Effects of gestational weight gain and HbA1C on fetal growth and birthweight in women with obesity

Abstract

Introduction

Fetal growth and birthweight are influenced by various endogenous and exogenous factors during pregnancy. Pregnant women with obesity face higher rates of complications, which can impact fetal growth. This study aims to explore how gestational weight gain and metabolic factors, such as HbA1C and adipokines, are associated with fetal growth in the second and third trimesters and with birthweight in Norwegian women with a pregestational BMI ≥35 kg/m².

Material and Methods

We conducted a prospective longitudinal cohort study at the Department of Gynecology and Obstetrics at Drammen Hospital from 2016 to 2019. We included 163 pregnant women (nulliparous 47.6%) with pregestational BMI ≥35 kg/m² and singleton pregnancy, excluding type 1 and 2 diabetes. Gestational weight gain (GWG) was calculated as the difference between pregestational weight and the last weight before delivery. Fetal growth was monitored via ultrasonography at five time points, alongside fasting metabolic samples withdrawn around 18 and 36 weeks' gestation. Statistical analyses were conducted using linear regression in Stata Version 17.0.

Results

Mean birthweight was within normal ranges (mean z‐score 0.117). However, the proportions of LGA (18.4%) and SGA (12.9%) were high. One in three women was diagnosed with GDM, of whom approximately one in three needed medical treatment. In the adjusted multivariate analyses, only HbA1C measured around gestational week 36 and the weekly weight gain remained positively associated with birthweight, biparietal diameter, and mean abdominal diameter.

Conclusions

In this population of pregnant women with pBMI ≥35 kg/m², weekly GWG calculated from weight measurement around 36 weeks' gestation was positively associated with birthweight (z‐score), mean abdominal diameter (MAD), and biparietal diameter (BPD). HbA1C was positively associated with birthweight (z‐score). Our results indicate that HbA1C and the modifiable GWG are exposures with the potential to influence fetal growth and birthweight. Nevertheless, current follow‐up protocols do not routinely emphasize these factors. Their evaluation therefore warrants further investigation.

Our results indicate that HbA1C measured in the third semester and the modifiable gestational weight gain are exposures with the potential to influence fetal growth and birthweight in this population of pregnant women with pBMI ≥35 kg/m².

Abbreviations

BPD

biparietal diameter

GDM

gestational diabetes mellitus

GWG

gestational weight gain

HbA1C

glycated hemoglobin

IQR

interquartile range

MAD

mean abdominal diameter

OGTT

oral glucose tolerance test

pBMI

pregestational body mass index

SD

standard deviation

Key points.

Our results indicate that HbA1C and the modifiable gestational weight gain are exposures with potential to influence fetal growth and birthweight in this population of pregnant women with pregestational BMI ≥35 kg/m². Current follow‐up practices may be improved.

1. INTRODUCTION

Fetal growth and birthweight may provide insights into placental function, fetal well‐being, and potential future health risks for both the child and the mother and are influenced by multiple endogenous and exogenous factors during pregnancy. ¹ , ² In addition to well‐known factors, such as gestational weight gain (GWG), parity, maternal age, body mass index (BMI), and fetal sex, metabolic factors have been shown to be associated with birthweight. Adiponectin, which is involved in reducing insulin resistance, has shown a negative association with birthweight, ³ , ⁴ , ⁵ while glucose, HbA1C, insulin, and leptin have been positively associated with birthweight. ⁶ , ⁷ , ⁸ , ⁹

Women with pregestational obesity have a higher incidence of numerous antepartum, intrapartum, and postpartum complications, compared to women with a normal range pregestational BMI (pBMI). ¹⁰ The LifeCycle Project, which included 196670 participants, demonstrated that more than 50% of women with a pBMI ≥35 kg/m² and more than 60% of women with a pBMI ≥40 kg/m² experienced pregnancy complications. ¹¹ Common complications included disturbances in maternal glucose metabolism, resulting in gestational diabetes (GDM) and deviations in fetal growth and birthweight, including intrauterine growth restriction (IUGR), small for gestational age (SGA), and large for gestational age (LGA) neonates. ⁴ , ¹⁰ , ¹¹ , ¹²

