonsmoker), and nausea during pregnancy (yes/no).
The outcome for this analysis, self-reported GWG, was obtained from the birth certificate and modeled 2 ways: continuous GWG in pounds and as a categorical variable according to 1990 IOM GWG recommendations based on the woman’s prepregnancy BMI. Prepregnancy BMI was calculated as (weight in kilograms)/(height in meters) 2 , using self-reported height and weight from PRAMS questionnaires, and categorized according to the current WHO guidelines. A woman was classified as gaining below, within, or above 1990 IOM recommendations based on her prepregnancy BMI. Weight gain within recommendations was defined as: 28-40 lb for underweight women (BMI <18.5 kg/m 2 ); 25-35 lb for women with a normal BMI (18.5 ≤ BMI <25 kg/m 2 ); 15-25 lb for overweight women (25 ≤ BMI <30 kg/m 2 ); and 15-25 lb for obese women (BMI ≥30 kg/m 2 ). For obese women, we used the maximum GWG of 25 lb recommended for overweight women because no maximum weight gain allowance was established for obese women in the 1990 IOM recommendations.
We calculated the mean and SE for GWG and the weighted prevalence and SE for 1990 IOM recommended GWG groups (below, within, and above) and for maternal and pregnancy characteristics. All estimates were calculated overall (2000 through 2009 combined) and by 2-year increments from 2000 through 2009. We used linear regression (for mean) and logistic regression (for categorical variables) models to examine trends in weight gain and in maternal and pregnancy characteristics. We conducted similar analyses on mean GWG and the prevalence of GWG below, within, and above IOM recommendations stratified by prepregnancy BMI. To estimate the magnitude of change in the prevalence estimates for statistically significant trends in GWG groups (below, within, above recommendations), the biennial percentage point change was estimated from the beta coefficient of the infant’s birth year. Lastly, we examined the adjusted trend from 2000 through 2009 in mean GWG using linear regression, with year of infant birth as the independent variable and adjusted for all maternal and pregnancy characteristics. Similarly, we used multivariable logistic regression to generate adjusted odds ratios (aOR) and 95% confidence intervals (CI) for gaining within (yes/no) IOM recommendations for GWG for each 2-year increment (2002 through 2003, 2004 through 2005, 2006 through 2007, and 2008 through 2009), compared with the reference group, 2000 through 2001. For all analyses we considered a P < .05 as statistically significant. All analyses were conducted with SAS 9.2 (SAS Institute, Cary, NC) and SUDAAN 10.0.1 (RTI International, Research Triangle Park, NC) to account for the PRAMS complex survey design and weighted to reflect population estimates. The CDC Institutional Review Board approved the PRAMS protocol, and all participating states approved the analysis plan for the study.
Results
Women in the sample had an overall mean GWG of 31.3 lb. Approximately 35.8% of women had GWG within IOM recommendations, 44.4% gained above recommendations, and 19.8% gained below recommendations ( Table 1 ). Of the 124,348 women in the final sample, the majority had a normal prepregnancy BMI (53.4%), were <30 years of age (61.3%), were non-Hispanic white (74.3%), had a post-high school education (57.7%), were multiparous (60.8%), were not enrolled in Medicaid at delivery (64.5%), were nonsmokers before pregnancy (74.7%), did not report preexisting or gestational diabetes (96.3%) or hypertension (95.4%), and reported nausea during pregnancy (73.1%).
