Objective
The purpose of this study was to investigate the association of trimester-specific gestational weight gain with offspring fetal growth, obesity risk, and cardiometabolic health outcomes from birth to 4 years of age.
Study Design
We conducted the present study with 977 mother-child pairs of the pregnancy cohort “Rhea” study in Crete, Greece. We measured birthweight, body mass index from 6 months to 4 years of age, waist circumference, skinfold thickness, blood pressure, and blood levels of lipids, C-reactive protein, and adipose tissue hormones at 4 years of age. We used multiple linear and log Poisson regression models to examine the association of exposure with continuous or binary outcomes, respectively.
Results
Greater rate of gestational weight gain in the first trimester of pregnancy (per 200 g/wk) was associated with increased risk of overweight/obesity from 2 years (relative risk [RR], 1.25; 95% confidence interval [CI], 1.09–1.42) to 4 years of age (RR, 1.15; 95% CI, 1.05–1.25), but not with birth size. Each 200 g/wk of weight gain in the first trimester of pregnancy was also associated with greater risk of high waist circumference (RR, 1.13; 95% CI, 1.04–1.23), high sum of skinfold thickness (RR, 1.15; 95% CI, 1.02–1.29), and higher diastolic blood pressure at 4 years of age (β, 0.43 mm Hg; 95% CI, 0.00–0.86). Greater rate of gestational weight gain during the second and third trimesters of pregnancy (per 200 g/wk) was associated with greater risk of large-for-gestational-age neonates (RR, 1.22; 95% CI, 1.02, 1.45) and higher levels of cord blood leptin (ratio of geometric means, 1.08; 95% CI, 1.00–1.17), but not with child anthropometry at later ages.
Conclusion
Timing of gestational weight gain may influence childhood cardiometabolic outcomes differentially.
Gestational diabetes mellitus, maternal obesity, and excessive weight gain during pregnancy, which are each markers of fetal overnutrition, are considered among the most important modifiable early-life risk factors of childhood obesity. From a public health perspective, gestational weight gain (GWG) recently has gained particular interest because interventions on GWG could benefit from the fact that they (1) target women for the short duration of pregnancy, (2) can take advantage of the frequent visits of women to their obstetricians, and, if successful, (3) can reduce maternal postpartum weight retention and the risk of maternal obesity that might complicate future pregnancies.
Excessive GWG has been associated with poor health outcomes for both mother and child over the short- and long-term. Two recent metaanalyses suggested that excessive GWG is associated with higher risk of offspring obesity throughout life ; more debatable are the results when offspring adiposity is assessed by measures other than body mass index (BMI). Some birth cohorts have attempted to disentangle the effect of GWG depending on the timing of the gain and have suggested that early pregnancy weight gain might be critical for the development of offspring obesity later in life. However, we are not aware of any studies that have examined the association of trimester-specific GWG with offspring obesity in the first 4 years of life.
Findings for the association between GWG and other offspring cardiovascular traits such as blood pressure or serum lipid profile have been less consistent, mainly because of confounding by offspring adiposity. We are not aware of any studies that have examined the association of trimester-specific GWG with offspring blood pressure, lipid profile, and adipose tissue hormones in children as young as 4 years old in a population of substantial size.
In the present study, we examined the association of GWG (total and trimester-specific) with offspring birthweight, postnatal growth, obesity, and a range of cardiometabolic risk factors at 4 years old (waist circumference, skinfolds, blood pressure, lipids, adiponectin, leptin, and C-reactive protein) in the “Rhea” pregnancy cohort in Crete, Greece.
Materials and Methods
We recruited mothers who became pregnant from February 2007 to January 2008 who were residents in the prefecture of Heraklion Crete, Greece. Research assistants invited women to provide blood and urine samples and to participate in a face-to-face interview at enrollment at <15 weeks of gestation. The next contact with the mothers was at 24 weeks of gestation, at birth, at 8-10 weeks after delivery, and for the child’s follow-up examination at 9 and 18 months, and at 4 years old. The study was approved by the Ethical Committee of the University Hospital of Heraklion (Crete, Greece), and all participants provided written informed consent.
