Interpregnancy weight change as a potential risk factor for large-for-gestational-age infants: the Japan Environment and Children’s Study

Abstract Objective This study aimed to estimate the impact of interpregnancy weight change from the first to the second pregnancy on the risk of infants being large for gestational age (LGA). Methods This nationwide prospective birth cohort analysis included 3245 women who delivered their first two live singletons between 2011 and 2014. Interpregnancy weight change was calculated as the difference between the prepregnancy body mass index (BMI) of the first and second pregnancies. LGA infants were compared among three interpregnancy weight change groups: weight loss (a BMI loss >1 unit), weight gain (a BMI gain >1 unit), and stable weight (BMI maintained within − 1 to <1 unit). Interpregnancy weight change was assessed in mothers with a BMI <25 and ≥25 kg/m2, and adjusted odds ratios (ORs) were calculated for LGA infants by multiple logistic regression. Results The incidence of LGA infants was 8.6% (279 out of 3245). Compared with the stable weight group, interpregnancy weight gain was associated with an increased risk of infants being LGA (adjusted OR: 1.69, 95% confidence interval: 1.21–2.36) in the normal BMI (<25 kg/m2) group. In contrast, in the overweight/obese BMI (≥25 kg/m2) group, interpregnancy BMI was not a significant risk factor for LGA infants. Conclusions Accurate risk stratification using interpregnancy BMI could assist the clinical management of women with a normal BMI who are at risk of delivering LGA infants.


Introduction
Being large for gestational age (LGA), defined as a birth weight above the 90th percentile, is associated with increased perinatal mortality, morbidity and a risk of developing obesity and insulin resistance in the future [1][2][3]. For mothers, the delivery of LGA infants increases the risk of postpartum hemorrhage, cesarean section, and third-and fourth-degree tears [4]. Similarly, LGA neonates are at an increased risk of shoulder dystocia, hypoglycemia, and neonatal intensive care unit admission [4][5][6]. Previous studies suggested that prenatal identification of LGA infants and appropriate delivery management of suspected LGA infants are effective methods to prevent delivery-related perinatal morbidity [4,7]. Therefore, it is important to determine the causes of infants being LGA and develop strategies to improve early detection. A high birth weight is affected by several factors, such as maternal prepregnancy weight, gestational diabetes mellitus (GDM), overt diabetes in pregnancy, male sex, multiparity, advanced maternal age, excessive maternal weight gain, and increased body mass index (BMI) [8][9][10][11][12][13][14]. Interpregnancy BMI, which is the difference between the first and second pregnancy maternal BMI recorded at the time of the first antenatal visit, has been used to assess the risk of infants being LGA in recent years [14][15][16]. In a recent systematic review and meta-analysis, interpregnancy BMI was positively associated with the risk of infants being LGA [14]. However, to the best of our knowledge, no study has assessed the relationship between BMI and LGA infants in a Japanese population. Furthermore, it remains unclear whether this relationship is relevant to the Japanese population since maternal physique differs by ethnicity and race [17]. Particularly, pregnant women in Japan are physically shorter and thinner than those in Western European countries [17]. Despite a plausible rationale for weight control as part of interpregnancy planning, a long-standing knowledge gap exists between healthcare providers and women of reproductive age. An improved understanding of the relationship between interpregnancy BMI and LGA infants might help solve this problem [14]. In contrast, lower maternal prepregnancy BMI increases the risk of preterm birth, low birth weight, and small-forgestational-age (SGA) infants [18]. Therefore, excessive interpretation of the BMI decrease is presumably undesirable in terms of perinatal complications. To reduce perinatal complications, avoiding only excessive interpregnancy weight gain is preferred (i.e. "stable interpregnancy BMI" is important). This study aimed to estimate the impact of interpregnancy weight change from the first to the second pregnancy on the risk of infants being LGA using data from a nationwide prospective birth cohort study with a large sample size, the Japan Environment and Children's Study (JECS).