A common risk‐stratification approach for detecting complication‐prone individuals, due to disturbances in glucose metabolism in pregnant women classified as overweight or obese, is testing for preexisting diabetes mellitus type 2 by an HbA1C‐test prior to 16 weeks' gestation, and for development of GDM by standardized oral glucose tolerance testing (OGTT) around 24–28 weeks' gestation. ¹³ , ¹⁴ Late onset GDM is not detectable by this strategy. ¹⁵ Increased HbA1C in the third trimester has been linked to LGA‐neonates, but this is not well documented in women without GDM and preexisting diabetes mellitus. ¹⁶ Women with pregestational obesity, in Norway, are recommended to engage in daily moderate physical activity >30 min, adhere to a balanced diet, ¹³ , ¹⁴ and follow US Institute of Medicine (IOM) recommendations for gestational weight gain (GWG) totaling 5–9 kg or 0.2 kg per week. ¹⁷ The optimal GWG for women with pregestational obesity has been an object of discussion. Some argue that the recommended GWG limits are too broad and excessively high. ⁴ , ¹² The GWG is complex and influenced by many factors, including maternal fluid retention, maternal gestational age, maternal medical conditions such as inflammatory diseases, medication, eating disorders, fetal growth, placenta size, and amniotic fluid, among others. ⁴ , ⁶ , ¹⁰ , ¹¹ , ¹⁷ The number of contributing influences makes GWG a potentially unreliable predictive factor, yet it is regularly applied in studies. Due to potential impact on fetal growth, regular fetal biometry check‐ups are recommended to assess whether estimated fetal weight and growth curves are within recommended limits. ¹³ , ¹⁸

In women with pregestational obesity, most previous studies have examined the associations between maternal BMI, GDM, and excessive GWG with infant birthweights. There is less knowledge concerning their associations with fetal growth in utero.

In the present study, we explored how weekly GWG and the metabolic consequences of high pBMI, particularly relating to glucose metabolism and adipokines, were associated with fetal growth in the second and third trimesters as well as with infant birthweight. The study was conducted in a Norwegian population of women with pBMI ≥35 kg/m².

2. MATERIAL AND METHODS

“Healthy mother – lifelong health” was a prospective, longitudinal cohort study conducted at Drammen Hospital, a regional referral center for high‐risk obstetrics with approximately 1.900 deliveries annually, serving a catchment area of approximately 500 000 inhabitants. We recruited pregnant women attending prenatal check‐ups at the hospital from March 2016 to December 2018, as delineated in our publication from 2024. ¹⁹ Briefly, the inclusion criteria were pBMI ≥35 kg/m² and referral by a general practitioner or a midwife before 22 weeks' gestation. We decided to use a pBMI cut‐off of ≥35 kg/m² rather than ≥30 kg/m² in line with the Norwegian national guidelines on antenatal care of pregnant women with pregestational obesity. ²⁰ Exclusion criteria were type 1 and 2 diabetes and multiple pregnancy. Each woman had at least eight antenatal visits with a midwife and at least four antenatal visits with an obstetrician during follow‐up (except for those with preterm births, before 37 weeks' gestation). The data collection included clinical data at study entry and from scheduled appointments with medical and obstetric history, self‐reported height and pregestational weight from which a pBMI was calculated. Preeclampsia was defined according to the Norwegian national obstetric guidelines issued in 2016 by the Norwegian Society of Obstetrics and Gynecology. ²¹

Gestational weight gain (GWG) was calculated by subtracting pregestational weight from the last measured weight before delivery. Weekly GWG was calculated by dividing the total GWG by the number of gestational weeks at birth. At inclusion, women were informed that the recommended GWG range was 5–9 kg, as advised by the American National Academy of Medicine (formerly the Institute of Medicine). ²²