| Characteristic | Year categories | ||||||
|---|---|---|---|---|---|---|---|
| Overall 2000 through 2009 n = 124,348 | 2000 through 2001 n = 24,118 | 2002 through 2003 n = 25,726 | 2004 through 2005 n = 24,346 | 2006 through 2007 n = 25,254 | 2008 through 2009 n = 24,904 | Trend 2000 through 2009 P value a | |
| % (SE) b | |||||||
| Mean gestational weight gain, lb c | 31.3 (0.06) | 31.2 (0.14) | 31.4 (0.13) | 31.2 (0.14) | 31.3 (0.14) | 31.4 (0.14) | .46 |
| IOM d | |||||||
| Below recommended guidelines | 19.8 (0.18) | 20.0 (0.42) | 19.3 (0.38) | 19.3 (0.39) | 20.3 (0.39) | 20.3 (0.38) | .14 |
| Within recommended guidelines | 35.8 (0.21) | 37.5 (0.52) | 37.0 (0.47) | 36.2 (0.48) | 34.3 (0.46) | 34.2 (0.45) | < .01 |
| Above recommended guidelines | 44.4 (0.22) | 42.5 (0.53) | 43.7 (0.48) | 44.5 (0.50) | 45.4 (0.49) | 45.5 (0.48) | < .01 |
| Prepregnancy body mass index | |||||||
| Underweight (<18.5 kg/m 2 ) | 4.7 (0.09) | 5.8 (0.25) | 5.3 (0.22) | 4.3 (0.20) | 4.2 (0.19) | 3.9 (0.18) | < .01 |
| Normal (18.5–24.9 kg/m 2 ) | 53.4 (0.22) | 55.6 (0.53) | 54.4 (0.48) | 53.3 (0.50) | 52.1 (0.49) | 51.8 (0.48) | < .01 |
| Overweight (25.0–29.9 kg/m 2 ) | 23.3 (0.19) | 21.7 (0.44) | 22.7 (0.40) | 22.9 (0.42) | 24.7 (0.43) | 24.3 (0.42) | < .01 |
| Class I obesity (30.0–34.9 kg/m 2 ) | 10.8 (0.14) | 9.8 (0.32) | 10.7 (0.31) | 11.3 (0.32) | 10.9 (0.30) | 11.2 (0.30) | < .01 |
| Class II obesity (35.0–39.9 kg/m 2 ) | 4.8 (0.10) | 4.4 (0.23) | 4.3 (0.20) | 5.0 (0.23) | 4.8 (0.21) | 5.2 (0.22) | < .01 |
| Class III obesity (≥40 kg/m 2 ) | 3.1 (0.08) | 2.8 (0.18) | 2.6 (0.15) | 3.2 (0.19) | 3.3 (0.18) | 3.5 (0.18) | < .01 |
| Age, y | |||||||
| 18-19 | 6.7 (0.12) | 7.2 (0.28) | 6.5 (0.25) | 6.2 (0.26) | 7.1 (0.27) | 6.6 (0.25) | .51 |
| 20-24 | 24.7 (0.19) | 24.9 (0.46) | 25.3 (0.42) | 24.9 (0.43) | 24.7 (0.42) | 24.0 (0.41) | .07 |
| 25-29 | 29.9 (0.20) | 29.2 (0.48) | 28.4 (0.44) | 29.5 (0.46) | 30.6 (0.45) | 31.6 (0.45) | < .01 |
| 30-34 | 24.5 (0.19) | 25.2 (0.47) | 25.6 (0.42) | 24.2 (0.43) | 23.7 (0.41) | 24.1 (0.41) | < .01 |
| ≥35 | 14.2 (0.15) | 13.5 (0.36) | 14.3 (0.33) | 15.1 (0.35) | 14.0 (0.32) | 13.8 (0.31) | .93 |
| Race/ethnicity | |||||||
| White, non-Hispanic | 74.3 (0.17) | 77.5 (0.38) | 76.0 (0.36) | 74.3 (0.42) | 73.3 (0.37) | 71.2 (0.37) | < .01 |
| Black, non-Hispanic | 9.6 (0.12) | 8.4 (0.24) | 10.1 (0.26) | 9.2 (0.31) | 9.7 (0.24) | 10.6 (0.24) | < .01 |
| Hispanic | 9.2 (0.12) | 8.2 (0.27) | 8.2 (0.23) | 9.7 (0.28) | 9.8 (0.27) | 9.8 (0.25) | < .01 |
| American Indian/Alaska Native | 1.5 (0.04) | 1.5 (0.10) | 1.3 (0.07) | 1.5 (0.09) | 1.6 (0.09) | 1.6 (0.08) | .19 |
| Asian/Pacific Islander | 4.5 (0.07) | 4.2 (0.16) | 4.3 (0.16) | 4.4 (0.15) | 4.7 (0.16) | 4.8 (0.15) | < .01 |
| Other | 0.9 (0.05) | 0.2 (0.05) | 0.2 (0.04) | 1.1 (0.12) | 1.0 (0.10) | 2.0 (0.14) | < .01 |
| Education, y | |||||||
| <12 | 12.0 (0.15) | 11.7 (0.36) | 12.5 (0.34) | 11.9 (0.35) | 12.3 (0.35) | 11.6 (0.32) | .56 |
| 12 | 30.3 (0.20) | 33.2 (0.51) | 31.9 (0.45) | 30.6 (0.46) | 28.9 (0.44) | 27.3 (0.43) | < .01 |
| >12 | 57.7 (0.22) | 55.1 (0.53) | 55.6 (0.48) | 57.5 (0.49) | 58.8 (0.48) | 61.2 (0.47) | < .01 |
| Parity | |||||||
| 0 | 39.2 (0.22) | 38.9 (0.52) | 39.0 (0.47) | 38.7 (0.49) | 39.9 (0.48) | 39.3 (0.47) | .30 |
| ≥1 | 60.8 (0.22) | 61.1 (0.52) | 61.0 (0.47) | 61.3 (0.49) | 60.1 (0.48) | 60.7 (0.47) | .30 |
| Insurance coverage at delivery | |||||||
| Medicaid | 35.5 (0.21) | 28.4 (0.47) | 32.2 (0.45) | 37.2 (0.49) | 38.1 (0.47) | 40.5 (0.47) | < .01 |