Study population
The JECS is an ongoing nationwide prospective birth cohort study, conducted at 15 Regional Centers (Hokkaido, Miyagi, Fukushima, Chiba, Kanagawa, Koshin, Toyama, Aichi, Kyoto, Osaka, Hyogo, Tottori, Kochi, Fukuoka, and South Kyusyu/Okinawa) in Japan. Details of the JECS have been described previously [19][20][21]. This study was conducted in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology statement. Pregnant women were recruited for this study between January 2011 and March 2014. The eligibility criteria were as follows: residence in the Study Areas at the time of recruitment, expected delivery date after August 2011, comprehension of the Japanese language, and completion of a self-administered questionnaire. A total of 104,062 fetal records were included in this cohort. This study used the jecs-ta-201901930-qsn dataset, which was released in October 2019 and revised in February 2020. Only the first two successive pregnancies were assessed to minimize the confounding impact of parity on pregnancy outcomes. Moreover, we included only mothers with complete obstetric and demographic data regarding the two eligible pregnancies. Finally, data were divided according to maternal prepregnancy BMI ("normal": <25 kg/m 2 , "overweight/obese": !25 kg/m 2 ) to examine the effects of interpregnancy weight change in the overweight/obese population compared with women with a normal BMI, in accordance with a previous study [14].

Ethics statement
The JECS protocol was reviewed and approved by the Ministry of the Environment's Institutional Review Board on Epidemiological Studies and the Ethics Committees of all participating institutions. Written informed consent was obtained from all participants. JECS was conducted in accordance with the Declaration of Helsinki and other national regulations and guidelines.

Data collection
The study participants completed questionnaires throughout pregnancy (i.e. during the first and second/third trimesters). Medical records at the time of registration and immediately after vaginal delivery or cesarean section were transcribed by physicians, midwives/nurses, and/or research Co-ordinators. Information regarding maternal factors was obtained from the questionnaires completed during pregnancy. Obstetric records included maternal age at delivery, birth weight, infant sex, presence of GDM or overt diabetes during pregnancy, presence of hypertensive disorder of pregnancy (HDP), the use of assisted reproductive technology (ART, in vitro fertilization or intracytoplasmic sperm injection), parity, gestational age at delivery, maternal stature, prepregnancy weight at the first and second pregnancies, and body weight at the last prenatal visit within one week of birth. Prepregnancy BMI was calculated according to the World Health Organization standard (body weight/square of height [kg/m 2 ]) [14]. GDM was diagnosed in the presence of at least one abnormal plasma glucose value (!92, 180, and 153 mg/dL for fasting, and 1-h and 2-h postprandial plasma glucose concentrations, respectively) after a 75-g oral glucose tolerance test [10]. HDP was defined as a blood pressure !140/90 mmHg [22]. Maternal weight gain during pregnancy was calculated by subtracting the patient's body weight from the value at the last prenatal visit [14]. Excessive maternal weight gain was defined according to prepregnancy BMI [23]. Women were categorized into four groups (G) for analysis, according to their prepregnancy BMI, i.e. G1, <18.5 kg/m 2 ; G2, 18.5 to <25.0 kg/m 2 ; G3, 25.0 to <30.0 kg/m 2 ; and G4, !30.0 kg/m 2 . Excessive weight gain was defined as >15 kg for G1; >13 kg for G2; >10 kg for G3; and >5 kg for G4 [23].

Exposure definitions
We calculated interpregnancy BMI change as the difference between the BMI at the beginning of the first and second pregnancies. We categorized BMI change as less than À1 (a BMI loss >1 unit), À1 to less than 1 (the reference: stable interpregnancy group), and 1 or more (a BMI gain >1 unit) units with reference to several relevant publications [14][15][16].