Fetal growth was monitored by ultrasonographic abdominal measurements conducted by trained midwives at the first visit and by trained obstetricians at the rest of the visits, utilizing Voluson E6 machines from GE Healthcare, with 5‐MHz curvilinear transducers. Biparietal diameter (BPD) and mean abdominal diameter (MAD) were measured around gestational weeks 18 (MAD0, BPD0), 24 (MAD1, BPD1), 32 (MAD2, BPD2), 36 (MAD3, BPD3), and around 39 (MAD4, BPD4). BPD was measured as described by Eik‐Nes, ²³ by placing the calipers at the outer border of the cranium on both sides, at the level of the cavum septum pellucidi and the thalamus. MAD was computed by measuring the anterior–posterior and transverse diameters of the fetal abdomen where the umbilical vein enters the liver, with calipers placed at the outer border of the skin. ²³ For each BPD and MAD measurement, a mean was derived from three separate readings.

The due date was determined by BPD or femoral length measured at approximately 18 weeks' gestation and no later than 22 weeks', using the eSnurra algorithm. ²⁴ SGA and LGA were defined as sex‐ and gestational age–adjusted birthweight less than the 10th percentile and greater than the 90th percentile, respectively, based on the algorithm by Skjaerven (2000). ²⁵

Blood samples were drawn after an overnight fast at two separate visits, the first at around 18 weeks' gestation (IQR: 15–21 weeks) and the second at around 36 weeks' gestation (IQR: 36–36 weeks). We measured leptin, HbA1C, and adiponectin levels. The oral glucose tolerance test (OGTT) was performed twice, the first at around 18 weeks' gestation and again at 28 weeks' gestation if the first test was normal; the blood sample was drawn 2 h after ingestion of a 75‐g glucose solution. This timing was based on the Norwegian national obstetric guidelines at the time of study inclusion. In the new Norwegian GDM criteria from 2017, the fasting glucose threshold was lowered from 6.9 to 5.3 mmol/L, and the 2‐h glucose threshold after OGTT was increased from 7.8 to 9.0 mmol/L. 1‐h glucose levels are not routinely measured in the Norwegian OGTT protocol. Women included before implementation of these new criteria were diagnosed according to the old threshold, while those assessed after were diagnosed according to the new thresholds.

Glycated hemoglobin (HbA1C) was analyzed using the Tosoh G8 (Tosoh Bioscience, San Francisco, CA, USA) at Drammen Hospital.

Levels of leptin and adiponectin were analyzed at the Hormone Laboratory at Oslo University Hospital. Leptin was analyzed by radioimmunoassay (Merck Millipore, Burlington, MA, USA) until September 2018 and thereafter by ELISA (Mediagnost, Reutlingen, Germany). Adiponectin was analyzed by radioimmunoassay (Merck Millipore, Burlington, MA, USA).

2.1. Statistical analysis

Continuous data are presented as the mean and standard deviation (SD; normally distributed data) or the median and interquartile range (IQR, 25th and 75th percentiles; skewed data), whereas categorical data are presented as numbers and percentages. The statistical significance level was set at 5%. We used linear regression to estimate associations between the 8 predictor variables and the five outcome variables, with 95% confidence intervals and corresponding p‐values, reflecting one‐unit changes in the predictor variables. The associations were assessed using crude and multivariate models. We tested the following outcome variables: birthweight (z‐score), SGA and LGA, MAD1‐4 and BPD1‐4.

We adjusted the outcome variables for gestational age (continuous) and sex (dichotomous, female and male). We applied a table‐wise Bonferroni correction, dividing the nominal significance level ($0.05$) by the number of association tests conducted within each table ($n=8$), with a new table‐wise significance level at 0.006. All statistical analyses were performed using Stata Version 17.0 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC).

