Maternal nutrition before and during pregnancy is a critical window of opportunity for optimizing maternal and child health outcomes (1–3). Women who enter pregnancy with a healthy body mass index (BMI between 18.5 and 24.9 kg/m²) are at a lower risk for adverse pregnancy outcomes such as pre-eclampsia, gestational diabetes, large or small for gestational age (LGA or SGA), and stillbirths (4–7). Poor maternal nutrition also has long-term implications for the offspring health by placing them at increased risk of diabetes, hypertension, hypercholesterolemia and heart disease later in life (8–12).

Excessive or inadequate gestational weight gain (GWG) during pregnancy has likewise been associated with increased risk adverse birth outcomes such as SGA, LGA, preterm birth and gestational diabetes (13–15). Prior work from the Women First Trial including over 2000 women from the Democratic Republic of Congo, Guatemala, India and Pakistan reported that women’s pre-conception BMI (PPBMI), early GWG from preconception to 12 weeks and from 12 to 32 weeks were independently associated with birth length and weight (16).

However, the long-term impact of GWG on maternal and child health outcomes is less well established. In a systematic review, excess GWG above the Institute of Medicine (IOM) recommendations was associated with an additional post-partum weight retention (PPWR) of 3 kg and 4.7 kg at 3 and 15 years postpartum, respectively, compared to those who gained within IOM recommendations (17). Women who gained below IOM recommendations had a 3 kg less of weight retention at 6 months postpartum; however, there was no long-term associations at 15 years (17). PPBMI was reported to have a negative association with PPWR, though GWG was considered the key driver of short and long term PPWR (18). However, these conclusions were based on limited data primarily from high income settings with inadequate adjustment for confounding factors (such as breastfeeding and physical activity, for example). Furthermore, there was imprecise measurement of key exposures as pre-pregnancy weight that was often based on retrospective self-report data and may be prone to reporting bias (19, 20). The measures of GWG also do not permit the examination of the possible effects of differences in the timing and/or patterns of weight gain across pregnancy. Thus, there remain questions on the relative contributions of maternal nutrition before and during pregnancy on weight retention and long-term health outcomes.

We have previously examined the relative role of maternal nutrition before and during pregnancy from the PRECONCEPT study and reported similar and independent associations of maternal PPBMI and GWG with infant size at birth (21). In addition, ultrasound data on fetal growth were used to pinpoint critical periods for GWG that were most influential on child growth. Early GWG in the first 20 weeks had 3 times the influence on birthweight compared to GWG ≥ 30 weeks (22); however, the long-term impacts remain unclear. Thus, our current objective was to examine the associations between PPBMI and timing of GWG on PPWR at 1, 2 and 6–7 years and maternal and child percent body fat and BMIZ at 6–7 years, using a prospective cohort of Vietnamese mothers and their children.

The baseline maternal and household characteristics, birth outcomes, and feeding practices were similar in our final analytic samples for maternal (n = 864) and child health (n = 1,381) outcomes Table 1. In brief, women were on average 26 years of age at preconception enrollment and nearly 80% worked as farmers. Around 10% of the infants were born preterm. Approximately 60% of infants were exclusively breastfeed, with the majority continuing to provide breastmilk for the first year (>95%) but not for the full 2 years (<5%). At the 6–7 years follow-up visit, children received on average 5 different food groups and had reported average physical activity levels. There were no significant differences in the analytic samples, with the exception of slight differences in maternal age at enrollment.

Average maternal BMI was 19.7 kg/m² at preconception with an average total GWG of 10.2 kg (Table 2). There was a steady increase in weight gain throughout pregnancy with an average weight gain per period of 1.7 kg (<20 weeks), 4.8 kg (20–29 weeks) and 4.3 (≥ 30 weeks). Women retained on average 1.1 kg, 1.5 kg and 4.3 kg at the 1, 2, and 6–7 years follow-up visits, respectively. At the 6–7 years follow-up visit, maternal percent body fat was 28.2 ± 6.2% and child percent body fat was 27.2 ± 4.5%.

We examined the relative association of maternal PPBMI and total CGWG z-scores with PPWR at 1, 2, and 6–7 years (Table 3). A one SD increase in PPBMI was associated with a decrease in PPWR at 1 (−0.21 [−0.37, −0.04]) and 2 years (−0.20 [−0.39, −0.01]) while a one SD increase in total CGWG was associated with an increase in PPWR at 1 (1.01 [0.85, 1.18]), 2 (0.95[0.76, 1.15]), and 6–7 years (1.05 [0.76, 1.34]). Additionally, early maternal CGWG in the first window (≤ 20 weeks) had the greatest association with PPWR at each time point: 1 (0.85 [0.68, 1.02]), 2 0.71 [0.49, 0.93]), and 6–7 years (1.03 [0.73, 1.33]). Maternal CGWG in the second window (<21–29 weeks) was also associated with PPBMI at the 1 year and 2- years time points (0.58 [0.37, 0.80] and 0.50 [0.24, 0.77] respectively), but to a lesser degree. Later CGWG (≥30 weeks); however, was not significantly associated with PPWR at any of the time periods.