Outcomes
LGA infants was the primary outcome of interest in our study and was defined as birth weight above the 90th percentile at each gestational week [24]. SGA infants were defined as infants with a weight below the 10th percentile at each gestational week [24].

Statistical analysis
One-way analysis of variance and the chi-square test were used to evaluate the association between LGA infants and potential confounding factors. One-way analysis of variance was used to analyze continuous variables, such as birth weight. The chi-square test was used for categorical variables, such as the incidence of obstetric complications.
Multiple logistic regression analyses were conducted to estimate the association between interpregnancy BMI change and LGA infants. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated after adjusting for potential confounders. All analyses were performed using SPSS Statistics for Windows software version 25 (IBM Corp., Armonk, NY). The significance level was set at p < .05.

Results
A total of 3566 women with their first two consecutive singleton deliveries were included in this cohort. There were 321 (9.0%) women who met the exclusion criteria. We excluded women whose first or second pregnancy was a miscarriage or stillbirth (n ¼ 113). We also excluded women with missing data (birth weight at first pregnancy, n ¼ 125; delivery after 42 weeks of gestation at first pregnancy, n ¼ 12; maternal age at second pregnancy, n ¼ 10; method of conception at second pregnancy, n ¼ 6; birth weight at second pregnancy, n ¼ 3; prepregnancy maternal weight at second pregnancy, n ¼ 1; maternal weight at delivery at second pregnancy, n ¼ 50; and delivery after 42 weeks of gestation at second pregnancy, n ¼ 1). Therefore, a total of 3245 (91.0%) women were eligible for inclusion in this study (Supplementary Figure 1). According to the prepregnancy BMI of the second pregnancy, 2883 (88.0%) mothers were normal weight (<25 kg/m 2 ), and 392 (12.0%) mothers were obese (!25 kg/m 2 ). The mean maternal ages at the first and second pregnancies were 28.2 ± 4.7 years and 30.1 ± 4.8 years, respectively. The mean maternal prepregnancy BMI at the first and second pregnancies were 21.0 ± 3.2 kg/m 2 and 21.4 ± 3.5 kg/m 2 , respectively. There were 1624 male infants (50.0%), 133 women with preterm delivery (4.1%), 49 women with ART (1.5%), and 563 (17.3%) cesarean second singleton deliveries. Supplementary Table 1 summarizes the clinical characteristics of the women enrolled in this study. In the second pregnancy of the normal BMI group (<25 kg/m 2 ), there was a statistically significant difference in maternal age, birth weight, weight gain during pregnancy, and prepregnancy BMI at the first and second pregnancies among the three groups according to BMI changes. In the second pregnancy of the overweight/obese BMI group (!25 kg/m 2 ), there was a statistically significant difference in maternal age and prepregnancy BMI at the first pregnancy among the three groups according to BMI changes. Table 1 presents the prevalence of SGA and LGA infants in the second pregnancy according to the interpregnancy BMI change. In the second pregnancy of the normal BMI group (<25 kg/m 2 ), there was a statistically significant difference in the prevalence of LGA, but not SGA, infants according to the interpregnancy BMI change. In the second pregnancy of the overweight/obese BMI group (!25 kg/m 2 ), there was no statistically significant difference in the prevalence of LGA and SGA infants according to the interpregnancy BMI change. In the second pregnancy of the normal BMI group (<25 kg/m 2 ), after adjusting for potential confounding factors, BMI gain (!1 unit) was associated with an increased risk of infants being LGA (adjusted OR: 1.69, 95% CI: 1.21-2.36) ( Table 2).
Conversely, in the second pregnancy of the overweight/obese BMI group (!25 kg/m 2 ), after adjusting for potential confounding factors, interpregnancy BMI was not significantly associated with LGA infants.