3. RESULTS

We included 163 women with a pBMI ≥35 kg/m², giving a median pBMI of 38.4 kg/m² (IQR 36.1–41.0). Demographic and clinical characteristics of the participants are presented in Table 1. In this cohort, 19.6% remained within the recommended GWG range of 5 to 9 kg; 30.7% gained less than 5 kg, and 49.7% gained more than 9 kg. The proportions of women with GDM and preeclampsia were higher than in the background pregnancy population in Norway during the same time period: 35.6% vs. 5.1% and 6.8% vs. 2.6%, respectively. ²⁶ Of the 163 women, 79 (47.6%) were tested for GDM based on the old criteria, while the remaining study participants, included after the GDM criteria update in 2017, were tested based on the new criteria. Of the 20 women who required medical treatment for GDM, 17 (85.0%) received insulin treatment, while three (15.0%) were treated with metformin. At least one comorbidity in addition to obesity was reported by 53% of participants, and 43% used one or two types of medication during the index pregnancy.

TABLE 1.

Descriptive characteristics of the study population.

	Mean	SD	Median	IQR	N	%
Women, n					163
Maternal characteristics
Age, years	30.0	4.9
Ethnicity:
African					6	3.7
Asian					14	8.6
European					143	87.7
Education >12 years					72	44.2
Primiparity					77	47.2
Smoking during pregnancy					10	6.1
pBMI, kg/m2			38.4	36.1–41.0
Weekly GWG, kg	0.2	0.2
GWG, kg	8.9	7.6
GDM					58	35.6
Preexisting hypertension					4	2.5
Gestational hypertension					6	3.7
Preeclampsia					11	6.8
Fetal characteristics
Sex, female					83	50.9
Gestational age at birth, weeks			40.1	38.9–40.9
Birthweight, grams	3574.9	651.9
Small for gestational age					21	12.9
Large for gestational age					30	18.4

Abbreviations: GDM, gestational diabetes mellitus (before 2017 criteria: fasting glucose 6.0–6.9 mmol/L or 2‐hourh 7.8–11.0 mmol/L; after 2017 criteria: 5.3–6.9 mmol/L, 2‐hourh 9.0–11.0 mmol/L); IQR, interquartile range; Large for gestational age, birthweight above the 90th percentile using z‐scores ²⁵ ; pBMI, pre‐pregnancy body mass index; Preeclampsia, hypertension and proteinuria in pregnancy after 20th gestational week; SD, standard deviation; Small for gestational age, birthweight below the 10th percentile using z‐scores ²⁵ ; Weekly GWG, gestational weight gain (calculated from pre‐pregnancy weight and last weight before delivery) divided by gestational age at birth (weeks).

Table 2 presents a timeline for the measurements of maternal weight and blood samples as well as fetal biometric ultrasound measurements. Table 3 presents the results of crude, adjusted, and multivariate linear regression analyses, including estimated differences in the means of eight maternal metabolic measures. The biometric measurements are illustrated in Figures 1 and 2 and show that the growth of both BPD and MAD became more diverse after 30 weeks' gestation (BPD2–4 and MAD2–4), with the spread appearing to be greater in MAD compared to the BPD. The 10th, 50th, and 90th percentiles in Figures 1 and 2 are derived from Norwegian growth charts published by Johnsen et al. (2006). ²⁷

TABLE 2.

Overview of data collection: Biometric ultrasound measurements, maternal weight measurements, and blood analyses.

Gestational Week (Visit)
Measurement Type	<16	~18 (0)	24 (1)	32 (2)	36 (3)	~39 (4)
BPD (mm), mean (SD)		44.8 (2.9)	62.6 (4.1)	85.8 (3.5)	93.6 (3.2)	97.7 (3.5)
MAD (mm), mean (SD)		42.2 (3.5)	63.0 (4.5)	90.1 (5.5)	102.8 (5.7)	112.7 (6.8)
HbA1C (%), mean (SD)	5.2 (0.3)				5.4 (0.4)
OGTT week, mean (SD)		18.3 (3.4)	27.5 (4.1)
Leptin (pmol/L), median (IQR)		4193 (3157–5125)			4376 (3394–5125)
Adiponectin (nmol/L), median (IQR)		246 (190–303)			262 (207–316)
Weight (kg), mean (SD)	108.9 (11.0)		110.2 (11.6)	113.0 (12.0)	116.1 (12.6)	117.7 (13.2)

Abbreviations: BPD, biparietal diameter; HbA1c, glycated hemoglobin; IQR, interquartile range; MAD, mean abdominal diameter; OGTT, oral glucose tolerance test; SD, standard deviation.