| Other insurance/uninsured | 64.5 (0.21) | 71.6 (0.47) | 67.8 (0.45) | 62.8 (0.49) | 61.9 (0.47) | 59.5 (0.47) | < .01 |
| Prenatal smoking | |||||||
| Smoker throughout pregnancy | 14.2 (0.16) | 14.7 (0.40) | 14.1 (0.35) | 14.6 (0.37) | 13.7 (0.35) | 13.9 (0.35) | .08 |
| Quit smoking before third trimester of pregnancy | 11.2 (0.14) | 10.6 (0.34) | 11.1 (0.31) | 10.8 (0.32) | 11.5 (0.32) | 11.8 (0.32) | < .01 |
| Nonsmoker | 74.7 (0.20) | 74.7 (0.48) | 74.9 (0.43) | 74.6 (0.45) | 74.8 (0.43) | 74.3 (0.43) | .54 |
| Gestational or preexisting hypertension | 4.6 (0.09) | 4.3 (0.20) | 4.7 (0.20) | 4.8 (0.21) | 4.5 (0.19) | 4.6 (0.20) | .61 |
| Gestational or preexisting diabetes | 3.7 (0.08) | 3.3 (0.19) | 3.5 (0.18) | 3.4 (0.18) | 4.3 (0.20) | 4.1 (0.19) | < .01 |
| Nausea during pregnancy | 73.1 (0.20) | 75.0 (0.47) | 75.1 (0.42) | 71.2 (0.46) | 72.1 (0.44) | 72.5 (0.43) | < .01 |
a P values for trend generated from unadjusted regression models
c Mean gestational weight gain trend indicated by pounds (SE)
d Gaining within IOM guidelines: 28–40 lb for underweight women (BMI <18.5 kg/m 2 ); 25–35 lb for normal-weight women (18.5 ≤ BMI <25 kg/m 2 ); 15–25 lb for overweight women (25 ≤ BMI <30 kg/m 2 ); and 15–25 lb for obese women (BMI ≥30 kg/m 2 ).
Among women in this sample, unadjusted mean GWG remained relatively constant from 2000 through 2001 (31.2 lb) to 2008 through 2009 (31.4 lb) ( P trend = .46) ( Table 1 ). There was a statistically significant 1.0 biennial percentage point decrease from 2000 through 2001 (37.5%) to 2008 through 2009 (34.2%) in the percentage of women who gained weight within IOM recommendations ( P trend < .01) ( Table 1 ). In addition, there was a significant 0.8 biennial percentage point increase from 2000 through 2001 (42.5%) to 2008 through 2009 (45.5%) in the percentage of women who gained weight above IOM recommendations ( P trend < .01) ( Table 1 ). The percentage of women who gained weight below IOM recommendations remained relatively constant from 2000 through 2001 (20.0%) to 2008 through 2009 (20.3%) ( P trend = .14). There was a statistically significant decrease from 2000 through 2009 in the percentage of women who were underweight ( P trend < .01) or had a normal prepregnancy BMI ( P trend < .01). Additionally, there were significant increases from 2000 through 2009 in the percentages of women who were overweight ( P trend < .01) or class I, II, and III obese ( P trend < .01 for all) before pregnancy ( Table 1 ). From 2000 through 2009, there were also statistically significant changes in all other variables examined, except parity and hypertension ( P trend < .05).
Overall, mean GWG decreased as prepregnancy BMI increased ( Figure 1 ). Normal-weight women had the greatest mean GWG, 33.3 lb (SE 0.07) and obese, class III women had the lowest mean GWG, 20.9 lb (SE 0.41). Overweight (31.3 lb) and obese, class I (27 lb) women both had a mean GWG >25 lb, the upper limit recommended for GWG for overweight and obese women. Additionally, GWG below recommendations was highest for obese, class III women (40%) followed by women who were underweight (32.6%). Underweight women (45.3%) and normal-weight women (41.2%) had the largest proportion of women gaining within IOM recommended levels. Class III obese women had the smallest proportion gaining within IOM recommended levels (25.7%). Overweight (64%) and class I obese (49.5%) women were the 2 groups with the largest proportions gaining above IOM recommendations ( Figure 1 ).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