Of 1363 singleton live births, 977 had complete data on maternal GWG, offspring birthweight, and BMI from 6 months to 4 years. Of these, 451 children had cord leptin measurements; 661 children had waist circumference and skinfold thickness measurements at 4 years; 518 children had blood pressure measurements, and 567 children provided blood samples at the 4-year follow-up examination. Trimester-specific GWG data were available for 595 mother-child pairs. We excluded births at <34 weeks of gestation (n = 20) to limit bias because of the dependence of GWG on gestational age.
GWG
At enrollment, research assistants measured mother’s height (centimeters) and weight (kilograms) in light clothing without shoes and obtained information on prepregnancy weight. We calculated first-trimester GWG rate to be the difference between self-reported prepregnancy weight and weight as measured at enrollment and divided it by the corresponding gestational age. We calculated second- and third- trimester GWG rate to be the difference between total and first-trimester GWG. We decided to examine changes in first-trimester and second- and third- trimester GWG per 200 g/wk in an effort to use a common rate of change that would be reasonable for both periods of pregnancy. This weekly rate corresponds to a 2.8-kg more weight gain up to 14 weeks of gestation and to 5.2-kg more weight gain for the remaining period of pregnancy, which is close to the recommended weight gain for the first trimester of pregnancy, and corresponds to one-half of the recommended weight gain for the second and third trimester of pregnancy for normal-weight women. We obtained information on total GWG after delivery (9 ± 2 months), based on a phone-interview with the mother. We examined total GWG as continuous variable (per 2 kg change) and as categoric (inadequate, adequate, or excessive).
Child anthropometry
Weight and length at birth were obtained from the hospital delivery logs and medical records. Large-for-gestational-age (LGA) neonates were defined as live-born infants above the 90th percentile of birthweight for gestational age in a referent population.
At child follow-up visits, trained research assistants measured weight and length (up to 2 years old) or height (from 2-4 years old) using validated scales (Seca 354 baby scale, Seca Bellisima 841; Seca Corporation, Hanover, MD) and stadiometers (Seca 210 measuring mat, Seca 213; Seca Corporation) according to standard operating procedures. Repeated measures of weight and length/height were also abstracted from the children’s health cards. We calculated BMI and converted raw values into sex- and age-specific standard deviation scores (SD scores) by using internally generated growth reference curves. Because of variation in children’s ages at measurement, we estimated BMI SD scores at exactly 6 months and 1, 2, 3, and 4 years of age with a sex- and age-specific, multilevel (mixed) linear model that was fitted with fractional polynomials and random effects for the child. To minimize the effect of children with implausible growth trajectories, we excluded children whose measurements were ≤5 SD or >5 SD from the mean at any age.
We defined rapid BMI growth the first 6 months of life as a BMI SD score gain >0.67; children with a BMI SD score gain ≤0.67 constituted the reference group. For defining childhood overweight/obesity at 2, 3, and 4 years old, we used the BMI cutoff points for sex and age that were proposed by the International Obesity Task Force.
Waist circumference was measured at 4 years old in duplicate to the nearest 0.1 cm in a standing position, at the high point of the iliac crest at the end of a gentle expiration, with the use of a measuring tape (Seca 201; Seca Corporation). Skinfold thickness was measured at 4 anatomic sites (triceps, subscapular, suprailiac, and thigh) on the right side of the body in triplicate to the nearest 0.1 mm with a calibrated caliper (Harpenden HSK- BI, CE-0120; Baty International, West Sussex, UK). Intraobserver and interobserver reliability assessments were undertaken according to previous methods. Intraobserver reliability was >0.98, and interobserver reliability was >0.82 for all anthropometric measurements. We used the 75th percentile of the study cohort distribution as a cutoff point to denote high values on waist circumference (≥55.5 cm) and on skinfolds (≥49.7 mm for girls and ≥41.7 mm for boys).
Child cardiometabolic risk factors
At the 4-year examination, trained research assistants measured systolic and diastolic blood pressure after 5 minutes rest in the seated position at the child’s right arm with a cuff of appropriate size for arm circumference using a Dinamap Pro Care 400 (Critikon Inc, Tampa, FL), which uses an oscillometric method. We used the average of 5 consecutive measurements that were taken with 1-minute intervals. Nonfasting blood samples were collected from the children at the end of the 4-year follow-up visit. Blood samples were processed within 2 hours, and the serum was stored at –80°C until analysis. Total cholesterol and HDL-cholesterol in serum were measured with standard enzymatic methods (Medicon Hellas SA, Gerakas, Greece). Leptin and adiponectin (Invitrogen, Carlsbad, CA) were measured by an enzyme-linked immunosorbent assay, and C-reactive protein was measured with a high-sensitivity homogenous immunoassay (ORS 6199; Beckman Coulter Inc, Fullerton, CA). Cord leptin, was measured as described previously. Inter- and intraassay coefficients of variation were <5%.