TABLE 3.

Overview of associations between exposure and outcome variables from univariate and multivariate linear regression analyses.

Exposure	MAD2 Univariate				MAD2 adjusted Multivariate				MAD4 Univariate				MAD4 adjusted Multivariate
	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value
Sex, female	162	0.061	(−1.648, 1.770)	0.944					124	0.144	(−2.293, 2.581)	0.907
Age, years	162	0.051	(−0.125, 0.228)	0.567	117	0.029	(−0.201, 0.260)	0.801	124	−0.021	(−0.281, 0.239)	0.871	98	0.070	(−0.253, 0.393)	0.669
Para0, nulliparous	162	−0.803	(−2.728, −0.022)	0.354	117	−0.939	(−3.069, 1.191)	0.384	124	−0.592	(−3.032, 1.848)	0.632	98	−0.328	(−3.178, 2.523)	0.820
pBMI	162	0.149	(−0.098, 0.395)	0.235	117	0.352	(0.056, 0.648)	0.020	124	−0.357	(−0.174, 0.001)	0.050	98	−0.107	(−0.497, 0. 283)	0.588
Weekly GWG, kg	162	5.784	(1.620, 9.948)	0.007	117	7.715	(2.616, 12.814)	0.003*	124	8.679	(2.312, 15.045)	0.008	98	7.783	(0.152, 15.413)	0.046
Leptin 18 weeks' gestation	131	−0.001	(−0.001, 0.000)	0.372	117	−0.001	(−0.002, 0.001)	0.023
Leptin 36 weeks' gestation									108	0.001	(−0.001, 0.001)	0.976	98	−0.001	(−0.001, 0.001)	0.808
Adiponectin 18 weeks' gestation	129	−0.008	(−0.018, −0.001)	0.096	117	−0.003	(−0.014, 0.007)	0.493
Adiponectin 36 weeks' gestation									108	−0.011	(−0.024, 0.003)	0.122	98	−0.006	(−0.019, 0.008)	0.402
GDM, yes	162	1.475	(−0.292, 3.243)	0.101	117	−0.222	(−2.523, 2.078)	0.848	124	0.673	(−1.881, 3.227)	0.603	98	−0.6309	(−3.652, 2.390)	0.679
HbA1C, <16 weeks' gestation
HbA1C 36 weeks' gestation	143	2.483	(0.057, 4.908)	0.045	117	1.587	(−1.666, 4.839)	0.070	108	3.028	(−0.230, 6.286)	0.068	98	3.419	(−0.475, 7.314)	0.085