Maternal PPBMI and CGWG was also associated with maternal and child percent body fat at 6–7 years postpartum (Table 4). A one SD increase in both PPBMI and total CGWG were associated with an increase of maternal percent body fat at 6–7 years (3.19 [2.85, 3.53] and 0.81 [0.43, 1.19] respectively). A one SD in PPBMI was also associated with an increase of child percent body fat at 6–7 years (0.79 [0.55, 1.02]). While total CGWG was not significantly associated with child percent body fat, when separated by timing of GWG, greater early (<20 weeks) and mid (20–29 weeks) CGWG was significantly associated with an increase in child percent body fat (0.42 [0.15, 0.69], 0.42 [0.09, 0.74], respectively. Furthermore, a one SD increase of PPBMI and total GWG were positively associated with child BMI z-score at 6–7 years (0.27 [0.21, 0.33] and 0.11[0.04, 0.17] respectively). For child BMIZ, CGWG in mid-pregnancy (<20–29 weeks) appeared to the most influential time period.

To examine these relationships further, we also divided participants into PPBMI and conditional GWG tertiles. As show in Figure 1, women in PPBMI tertile 3 (the largest PPBMI) had significantly lower PPWR at 1 year (−0.55 [−0.94, −0.15]) and 2 years (−0.55 [−1.01, −0.09]) but not at 6–7 years compared to women in PPBMI tertile 1 (the smallest PPBMI). Women with the largest CGWG (tertile 3) had significantly greater PPWR at 1 year (2.18 [1.76, 2.60]), 2 years (2.29 [1.80, 2.73]), and 6–7 years (2.57 [1.88, 3.27]) compared to women who gained the least amount of weight (tertile 1). Likewise, greater PPBMI (tertile 3) was positively associated with both maternal (7.03 [6.13, 7.93]) and child (1.24 [0.67, 1.81]) percent body fat at 6–7 years compared to PBMI tertile 1 (Figure 2). Greater maternal CGWG was associated with significantly higher maternal (2.23 [1.26, 3.20]) but not child percent body fat.

In our prospective Vietnamese cohort, maternal nutrition before and during pregnancy had long-term implications for PPWR and maternal and child body composition. Maternal PPBMI was negatively associated with PPWR at 1 and 2 years but not at 6–7 years; in contrast, maternal PPBMI remained a significant predictor of both maternal and child percent body fat at 6–7 years. Total CGWG was positively associated with PPWR at 1, 2, and 6–7 years as well as maternal percent body fat and child BMIZ at 6–7 years and most notably CGWG in both the early (<20 weeks) and mid (20–29 week) periods were positively associated with child percent body fat at 6–7 years. Early CGWG (<20 weeks) also had the strongest relative associations with PPWR at all time points compared to later weight gain, signaling that not only the total amount of weight gained during pregnancy but also the timing of weight gain may be an important consideration for future programs to address.

Our findings corroborate the findings from previous studies that showed positive associations between GWG and PPWR and negative associations between PPBMI and PPWR (17, 18) but were done mainly in high-income countries. An exception is a recent previous study among Vietnamese women, that reports that women with an underweight PPBMI had a significantly higher PPWR at 1 year compared to women with a normal PPBMI (adjusted mean PPWR of 3.7 kg vs. 2.34 kg, respectively) (33). Likewise, women with excessive GWG had a significantly higher PPWR at 1 year compared to those with adequate GWG (adjusted mean PPWR of 5.1 kg vs. 2.9 kg, respectively) (33). Our findings expand upon this work to utilize an advanced analytic approach that allowed us to examine the independent effects of maternal nutrition before and during pregnancy and compare the relative strength of associations using standardized z-scores. In our cohort of Vietnamese women, a one SD increase in total CGWG had nearly a 5-fold greater influence on PPWR compared to PPBMI.

A novel aspect of our study was the ability to further examine the unique and independent influence of the role of timing of GWG on PPWR. While growing evidence has indicated the potential importance of timing of GWG on infant size a birth (16, 22, 34, 35), to our knowledge, this is the first study to prospectively follow women from preconception throughout early childhood to examine the long-term associations with PPWR in a middle-income country. In our cohort early CGWG (<20 wk) had a 6 to 7-fold greater association with PPWR compared with later CGWG (≥30 wk). The mechanisms underlying these associations with early weight gain are unclear. One possible explanation is that the anabolic metabolic adaptations during early pregnancy may be related to increases in insulin sensitivity and/or increased maternal fat stores (36). In contrast, later pregnancy is a catabolic state that is associated with increased insulin resistance to facilitate nutrient transfer and fetal growth (36).