Discussion
In this nationwide prospective birth cohort study, the following two important clinical findings were observed: (1) an interpregnancy BMI increase of !1 kg/m 2 in mothers with a normal BMI (<25 kg/m 2 ) in the second pregnancy increased the risk of infants being LGA but did not affect the prevalence of SGA infants, and (2) the interpregnancy BMI in mothers with overweight/obese BMI (!25 kg/m 2 ) in the second pregnancy did not affect the risk of infants being LGA or SGA.
There is a consensus that birth weight is directly related to insulin sensitivity, indicating that maternal glucose metabolism plays an important role in fetal growth [11,25,26]. If maternal glycemic control is impaired and maternal serum glucose levels are high, glucose crosses the placenta [11,25,26]. However, maternally derived insulin does not cross the placenta [11,25,26]. As a result, in the second trimester, the fetal pancreas, which is now capable of secreting insulin, starts to respond to hyperglycemia and secretes insulin autonomously (i.e. regardless of glucose stimulation) [11,25,26]. This combination of hyperinsulinaemia (insulin as a main anabolic hormone) and hyperglycemia (glucose as a main anabolic fuel) leads to an increase in fetal fat and protein storage, resulting in LGA infants [11]. Previous studies reported that overweight or obese pregnant women have an approximately 50% increase in insulin insensitivity compared with normal-weight pregnant women [27], and the risk of developing GDM has been positively associated with obesity [15]. Considering the pathophysiology of maternal glucose metabolism, it seems reasonable to speculate that a BMI increase is significantly associated with LGA infants in women with a normal BMI (<25 kg/m 2 ) in the second pregnancy. Moreover, obesity itself is a risk factor for LGA infants regardless of the interpregnancy BMI. As a result, interpregnancy BMI did not contribute to the development of LGA infants in mothers with an overweight/obese BMI (!25 kg/m 2 ) in the second pregnancy.
Although slight differences in the cutoff values of increased interpregnancy BMI and the study populations were previously reported, several studies have indicated an association between increased interpregnancy BMI and LGA infants [14][15][16]. The findings of our study are consistent with these previous findings only for the normal BMI group (<25 kg/m 2 ) in the second pregnancy [14][15][16]. However, our study analyzed more risk factors for LGA infants with a relatively large sample size in a Japanese population. Although it is well known that glucose metabolism varies by race [1,12], to the best of our knowledge, this is the first study to investigate the association between increased BMI and LGA infants in a Japanese population. In contrast, a population-based retrospective cohort analysis of 10,444 obese women in the United 14 Notes: Data are presented as numbers (%). BMI: body mass index; LGA: large-for-gestational-age; SGA: small-for-gestational-age. States demonstrated that weight loss in obese women reduced the risk of subsequent birth of LGA infants without increasing the risk of SGA infants [28]. This discrepancy between our study and this previous study [28] might be related to differences in the study populations (i.e. unlike our study, the obese group of this previous study consisted of women with prepregnancy BMI !30 kg/m 2 [28]). According to the sample size calculation method previously reported and widely used for logistic regression analysis, "Ten Events Per Variable" is a widely adopted minimal guideline criterion [29]. Based on this method, if we performed our study with the inclusion of only women with prepregnancy BMI !30 kg/m 2 , at least 60 LGA infants in the second pregnancy would be required (six variables were used: excessive weight gain, male infant, advanced maternal age, LGA infant at first pregnancy, glucometabolic disorder, and interpregnancy BMI). As our study only included 20 LGA infants from mothers with a BMI !30 kg/m 2 at the second pregnancy, we could not perform an analysis with sufficient power to detect a risk difference. However, at the second pregnancy, there was a statistically significant difference in the prevalence of LGA infants in the BMI <30 kg/m 2 group. On the other hand, there was no significant difference in the prevalence of LGA infants in the BMI !30 kg/m 2 group (Supplementary Table 2). This result was similar to that of our analysis presented in Table 1, which classified women using a BMI threshold of 25 kg/m 2 . Pregnant women in Japan are physically thinner than those in Western European countries [17], and only 92 (2.8%) obese women (BMI !30 kg/m 2 ) were included in this study. The recruitment of a sufficient number of obese pregnant women for analysis in Japan is difficult.