Exposure	BPD2 Univariate				BPD2 Adjusted Multivariate				BPD4 Univariate				BPD4 Adjusted Multivariate
	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value
Sex, female	160	−0.933	(−2.018, 0.152)	0.091					123	−1.066	(−2.317, −0.184)	0.094
Age, years	160	−0.011	(−0.125, 0.103)	0.850	116	0.031	(−0.099, 0.162)	0.634	123	−0.024	(−0.159, 0.111)	0.725	97	0.098	(−0.06, 0.256)	0.222
Para0, nulliparous	160	0.666	(−0.235, 0.083)	0.230	116	0.734	(−0.477, 1.946)	0.232	123	0.832	(−0.428, 2.093)	0.194	97	0.890	(−0.514, 2.294)	0.211
pBMI	160	−0.076	(−0.235, 0.083)	0.347	116	0.031	(−0.140, 0.203)	0.720	123	−0.184	(−0.372, 0.004)	0.055	97	0.020	(−0.177, 0.218)	0.839
Weekly GWG, kg	160	4.104	(1.422, 6.787)	0.003*	116	2.930	(−0.029, 5.889)	0.052	123	4.510	(1.143, 7.877)	0.009	97	5.0890	(2.020, 9.761)	0.003*
Leptin 18 weeks' gestation	130	−0.001	(−0.001, 0.001)	0.030	116	−0.001	(−0.001, −0.001)	0.002*
Leptin 36 weeks' gestation									107	−0.001	(−0.001, 0.001)	0.243	97	−0.001	(−0.001, 0.001)	0.054
Adiponectin 18 weeks' gestation	128	−0.006	(−0.011, −0.001)	0.048	116	−0.004	(−0.010, 0.001)	0.172
Adiponectin 36 weeks' gestation									107	−0.011	(−0.018, −0.004)	0.003*	97	−0.009	(−0.015, −0.002)	0.009
GDM, yes	160	0.219	(−0.923, 1.362)	0.705	116	−0.253	(−1.560, 1.055)	0.702	123	−0.444	(−1.767, 0.894)	0.508	97	−1.318	(−2.796, 0.159)	0.080
HbA1C, <16 weeks' gestation
HbA1C 36 weeks' gestation	142	0.536	(−1.000, 2.071)	0.491	116	1.161	(−0.681, 3.002)	0.214	107	0.361	(−1.415, 2.136)	0.688	97	1.536	(−0.372, 3.444)	0.113

	Birthweight (z‐score) Univariate				Birthweight (z‐score) Multivariate
Exposure	Number of observations	Coefficient	Confidence Interval, 95%	p value	Number of observations	Coefficient	Confidence Interval, 95%	p value
Age, years	163	−0.023	(−0.059, 0.013)	0.207	122	0.007	(−0.037, 0.051)	0.757
Para0, nulliparous	163	−0.372	(−0.726, −0.018)	0.039	122	−0.182	(−0.597, 0.234)	0.389
pBMI	163	−0.004	(−0.056, 0.048)	0.868	122	0.007	(−0.050, 0.065)	0.802
Weekly GWG, kg	163	1.661	(0.805, 2.517)	<0.001*	122	1.481	(0.383, 2.578)	0.009
Leptin 36 weeks' gestation	133	−0.001	(−0.001, 0.001)	0.772	122	−0.001	(−0.001, 0.001)	0.322
Adiponectin 36 weeks' gestation	133	−0.001	(−0.003, 0.001)	0.161	122	−0.001	(−0.003, 0.001)	0.224
GDM, yes	163	0.225	(−0.147, 0.598)	0.234	122	−0.164	(−0.604, 0.277)	0.463
HbA1C 36 weeks' gestation	143	0.623	(0.129, 1.117)	0.014	122	0.668	(0.096, 1.240)	0.022

Note: * = Significant (p value 0.006) after table‐wise Bonferroni correction (e.g., α/number of tests); BPD 2 and 4, biparietal diameter measured around 32 and 39 weeks' gestation; GDM, a dichotomous variable indicating gestational diabetes (1 = no, 0 = yes); MAD 2 and 4, mean abdominal diameter measured around 32 and 39 weeks' gestation; Para0, a dichotomous variable indicating parity (0 = multiparous, 1 = nulliparous); pBMI, pre‐pregnancy body mass index; Weekly GWG, gestational weight gain (calculated from pre‐pregnancy weight and last weight before delivery) divided by gestational age at birth (weeks); Birthweight, z‐scores. ²⁶

FIGURE 1.

Intrauterine biparietal diameter growth is shown from early to late gestation. Sequential prenatal visits are marked by color: red (BPD 0), blue (BPD 1), olive (BPD 2), green (BPD 3), and pink (BPD 4). Percentiles from Johnsen (2006) are purple (10th), yellow (50th), and slate blue (90th).

FIGURE 2.

Intrauterine mean abdominal diameter growth is shown from early to late gestation. Sequential prenatal visits are marked by color: red (MAD 0), blue (MAD 1), olive (MAD 2), green (MAD 3), and pink (MAD 4). Percentiles from Johnsen (2006) are purple (10th), yellow (50th), and slate blue (90th).