Statistical methods
We used linear regression models to estimate the β coefficients for the association between GWG and continuous outcomes and log-Poisson regression models to estimate the relative risks (RRs) for the association between GWG and binary outcomes. We applied generalized additive models to explore the shape of the relationships between GWG and continuous outcomes, with adjustment for confounders. Generalized additive models indicated that the associations between GWG (total or trimester-specific) and the outcomes did not deviate from linearity. Leptin, adiponectin, and C-reactive protein were log transformed to normalize their distributions. The resultant regression coefficients were exponentiated to give a ratio of geometric means per change in exposure. To assess differences in BMI trajectories from birth to 4 years for each stratum of excessive, adequate, and inadequate GWG, we constructed mixed effects linear regression models using BMI for age percentiles of our population. Model parameters were estimated with restricted maximum likelihood.
We considered the following potential confounders in our models: maternal age at delivery, educational level at recruitment (low, ≤9 years; medium, >9 years up to attending postsecondary school education; high, attending university or having a university/technical college degree), parity, smoking during pregnancy, prepregnancy BMI (kilograms per square meter), paternal BMI (kilograms per square meter), gestational length (for models that used birth size as an outcome or total GWG as an exposure), duration of any breastfeeding (months), sex, age at outcome assessment, energy intake at 4 years old (kilocalories per day) by a validated food frequency questionnaire), time watching television at 4 years old (<30 minutes, 1-2 hours, ≥3 hours per day), and child’s BMI at 4 years old (for models that used cardiometabolic risk factors as outcomes). Level of completeness for covariates included in the main models was >95%.
To assess the potential modifying effects of child sex (male, female), maternal prepregnancy BMI (as continuous variable or as categoric [<25 kg/m 2 or ≥25 kg/m 2 and as <25 kg/m 2 , 25-30 kg/m 2 , ≥30 kg/m 2 ]), and breastfeeding duration (<1 month, 1-6 months, >6 months), we included appropriate interaction terms in regression models and stratified the sample. We also repeated analyses after excluding (1) children who were born preterm (<37 weeks of gestation), (2) children who were born small for gestational age, (3) pregnant women who were diagnosed with gestational diabetes mellitus, and (4) preeclamptic pregnancies. We used Stata SE for all analyses (version 13; StataCorp, College Station, TX).
Results
GWG (total and trimester-specific) by maternal and child characteristics are presented in Table 1 . Women who were highly educated, who were multiparous, with higher prepregnancy BMI, and who experienced gestational diabetes mellitus had lower GWG throughout pregnancy. Women of Greek origin and current smokers in pregnancy had higher first-trimester GWG. According to the 2009 Institute of Medicine recommendations, 45% of the women gained excessive weight; 32% of the women gained adequate weight, and 23% of the women gained inadequate weight during their pregnancy. The prevalence of LGA neonates was 17%; 31% of the infants had rapid BMI growth the first 6 months of life. The prevalence of overweight and obesity was 10% and 1% at 2 years, 16% and 3% at 3 years, and 21% and 4% at 4 years, respectively.