There was no statistically significant difference in the prevalence of SGA infants according to interpregnancy BMI in the normal BMI (<25 kg/m 2 ) or obese/overweight BMI (!25 kg/m 2 ) groups at the second pregnancy. However, it is difficult to provide detailed comments on the relationship between BMI and SGA infants in this study because we only analyzed the prevalence of SGA infants according to interpregnancy BMI. Unlike European countries and the United States, where maternal obesity is a health problem [30], the proportion of underweight women of childbearing age (i.e. BMI <18.5 kg/m 2 ) is increasing in Japan [31]. The prevalence of underweight individuals has been reported to be 11.5% in Japan [32], 7.4% in the United States [33], and 4.3% in the United Kingdom [34]. In this study, 510 (14.3%) women were underweight.
In Japan, where being underweight and having poor weight gain during pregnancy cause various perinatal problems [18,31], the association between BMI and SGA infants requires careful interpretation. In the future, we would like to conduct a detailed analysis of the association between BMI and SGA infants, including potential confounding factors. To the best of our knowledge, no reports have claimed that being underweight before pregnancy is a risk factor for LGA. However, considering the high incidence of perinatal complications in underweight women, we examined the association between interpregnancy weight change from the first to the second pregnancy and LGA in the 2349 women whose BMI was between 18.5 and 25.0 kg/m 2 (we excluded 510 underweight women [<18.5 kg/m 2 ] from the 2859 normal weight [<25 kg/m 2 ] women). As a result, after adjusting for potential confounding factors, BMI gain (!1 unit) was associated with an increased risk of infants being LGA (adjusted OR: 1.60, 95% CI: 1.13-2.27), and BMI loss (< À1 unit) was not associated with an increased risk of infants being LGA (adjusted OR: 0.89, 95% CI: 0.50-1.56) (data not shown). In other words, the results of the association between interpregnancy weight change and LGA did not change regardless of the inclusion of underweight (<18.5 kg/m 2 ) women in the study population.
The strength of this study was the inclusion of data from a nationwide cohort study with a large sample size. The JECS, which has 100,000 participants, is the largest nationwide birth cohort study in Japan, and it is considered representative of the general population. Furthermore, the outcome measurements were reliable because the pregnancy and delivery information were based on medical records transcribed by doctors, research coordinators, nurses, and midwives. However, this study had several limitations. First, accurate data regarding several genetic disorders, such as the Beckwith-Wiedemann syndrome [35] and the Simpson-Golabi-Behmel syndrome [36], which are potential risk factors for LGA infants, were not considered in this study. However, it is difficult to accurately identify all genetic disorders in a nationwide cohort study. Second, the JECS study is composed almost exclusively of Japanese participants, with only a very small number of non-Japanese. The generalizability of our findings may be limited by the homogeneity of this cohort, which included almost exclusively Japanese women. Finally, although this is not a limitation, BMI gain was significantly more common in the overweight/obese group than in the normal weight group (62.2% vs 21.8%; p < .001). In Japan, many individuals of reproductive age strongly desire to be slim and work toward it [18]. This study's findings may reflect the fact that this desire is more pronounced among those in the normal weight group as compared to those in the overweight/obese group. The association between prepregnancy BMI and BMI gain may be one of the key themes for improving prognosis in perinatal care in the future.
In conclusion, we found an increased risk of infants being LGA when mothers had a normal BMI (<25 kg/m 2 ) at the second pregnancy, with an interpregnancy BMI increase of !1 kg/m 2 in the Japanese population. Although counseling women regarding the importance of proper interpregnancy BMI requires a detailed study of the association between interpregnancy BMI and SGA infants, accurate risk stratification using interpregnancy BMI could assist the clinical management of women with normal BMI who are at risk of delivering LGA infants.