First, starting with standardized birthweight (z‐score) as the outcome variable, we found that birthweight was positively associated with weekly GWG and HbA1C and negatively associated with nulliparity. However, in multivariate analyses, birthweight remained associated only with weekly GWG and HbA1C, but after table‐wise Bonferroni correction, none of these associations remained statistically significant.

Secondly, focusing on birthweight categories, we observed that, although the mean birthweight was within the normal range, the proportion of LGA infants in our study population was 18.4%, more than double the 8.6% observed in the Norwegian background population during the same timeframe (unpublished data from the Medical Birth Registry of Norway, received 25 March 2025). The proportion of SGA infants (12.9%) was also higher compared to the background population (9.0%) (unpublished data from the Medical Birth Registry of Norway, received 25 March 2025). Maternal age was positively associated with SGA, while both weekly GWG and HbA1C measured at 36 weeks were negatively associated with SGA in crude analyses. In multivariate linear regression analyses, weekly GWG and HbA1c measured at around 36 weeks' gestation were positively associated with LGA. After table‐wise Bonferroni correction, only the association between LGA and weekly GWG remained statistically significant, as presented in supporting information (Table S1).

Looking into intrauterine fetal growth parameters, we found that both adiponectin and leptin were negatively associated with BPD. Although this association remained statistically significant in the univariate analysis after table‐wise Bonferroni correction, the effect size was very small, and their clinical relevance is assumed to be limited (Table 3 and Table S1). Weekly GWG from around 32 weeks and throughout the pregnancy was positively associated with BPD and MAD; however, the association remained significant only with BPD after table‐wise Bonferroni correction.

HbA1C measured around 32 and 36 weeks was positively associated with MAD2 and MAD3 in univariate analyses, and the association remained significant after table‐wise Bonferroni correction (Table 3 and Table S1).

BMI was positively associated with MAD2 in the multivariable model and negatively associated with MAD4 in the univariate analysis. The fact that the directions of association are opposite, that similar associations were not observed at other time points in pregnancy, and that the findings did not remain significant after Bonferroni correction all suggest that these are chance findings.

4. DISCUSSION

In this observational longitudinal study of 163 pregnant women with pBMI ≥35 kg/m², we found that weekly GWG and HbA1C measured around 36 weeks' gestation were positively associated with birthweight, as well as intrauterine BPD measurements and partly with MAD. There was no consistent or robust pattern of associations between pBMI and any of the adjusted growth outcomes. Despite birthweight being within normal ranges (mean z‐score 0.126), the proportion of especially LGA, and SGA, was high compared to the background population. Moreover, there was a high incidence of gestational diabetes mellitus (GDM), with one in three women diagnosed with GDM, of whom approximately one in three needed medical treatment.

Multiple studies suggest that birthweight is influenced by a wide variety of factors. ⁵ , ⁷ , ¹⁰ , ¹¹ Relevant prior findings include a positive association between birthweight and factors such as maternal pBMI, multiparity, and excessive GWG. However, our results did not support all these findings. Instead, we demonstrated that positive associations between birthweight and multiparity were eliminated in multivariate analyses, while positive associations between birthweight and weekly GWG and HbA1C persisted, though not after Bonferroni correction. The association between HbA1c and birthweight was positive after adjusting for GDM. All GDM cases received glucose‐lowering interventions with medication or with diet and blood‐glucose monitoring, which may have reduced their glucose and HbA1c and thus diminished the differences. A Norwegian study by Carlsen et al. (2022) demonstrated that HbA1C levels were linearly related to infant size at birth in women without diabetes. ²⁸

Many previous studies examining the associations between maternal BMI and neonatal birthweight primarily involved populations of normal weight and low‐risk mothers, lacking high‐risk populations altogether or including only a small proportion of obese and high‐risk mothers. ⁵ , ⁷ , ¹⁰ , ¹¹ , ²⁸ Prior studies ⁵ , ⁷ , ²⁸ have shown an association between pBMI and birthweight in study populations comprising women with normal‐ and/or overweight women (pBMI <35 kg/m²). A possible explanation for the discrepancy between these findings and the lack of an association between pBMI and birthweight in our study may be because all the included women were obese (pBMI >35 kg/m²). The effect of BMI on birthweight may plateau at such high BMI levels, and it is possible that the association between pBMI and birthweight is most clearly observed across a wider pBMI spectrum.

