Late pregnancy reversal from excessive gestational weight gain lowers risk of childhood overweight—A cohort study

Authors


  • Disclosure: The authors declared no conflict of interest.

  • Funding agencies: This study was supported by German Federal Ministry of Education and Research (BMBF) Grant 01GI0820 (RvK, RE) and Grant 0315088 (RE); Bavarian State Ministry of the Environment and Public Health Grant LP00124 (RE); and LMUinnovativ research priority project MCHealth (sub-project II) (OB).

Abstract

Objective

Whether reversal to adequate gestational weight gain (GWG) in the third trimester reverses the risk for childhood overweight associated with excessive GWG is assessed.

Design and Methods

In a retrospective cohort study in 6,665 mother–child pairs, pre-pregnancy weight and the temporal course of GWG were collected from medical records. Overweight as defined by International Obesity Task Force was assessed at a mean age of 5.8 years. Main exposures were exceeding week-specific cut-off values for GWG in the third trimester or any previous trimester. Logistic regression models, adjusted for possible confounding factors, were used to predict the risk of childhood overweight from excessive GWG in the third trimester with stratification by excessive GWG in previous trimesters.

Results

In the final model, women who avoided excessive GWG in the third trimester had children with a 31% (odds ratio [OR]: 0.69, 95% confidence interval [CI]: 0.59, 0.82) lower probability being overweight. A similar association was observed for reversing from excessive GWG in the first or second trimester to normal GWG in the third trimester: 27% (OR: 0.73, 95% CI: 0.53, 0.99).

Conclusions

Avoidance of excessive GWG in the third trimester is associated with lower risk of childhood overweight even in case of excessive GWG in the first or second trimester.

Introduction

The prevalence of childhood overweight and obesity is rising rapidly [1, 2]. Childhood overweight is associated with serious consequences for health and well-being both during childhood years and in adulthood [3]. Effective early preventive strategies are therefore urgently needed. A growing body of longitudinal research indicates that childhood overweight [8] and higher proportions of body fat in children [8, 14] are associated with excessive gestational weight gain (GWG) via mechanisms referred to as foetal programming [8].

Revised guidelines on GWG according to maternal pre-pregnancy weight were published in 2009 by the Institute of Medicine (IOM)/National Research Council (NRC) [15]. The guidelines provide trimester-specific cut-off values for inadequate and excessive GWG among pre-pregnancy BMI categories from underweight to obese. As a result, it is now possible to identify women at risk of excessive total GWG in the first or second trimester with a fair to good positive predictive value as demonstrated recently [16].

Therefore, targeted interventions for women with excessive GWG in the first or second trimester appear possible. Interventions based on increasing physical activity [17, 18]; dietary advice [19]; or by a combination of physical activity, supplementary weight monitoring, and nutritional counselling have been shown to be effective in reducing total GWG [20]. A recent meta-analyses suggested that dietary measures might be most effective [21].

It is unknown, however, whether a shift to normal GWG in the third trimester from excessive GWG in the first two trimesters would be sufficient to reduce the long-term risk of childhood overweight that is related to persistent excessive GWG. We therefore analyzed the association between exceeding week-specific IOM/NRC cut-off values in different trimesters with the development of childhood overweight at the time of school entry. Data were obtained from 6,665 mother–child pairs in a retrospective cohort study conducted in Germany. The analyses presented here provide unique information about whether avoidance of excessive GWG in the third trimester, independent of excessive GWG in previous trimesters, is associated with a reduction in the risk of overweight in the offspring.

Methods

Study population and data sources

All children attending the compulsory 2010 school entry health examination in six districts in Bavaria were invited to participate in this study. Detailed information about sample recruitment and the study has been published previously [16]. Measurements of the heights and weights of children were carried out during the school entry health examinations in six regions of Bavaria, Germany, from October 2009 to June 2011. The child's weight was measured, wearing underwear, with a calibrated electronic scale (Seca, Birmingham, UK) with an accuracy of 0.1 kg. Height was measured, without wearing shoes, with an accuracy of 0.1 cm using a stadiometer (Seca, Birmingham, UK).