| Characteristic | Subjects, % | GWG a | ||
|---|---|---|---|---|
| Total (n = 977), kg | First-trimester (n = 595), g/wk | Second- and third- trimester (n = 595), g/wk | ||
| Study population, median interquartile range | 100 | 14 ± 8 | 153 ± 260 | 449 ± 260 |
| Maternal characteristics | ||||
| Maternal origin | ||||
| Greek | 92 | 14.1 ± 5.7 | 184 ± 265 b | 451 ± 205 |
| Other | 8 | 13.3 ± 5.8 | 60 ± 241 b | 450 ± 200 |
| Maternal educational level | ||||
| Low | 17 | 13.2 ± 5.9 b | 206 ± 292 | 405 ± 218 b |
| Medium | 51 | 14.4 ± 6.0 b | 178 ± 276 | 470 ± 213 b |
| High | 32 | 13.8 ± 5.1 b | 162 ± 232 | 439 ± 183 b |
| Marital status | ||||
| Married | 89 | 13.9 ± 5.7 b | 168 ± 257 | 445 ± 208 |
| Other than married | 11 | 15.3 ± 5.4 b | 233 ± 304 | 490 ± 175 |
| Maternal age at delivery | ||||
| <30 y | 47 | 14.7 ± 5.9 b | 197 ± 291 | 472 ± 215 b |
| ≥30 y | 53 | 13.4 ± 5.5 b | 159 ± 236 | 431 ± 193 b |
| Parity | ||||
| Primiparous | 43 | 14.8 ± 5.8 b | 183 ± 300 | 469 ± 203 b |
| Multiparous | 57 | 13.4 ± 5.6 b | 165 ± 233 | 438 ± 207 b |
| Maternal smoking in pregnancy | ||||
| Smoker | 26 | 15.3 ± 6.0 b | 224 ± 230 b | 480 ± 207 |
| Nonsmoker | 74 | 13.5 ± 5.6 b | 156 ± 270 b | 442 ± 204 |
| Prepregnancy body mass index | ||||
| Obese | 11 | 10.9 ± 6.9 b | 127 ± 318 | 372 ± 249 b |
| Overweight | 21 | 13.4 ± 5.5 b | 166 ± 274 | 447 ± 193 b |
| Normal | 63 | 14.6 ± 5.3 b | 184 ± 250 | 464 ± 195 b |
| Underweight | 5 | 15.3 ± 6.8 b | 222 ± 263 | 453 ± 242 b |
| Institute of Medicine categories for total GWG | ||||
| Excessive | 45 | 18.6 ± 4.6 b | 239 ± 270 b | 585 ± 170 b |
| Adequate | 32 | 11.9 ± 2.6 b | 123 ± 236 b | 407 ± 122 b |
| Inadequate | 23 | 7.9 ± 2.9 b | 128 ± 264 b | 245 ± 161 b |
| Physically active in pregnancy | ||||
| Yes | 8 | 13.5 ± 5.1 | 135 ± 196 | 462 ± 172 |
| No | 92 | 14.0 ± 5.8 | 181 ± 269 | 449 ± 208 |
| Energy intake in pregnancy, kcal/d (kcals/day) | ||||
| ≥2000 | 45 | 14.5 ± 6.3 b | 204 ± 302 | 458 ± 232 |
| <2000 | 55 | 13.5 ± 5.4 b | 161 ± 245 | 445 ± 193 |
| Gestational diabetes mellitus | ||||
| Yes | 9 | 12.0 ± 6.5 b | 111 ± 363 | 353 ± 242 |
| No | 91 | 14.1 ± 5.7 b | 144 ± 267 | 460 ± 197 |
| Mode of delivery | ||||
| Cesarean | 51 | 13.8 ± 5.8 | 191 ± 253 | 443 ± 202 |
| Vaginal | 49 | 14.2 ± 5.7 | 158 ± 274 | 459 ± 207 |
| Child characteristics | ||||
| Preterm (34-<37 wk) | ||||
| Yes | 9 | 13.2 ± 5.6 | 173 ± 235 | 437 ± 196 |
| No | 91 | 14.1 ± 5.7 | 176 ± 266 | 452 ± 205 |
| Gender | ||||
| Male | 50 | 14.3 ± 5.5 | 190 ± 269 b | 456 ± 208 |
| Female | 50 | 13.7 ± 5.9 | 162 ± 258 b | 445 ± 201 |
| Large-for-gestational-age neonate | ||||
| Yes | 17 | 15.6 ± 6.2 b | 209 ± 304 | 492 ± 216 b |
| No | 83 | 13.7 ± 5.6 b | 169 ± 248 | 443 ± 202 b |
b Statistically significant at P < .05 for comparisons by Mann-Whitney U -test, Kruskal-Wallis.