Another major contribution of our study is the examination of long-term associations with maternal and child body composition. In our cohort of Vietnamese women, PPBMI had nearly a 4-fold stronger relative association with maternal and child percent body fat at 6–7 years compared with total CGWG. We furthermore report new findings that early CGWC (<20 wks) had the strongest relative association with both maternal and child percent body fat, indicating the importance of early maternal nutrition for long-term body composition. While prior research has indicated that excessive GWG is associated greater gains in fat mass during pregnancy, few studies have tracked the long-term impact on body composition (37–39). In a prospective cohort of African American and Dominican mothers from New York, excessive GWG was associated with higher percent body fat and PPWR at 7 years among women who were underweight, normal and some overweight but not obese prior to pregnancy (39). The evidence on the role of timing of GWG on later body composition has been mixed. In the Avon Longitudinal Study of Parents and Children, among normal weight women, mid-pregnancy (19–28 wks) GWG was more strongly related to maternal adiposity and blood pressure at 16 years postpartum (40). In contrast, in a Boston pregnancy cohort, first trimester specific weight gain compared to later weight gain was most strongly associated with higher weight retention, maternal waist circumference and blood pressure (41). Even less is known on the long-term associations with child body composition, though greater GWG in the first trimester was associated with increased offspring overweight and obesity, higher waist circumference, skin fold thickness and blood pressure at 4 years in a prospective cohort from Greece (42). Further work is needed to understand context and timing specific findings.

Key strengths of our cohort include the prospective cohort design, accurate assessment of gestational age and maternal body size measurements from pre-pregnancy through pregnancy and post-partum. Most of the research to date estimate PPBMI based on maternal recall or measurements that are made at the first antenatal care visit when some women may have already started gaining weight. This is an important limitation particularly in light of the current findings on the importance of early GWG. Given the study design, i.e., longitudinal follow-up of a birth cohort, the results are subject to a number of biases related to attrition, non-response bias, representativeness of the sample. Although we had low loss to follow-up for the children and the final sample was comparable to those lost to follow-up. Our findings may not be generalizable to other settings with a greater prevalence of preconception obesity or excessive GWG, as in our context nearly a third of women entered pregnancy underweight and over two thirds gained below IOM recommendations (22). Nevertheless, our study provides valuable insights for similar low-resource settings undergoing nutrition transition with dual burdens of under and over nutrition (43). Most notably, we present novel data on maternal and child percent body fat at 6–7 years. While there are inherent limitations of the using BIA for assessing body composition (including potential role of hydration status, food intake, and previous physical exercise); the use of validated population-specific prediction equations strengthens our findings (44, 45). However, the lack of data on changes in maternal body composition or in the patterns of fat deposition during pregnancy limits our ability to examine how GWG is associated with changes in body composition and how this may be influenced by the timing of weight gain. Maternal fat accretion accounts for a substantial portion of GWG and consists of mammary tissue expansion, visceral adipose tissue, and subcutaneous adipose tissue (46–48). Visceral adipose fat surrounds intra-abdominal solid organs (49) and produces more pro-inflammatory adipokines and cytokines than subcutaneous adipose tissue (50–52) and is also associated with increased gestational metabolic disease including gestational diabetes and hypertension (53–56). A deeper understanding on the patterns of fat deposition could help explain underlying mechanisms of maternal nutrition and obesity and chronic disease risk. Furthermore, additional research with more detailed dietary intake data would be valuable to understand the role of both maternal and child diet.

In summary, we present novel findings on the long-term associations of maternal nutrition before and during pregnancy on PPWR as well as maternal and child body composition at 6–7 years in a study setting that is experiencing the nutrition transition represented by the dual burden of under and overnutrition and associated changes in diet quality and lifestyle (57). This work has important implications for targeting women of reproductive age as early GWG may represent a critical window when modifying maternal nutrition can have long-term consequences for maternal and child nutrition and be part of the strategies required to address the burden of non-communicable diseases like diabetes and cardiovascular disease that are the leading causes of morbidity and mortality globally (58, 59). Further research and replication of the unique analytic approach to examine the independent time effects of GWG in other settings may be merited and will help inform future revisions of GWG recommendations for different populations.

The data analyzed in this study is subject to the following licenses/restrictions: Data available upon request and approval of study team. Requests to access these datasets should be directed to UR, uramakr@emory.edu and PN, P.H.Nguyen@cgiar.org.

The studies involving human participants were reviewed and approved by Ethical Committee of Institute of Social and Medicine Studies in Vietnam and Emory University’s Institutional Review Board, Atlanta, Georgia, USA. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

MY, PN, RM, and UR: designed the research. PN, LT, and LK: conducted the field research. MY, LT, LK, and PN: analyzed data. MY, PN, SH, and UR: wrote the manuscript. MY: had primary responsibility for the final content of the manuscript. All authors reviewed and provided critical feedback.

This study was supported by Nestle Foundation, Micronutrient Initiative, Mathile Institute for Advancement of Human Nutrition.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.