Data on maternal pre-pregnancy weight and height, GWG in the first, second, and third trimesters, diabetes, and gestational diabetes were obtained from the “maternity pass.” In Germany, a “maternity pass” is issued to every first pregnant woman at her first antenatal visit to the gynecologist. The mothers keep “the maternity pass” as a personal document for this and all subsequent pregnancies. The document contains relevant information for the pregnancy such as pregnancy risk factors, rubella antibody titer, ultrasound screening and includes data on serial weight measurements over the course of pregnancy documented by the consulted physician. On the day of the school entry health examination, trained study nurses copied weight-related data from the “maternity pass.”

Information on sociodemographic and risk factors for childhood overweight was collected using a questionnaire that was completed by the child's parents at the time of examination.

The study protocol was approved by the local ethics committee, and signed informed consent was obtained from all the participants.

Of 21,454 school entry health examinations conducted during the study period, 11,730 of the children's mothers agreed to take part in the study (response rate 54.7%) (Figure 1). In this analysis, we included only women with singleton pregnancies, term births, no diagnosis of diabetes mellitus or gestational diabetes mellitus as well as complete data on trimester-specific weights, their child's anthropometric measurements at school entry, and potential confounding factors. GWG values below −10 kg for the whole pregnancy period or in the first, second, and third trimesters were removed, as they appeared implausible. If there was more than one weight per trimester reported, we used the earliest measurement for the analysis. Third-trimester weight measurements later than gestational week 40 were set to missing as well as four implausible values for trimester-specific GWG. This resulted in 6,665 mother–child pairs available for the final analysis.

Figure 1.

Subject flow diagram.

Outcome and explanatory variables

The main exposure was a trimester-specific pattern of excessive GWG. The trimesters were defined as: first trimester between week 1 and 13, second trimester between week 14 and 26, and third trimester between week 27 and 40. We used the week-specific cut-off values for each pre-pregnancy BMI group provided by the IOM/NRC guidelines [15] to determine excessive GWG in each trimester. A detailed description of this procedure has been reported elsewhere [16]. We calculated cut-offs for inadequate and excessive GWG fore each week according to IOM/NRC guidelines. During pregnancy, a linear progression of GWG was assumed with different slopes in the first trimester and the following two trimesters. For example, for normal-weight women, the upper cut-off point of normal GWG in the first trimester is 3 kg. Subtracted from the upper cut-off of total GWG (16 kg) at the end of pregnancy, the upper limit in the remaining 27 wk is a total of 13 kg. This results in a weekly gain of 0.23 kg in the first trimester and 0.48 kg in the second and third trimesters. A detailed description of this procedure has been reported elsewhere [16].

The main outcome variable was child's overweight at school entry. We applied the age- and sex-specific BMI cut-off values of the International Obesity Task Force (IOTF) [22] to classify whether a child was overweight. The term “childhood overweight” includes both overweight and obesity.

The following potential confounding factors, which were ascertained with questionnaires from the KiGGS study [12] and a recent Bavarian study on breastfeeding [23], were associated with outcome and exposure and therefore included in the analysis: birth weight (in grams), maternal age (in years), maternal smoking during pregnancy, breastfeeding, parental socioeconomic status (SES), child's physical activity, and maternal pre-pregnancy BMI on a continuous scale. Maternal smoking during pregnancy was dichotomized as “at no time during pregnancy” or “any time during pregnancy.” Breastfeeding was dichotomized as “at least one month exclusively without interruption” and “less than one month exclusively.” [23] Parental SES was defined using an additive index based on maternal and paternal educational background and current type of maternal and paternal employment [24]. Physical activity of the child was characterized by outdoor physical activity in winter and/or summer and sports in and/or outside a sports club. We dichotomized the four variables to discriminate best between frequent or rare physical activity in our sample.

To determine whether avoidance of excessive GWG in the third trimester was associated with a reduced risk of childhood overweight, the main analysis compared women who gained excessively during the third trimester with those who did not, irrespective of their gains in earlier trimesters. To assess whether this association was independent of whether women had exceeded the IOM/NRC GWG recommendations in either or both earlier trimesters, we additionally performed stratified analyses based on these exposures. Figure 2 depicts the distribution of individuals in the exposure categories and strata considered.

Figure 2.

Distribution of individuals in the exposure categories and strata considered. GWG, gestational weight gain.

Statistical analysis

We used t-tests, U-tests and χ2-tests (as appropriate) for comparisons between characteristics of included and excluded participants.