GWG and childhood obesity
Table 2 presents the associations of total and trimester-specific GWG with offspring size up to 4 years and the risks of LGA neonates, rapid BMI growth in the first 6 months of life, and child obesity/adiposity from 2-4 years of age. Figure 1 shows the modeled BMI for age percentile growth trajectory from birth to 4 years by strata of excessive, adequate, or inadequate GWG. Figure 2 shows the predicted probability and 95% confidence intervals (CIs) of childhood overweight/obesity at 4 years old by trimester-specific GWG rate.
| Outcome | Measure of association | GWG | ||
|---|---|---|---|---|
| Total (per 2 kg) | First trimester (per 200 g/wk) | Second and third trimesters (per 200 g/wk) | ||
| Birth and infancy | ||||
| Standard deviation score | ||||
| Birthweight | β (95% CI) | 0.05 (0.03–0.07) a | 0.04 (-0.01–0.10) | 0.07 (0.00–0.14) a |
| BMI at 6 mo | β (95% CI) | 0.04 (0.01–0.06) a | 0.07 (0.01–0.13) a | 0.03 (–0.03 to 0.10) |
| Large-for-gestational-age neonates | RR (95% CI) | 1.12 (1.07–1.17) a | 1.10 (0.97–1.24) | 1.22 (1.02–1.45) a |
| Rapid BMI growth in the first 6 mo b | RR (95% CI) | 1.05 (1.02–1.08) a | 1.05 (0.98–1.13) | 1.10 (0.99–1.23) |
| Early childhood (1-4 y) | ||||
| BMI standard deviation score | ||||
| 1 y | β (95% CI) | 0.03 (0.01–0.06) a | 0.07 (0.01–0.13) a | 0.03 (–0.04 to 0.10) |
| 2 y | β (95% CI) | 0.03 (0.01–0.06) a | 0.07 (0.01–0.13) a | 0.03 (–0.04 to 0.10) |
| 3 y | β (95% CI) | 0.03 (0.01–0.06) a | 0.07 (0.01–0.13) a | 0.03 (–0.04 to 0.10) |
| 4 y | β (95% CI) | 0.03 (0.01–0.06) a | 0.07 (0.01–0.13) a | 0.03 (–0.04 to 0.10) |
| Overweight/obese c | ||||
| 2 y | RR (95% CI) | 1.11 (1.05–1.17) a | 1.25 (1.09–1.42) a | 1.08 (0.82–1.40) |
| 3 y | RR (95% CI) | 1.07 (1.02–1.11) a | 1.18 (1.07–1.31) a | 0.97 (0.83–1.15) |
| 4 y | RR (95% CI) | 1.05 (1.01–1.08) a | 1.15 (1.05–1.25) a | 1.00 (0.87–1.14) |
| Adiposity measurements, 4 y | ||||
| Waist circumference, cm | β (95% CI) | 0.13 (–0.001 to 0.28) | 0.35 (0.01–0.69) a | 0.03 (–0.38 to 0.46) |
| Sum of skinfolds, mm | β (95% CI) | 0.42 (0.00–0.85) a | 0.72 (–0.33 to 1.78) | 0.21 (–1.09 to 1.53) |
| Waist circumference ≥75th percentile | RR (95% CI) | 1.08 (1.04–1.13) a | 1.13 (1.04–1.23) a | 1.12 (0.94–1.33) |
| Sum of skinfolds ≥75th percentile | RR (95% CI) | 1.04 (0.98–1.09) | 1.15 (1.02–1.29) a | 0.98 (0.82–1.17) |
a Statistically significant at the 5% level coefficients
b Defined as a gain in standard deviation score for BMI >0.67 SD (reference category ≤0.67)
c Defined with use of the BMI cutoff point for sex and age that was proposed by the International Obesity Task Force.
Total GWG
Higher total GWG was associated with higher birthweight SD score (β, 0.05; 95% CI, 0.03–0.07) and greater risk of LGA neonates (RR, 1.12; 95% CI, 1.07–1.17) in the fully adjusted models ( Table 2 ). Total GWG consistently was associated positively with child BMI SD score across the study period and correspondingly with greater risk of rapid BMI growth in the first 6 months of life and of overweight/obesity at 2, 3, and 4 years of age ( Table 2 ). It was also associated with greater risk of high waist circumference (RR, 1.08; 95% CI, 1.04–1.13) and high skinfold thickness measurements at 4 years of age (β, 0.42 mm; 95% CI, 0.00–0.85).