Multivariate logistic regression models were used to analyze the overall and stratified associations between excessive GWG in the third trimester and childhood overweight. Crude and adjusted analyses were conducted. Model 1 was adjusted for maternal age, maternal smoking during pregnancy, SES, breastfeeding, and child's physical activity, and model 2 was additionally adjusted for maternal pre-pregnancy BMI. Odds ratios (ORs) were calculated and binominal 95% confidence intervals (CIs) were determined. We tested for the interaction of the effect of excessive third trimester GWG on childhood overweight by excessive GWG in the first two trimesters.

Statistical significance of effects was determined by P-values of <0.05 or by 95% CI. All statistical analyses were conducted using the software package R 2.12.1 (http://cran.r-project.org).

Results

Sample description

On average, women excluded from the analysis were slightly younger, had higher pre-pregnancy BMI values and higher GWG in the first trimester, were less likely to have excessive total GWG, more likely to have smoked during pregnancy, and more likely to have lower SES than the women who were included in the study (Table 1). Children excluded from the analysis had slightly higher BMI values, were more likely to be overweight, more likely to watch TV for more than 1 h daily and were less physically active than children who were included in the study.

Table 1. Characteristics of subjects included in and excluded from the study
CharacteristicsAll (n = 9,824)Eligible and excluded (n = 3,159)Eligible and included (n = 6,665)
Mean/ ProportionLower 95% CIUpper 95% CIMean/ ProportionLower 95% CIUpper 95% CIMean/ ProportionLower 95% CIUpper 95% CI
  1. BMI, body mass index; CI, Confidence interval; GWG, gestational weight gain; PA, physical activity; SES, socioeconomic status; SGA and LGA, small and large for gestational age, respectively.
  2. aGWG per week can be approximated by dividing GWG by trimester by 13 weeks.
Maternal age (years)28.928.829.028.628.528.729.129.029.2
Primiparous51.350.152.648.546.450.752.851.454.3
Pre-pregnancy BMI (kg/m2)23.423.323.523.523.423.623.423.323.5
GWG (kg)         
Total14.914.815.014.814.614.915.014.915.1
Trimester 1a1.21.21.31.51.41.61.11.11.2
Trimester 2a7.57.47.67.57.47.67.57.57.6
Trimester 3a13.613.513.713.413.313.513.613.513.7
Excessive GWG         
Total53.552.554.552.350.154.453.952.755.1
Trimester 127.126.228.026.524.728.327.326.328.4
Trimester 245.444.346.444.842.746.845.644.446.8
Trimester 352.151.153.251.048.953.152.551.353.7
Smoking in pregnancy12.011.412.714.613.315.910.910.111.6
Low parental SES33.232.234.139.237.341.030.729.631.8
Child BMI (kg/m2)15.515.515.615.615.615.715.515.515.5
Overweight children11.010.411.712.411.213.610.49.711.1
Child age (years) Female sex5.85.85.85.85.85.85.85.85.8
Female sex48.647.649.648.646.850.348.747.549.9
Child birth weight (g)3412.93404.03421.83406.43397.33415.53415.83407.03424.6
SGA9.89.210.410.89.711.99.38.610.0
LGA8.47.99.08.67.69.68.37.79.0
Breastfeeding (> 1 month full)73.272.374.172.871.174.673.372.374.4
Low physical activity of the child5.65.16.07.86.88.94.74.25.2

In the study population, most women were primiparous (52.8%) with a mean BMI in the normal range (23.4 kg/m2). Total GWG exceeded the IOM/NRC recommendations in more than half of the women. Children were on average 5.8 years old and 10.4% of the children were overweight at the school entry health examination.

On average, weight measurements in the first, second, and third trimesters were obtained around the 7th, the 23rd, and the 36th week of gestation (trimester 1: mean= 7.5, SD= 2.3; trimester 2: mean= 23.7, SD=1.3; trimester 3: mean= 35.7, SD= 0.8).

Association of excessive GWG in different trimesters and childhood overweight

Of 6,665 pregnancies considered, 3,501 (52.5%) had excessive GWG in the third trimester. In most of these, 2,776 (79.3%) excessive GWG had also been observed in either of the two previous trimesters or both. Among 3,164 without excessive GWG in the third trimester, 691 (21.8%) had reversed from excessive GWG in either of the two previous trimesters or both (Figure 2).

Among all women, we observed a significant association of excessive GWG in the third trimester with childhood overweight in both the crude and adjusted models. Regardless of possible excessive GWG in previous trimesters, children of mothers without excessive GWG in the third trimester had a lower risk of overweight at the school entry health examination (OR: 0.63, 95% CI: 0.54, 0.75) (Table 2, model 1). After adjusting for maternal pre-pregnancy BMI (model 2), the reduction in risk of childhood overweight associated with normal GWG in the third trimester was still significant (OR: 0.69, 95% CI: 0.59, 0.82).

Table 2. Association of excessive GWG in the third trimester and childhood overweight
Excessive gestational weight gain (GWG)CrudeModel 1aModel 2b
OR95%CIOR95%CIOR95%CI
  1. Odds ratios and 95% confidence intervals are provided for the overall analysis and also for analyses stratified by excessive GWG in preceding trimesters.
  2. aAdjusted for maternal age, maternal smoking during pregnancy, SES, breastfeeding, and physical activity.
  3. bAdjusted for confounding factors mentioned in ‘a’ and additionally for pre-pregnancy BMI.
Overall         
Excessive GWG in third trimester “yes” (n = 3,501; reference) vs. “no” (n = 3,164)0.610.520.720.630.540.750.690.590.82
Stratified         
“No excessive GWG in any trimester” (n = 2,473) vs. “Excessive GWG in third trimester only” (n = 725; reference)0.650.500.850.660.500.860.750.570.99
“Excessive GWG in first or second trimester only” (n = 691) vs. “Excessive GWG in third and any previous trimester” (n = 2,776; reference)0.590.440.790.640.470.850.730.530.99

There was no interaction of the effect of third trimester excessive GWG on overweight in the offspring by GWG in previous trimesters. Among women who gained excessively in either the first or second trimester or both, children of mothers whose third trimester GWG did not exceed the IOM/NRC recommendations in the third trimester had a lower risk of overweight at the school entry examination (OR: 0.64, 95% CI: 0.47, 0.85) than those who did (Figure 2; Table 2, model 1). After adjusting for maternal pre-pregnancy BMI (model 2), this association was attenuated (OR: 0.73, 95% CI: 0.53, 0.99).

Similarly, among women who did not gain excessively in the first or second trimester, children of mothers who also did not gain excessively in the third trimester had a lower risk of overweight at the school entry health examination (OR: 0.66, 95% CI: 0.50, 0.86) than those who did (model 1). After adjusting for maternal pre-pregnancy BMI (model 2), this association was attenuated (OR: 0.75, 95% CI: 0.57, 0.99).

Discussion

Avoidance of excessive GWG in the third trimester was associated with a lower risk of childhood overweight irrespective of potentially excessive GWG in previous trimesters in this large sample of German mothers and children. This association was not explained by confounding by maternal pre-pregnancy BMI. The strength of this association was almost identical for women who had not experienced excessive GWG in any previous trimester and those with excessive GWG in the first or second trimester or both. These findings provide a rationale for developing interventions on limiting GWG to avoid excessive weight gain in the third trimester.

The observation that excessive GWG in the third trimester is associated with a risk for childhood overweight while its avoidance is protective matches the available evidence for the association between excessive total GWG and childhood overweight [8] and reflects the high positive predictive value of excessive third trimester for total excessive GWG (81.4%, 85.0%, 90.0% and 96.1% for obese, overweight, normal weight, and underweight mothers, respectively) demonstrated recently [16]. Importantly, the association could not be explained by maternal BMI. Although categorization of excessive and inadequate GWG is done by mothers' preconception BMI category, there might still be some residual confounding by maternal BMI because of its strong association with childhood BMI [11, 25]. Indeed, that is what we observed. When pre-pregnancy BMI was added in model 2, the risk was attenuated (from OR = 0.61 to 0.69 for the overall association) with 95% CIs not including 1. Similarly, in the analyses stratified by GWG in the first two trimesters, including pre-pregnancy BMI in model 2 reduced the association by a similar amount again with 95% CIs not including 1. The wider 95% CIs in the stratified analyses reflect smaller numbers in the strata. It is important that adjusting for pre-pregnancy BMI may attenuate but does not eliminate the association between excessive GWG in the third trimester and childhood overweight. This clear finding adds materially to the information available with which to assess the long-term consequences of gaining within the GWG guidelines. The quantitative risk analysis that was commissioned part of the development of the 2009 IOM/NRC guidelines [15] showed that the association between excessive GWG and childhood obesity was likely to be a major route by which excessive GWG establishes long-term health consequences. This finding provides stronger evidence for this estimation than was available at the time of the committee's work. The presumed underlying mechanism for this association is perinatal programming [28, 29].