Institute of Medicine categories of GWG
Offspring of mothers with excessive GWG had 60% increased risk of being LGA neonates (95% CI, 1.14–2.23; Appendix ; Supplementary Table 1 ) and were in higher BMI percentiles across the study period, compared with children of mothers with adequate GWG ( Figure 1 ). Offspring of mothers who gained inadequate GWG, compared with the reference group, had lower birthweight SD score (β, –0.16; 95% CI, –0.32 to –0.001; Supplementary Table 1 ) and remained in lower BMI percentiles up to 4 years of age, compared with the reference group of adequate GWG ( Figure 1 ).
Trimester-specific GWG
Greater first-trimester GWG rate was associated with higher child BMI SD score at 6 months and 1, 2, 3 and 4 years of age ( Table 2 ). Each 200 g/wk increase in first-trimester GWG was associated with higher risk of overweight/obesity at 2 years (RR, 1.25; 95% CI, 1.09–1.42), at 3 years (RR, 1.18; 95% CI, 1.07–1.31) and at 4 years (RR, 1.15; 95% CI, 1.05–1.25). Overall, the first-trimester GWG rate showed a positive linear relationship with the probability of overweight/obesity at 2, 3, and 4 years of age ( Figure 2 ). Higher first-trimester GWG was also associated with increased risk of high waist circumference (RR, 1.13; 95% CI, 1.04–1.23), and skinfold thickness measurements at 4 years of age (RR, 1.15; 95% CI, 1.02–1.29; Table 2 ).
Greater second- and third- trimester GWG rate was associated with increased birthweight SD score (β, 0.07; 95% CI, 0.00–0.14)] and risk of LGA neonates (RR, 1.22; 95% CI, 1.02–1.45), but not with child BMI SD scores at later ages ( Table 2 ). Rate of second- and third-trimester GWG showed no association with the probability of overweight/obesity at 2, 3, and 4 years of age ( Figure 2 ; Table 2 ).
Cardiometabolic risk factors
Total and trimester-specific GWG generally were not associated with lipids, adiponectin, or C-reactive protein at 4 years of age ( Table 3 ). We observed a positive association of first-trimester GWG with diastolic blood pressure at 4 years old, even with adjustment for the child’s BMI (β, 0.43 mm Hg; 95% CI, 0.00–0.86; Table 3 ). Each 200 g/wk increase in second- and third-trimester GWG was associated with higher cord leptin levels (ratio of geometric means, 1.08; 95% CI: 1.00–1.17), but with lower leptin levels at 4 years of age (ratio of geometric means, 0.89; 95% CI, 0.83–0.97; Table 3 ).
| Cardiometabolic outcome | GWG | ||
|---|---|---|---|
| Total (per 2 kg) | First trimester (per 200 g/wk) | Second and third trimesters (per 200 g/wk) | |
| Systolic blood pressure, mm Hg | −0.15 (−0.38 to 0.07) | 0.09 (–0.68 to 0.48) | 0.05 (–0.68 to 0.78) |
| Diastolic blood pressure, mm Hg | 0.00 (–0.15 to 0.17) | 0.43 (0.00–0.86) a | −0.059 (−0.61 to 0.49) |
| Total cholesterol, mg/dL | −0.57 (−1.46 to 0.32) | −0.04 (−2.42 to 2.33) | −1.91 (−4.88 to 1.06) |
| High-density lipoprotein, cholesterol, mg/dL | −0.10 (−0.43 to 0.23) | 0.31 (−0.57 to 1.19) | −0.89 (−1.99 to 0.20) |
| C-reactive protein, ratio geometric mean b | 1.01 (0.96–1.05) | 1.01 (0.90–1.14) | 1.03 (0.89–1.20) |
| Adiponectin, ratio geometric mean b | 0.99 (0.97–1.01) | 0.98 (0.93–1.03) | 0.95 (0.89–1.02) |
| Leptin, ratio geometric mean b | |||
| Cord blood | 1.02 (1.00–1.05) a | 0.99 (0.93–1.05) | 1.08 (1.00–1.17) a |
| 4 y | 0.97 (0.95–1.00) a | 1.01 (0.95–1.08) | 0.89 (0.83–0.97) a |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