The even more important new finding is, however, that a lower risk of childhood overweight with the avoidance of excessive GWG in the third trimester may also be achieved in case of excessive GWG in previous trimesters. In previous studies, investigators [8, 30, 31] found that absolute, week-specific GWG in the first and second trimesters was more relevant for the risk of later overweight in the offspring than week-specific GWG in the third trimester. These findings would suggest that most of the association of GWG on overweight in the offspring is determined in the first two trimesters. Our observations seem to contradict this finding, but may actually result from our use of a different and probably more appropriate categorization of GWG. While some GWG is physiological, a GWG exceeding the physiological requirements may be associated with adverse effects for the mother and offspring. Using “excessive” GWG, a measure of relative instead of absolute GWG, captures the proportion of higher GWG that is not associated with optimal outcomes and, thus, is the portion of GWG that should be avoided. Since reversing from excessive GWG in the first two trimesters to normal GWG in the third trimester reduced the risk for childhood overweight to a similar degree as absence of excessive GWG throughout pregnancy, identification of excessive GWG in the first two trimesters does not justify complacency: an increased risk for overweight in the offspring might still be reversed by appropriate interventions. Without such interventions, only 20% (691/3,467) pregnancies with excessive GWG in the first or second trimester will reverse to normal GWG in the third trimester.

Our study has some strengths and limitations. The data were generated in a population-based sample that included children in rural and urban regions of Bavaria, Germany. There are only few recent publications applying the new IOM/NRC guidelines. The observed prevalence of excessive GWG was only slightly higher than in the recent publications applying the new IOM/NRC guidelines [14, 32]. Nonparticipation and missing data may limit the generalizability of our findings. Some eligible mother–child dyads had to be excluded because of missing or implausible data on GWG or potential confounders. The differences, however, were small and inconsistent (excluded mother–child dyads reported a higher first trimester GWG but a lower total GWG than mother–child dyads included). The prevalence of childhood overweight observed in this study population was similar to other data from Germany [33, 34]. Selection bias might also be issue. Regarding the most important finding of this study, demonstrating that reversal from excessive GWG in the first two trimesters to normal GWG in the third reduces the risk for overweight, selection bias, however, would have to operate a an effect modifier, which does not appear likely.

Exposure data were ascertained from medical records and outcome data were measured on the day of the examination. All GWG data, apart from pre-pregnancy weight, which was recalled at the first antenatal visit, were measured at the antenatal visits. Although the total sample size of the study was high, it was not high enough, however, to disentangle the effects of first and second trimester excessive GWG and to allow assessing potential dose effects regarding excessive GWG in both trimesters. Unfortunately, no data on diet or physical activity during pregnancy have been collected in this retrospective cohort study. Therefore, no empirical evidence as to potentially effective dietary or physical activity habits during pregnancy could be obtained.

In conclusion, avoidance of excessive GWG in the third trimester reduces the risk of childhood overweight irrespective of excessive GWG in previous trimesters. Therefore, any interventions focusing on either women with excessive GWG in the first two trimesters or on all women to avoid excessive GWG in the third trimester appear warranted. Because excessive GWG in the third trimester was observed in more than half of the pregnancies and because the strength of the effect of GWG on childhood overweight was of considerable size, such interventions may constitute an important preventive strategy against childhood overweight.

Acknowledgments

We thank the local health authorities, study nurses, and the families for their participation and support.

  The author's responsibilities were as follows: RvK and RE designed research, RE was responsible for the data acquisition, AC and OB performed statistical analysis, KMR made important suggestions for the analysis of the data, RvK, AC, and KMR wrote the paper. RvK has primary responsibility for final content. All authors read and approved the final manuscript. This work is a part of the PhD project of AC. The study was part of the PEPO consortium of the Competence Network Obesity supported by the German Federal Ministry of Education and Research (BMBF) (project coordination and data analysis). It was also supported by the Bavarian State Ministry of the Environment and Public Health (data acquisition at school entry health examination). OB was supported by LMUinnovativ research priority project MCHealth (sub-project II).

Ancillary