Our findings in multivariate analysis of a positive association with HbA1C, in contrast to randomly measured fasting glucose and birthweight, were not surprising, as HbA1C reflects average glycemia over 2–3 months in non‐pregnant individuals, making it less sensitive to short‐term glucose fluctuations. Hence, fasting glucose levels may provide limited insight into the glycemic status of the pregnant woman compared to measured HbA1C. ²⁸ , ²⁹ However, it should be noted that the assumption that HbA1c reflects glycemia over 2–3 months has not been fully validated in pregnancy, as several pregnancy‐related factors and anemia may influence HbA1c levels. ¹⁶ , ³⁰ A recent meta‐analysis by Mañé et al. reported that first‐trimester HbA1c >5.7% was found to be associated with an increased risk of LGA and macrosomia, suggesting that early elevated HbA1c may identify pregnancies at higher risk. ³⁰ Kiefer et al. showed that in pregnancies with pregestational diabetes, measurement of HbA1C levels between 28 and 37 weeks' gestation significantly improved the prediction of LGA neonates. ³¹ Furthermore, Fonseca et al. found that an increased HbA1C level during the third trimester was associated with LGA neonates in women with GDM. ¹⁶ Our findings suggest that, for women with obesity who present a challenge for ultrasound estimation of fetal weight due to poorer imaging quality, ³² future studies could look into how a combination of a single third trimester HbA1C measurement and the calculated GWG may add clinically relevant information to ultrasound findings. If assessed early in the third trimester, there would remain an opportunity to improve glycemic control, GWG, and encourage lifestyle changes, potentially strengthening prenatal care strategies aimed at optimizing birth outcomes.

In clinical practice, it can be tempting to omit weight measurements in women who already have weight‐related difficulties and are reluctant to be weighed. However, it is important to remember that close attention to gestational weight gain is especially important in this group of women. The complexity of GWG, including the contribution of fluid retention/edema versus true caloric surplus, must be considered when interpreting weight changes. GWG offers a tangible opportunity for risk reduction during pregnancy: although not easily altered in all women, it can be actively managed when mothers receive timely information and support, in contrast to many other exposures that are not easily altered. Our multivariable analyses suggest that weekly GWG and HbA1C have a large estimated effect on adjusted birthweight; each 0.1 kg/week increase in GWG was associated with a 74.8 g higher adjusted birthweight (95% CI 28.3–121.3 g, p = 0.002*), whereas each one‐unit increase in HbA1c % corresponded to a 324.8 g increase (95% CI 82.5–567.1 g, p = 0.009) (Table S1). Fasting glucose provides a transient state of maternal glycemic status and is a well‐established method for assessing the management of GDM. However, its applicability to broader fetal growth assessment is debatable. HbA1c may offer a practical alternative because it would require only testing early and late in pregnancy, a possibility that merits further investigation. Possibly, standardized levels for third‐trimester HbA1c for women without anemia could be an additional tool for the physician to identify an increased risk for LGA infants.

In our study group, the incidence of LGA infants was more than doubled (18.4%) compared to the Norwegian background population (8.6%). In Norway, there has been a decrease in the incidence of LGA infants from 13.2% in 1999–2003 ³³ to 9.9% in 2014–2018 which has been attributed to improvements in prenatal care, including the diagnosis and management of GDM. ³³ According to SuárezIdueta1, ³⁵ LGA rates vary between 8 and 25% between countries, and the diversity in the rates can be explained by factors such as the use of different growth charts and measurement methods, variations in ethnicity, and differences in prenatal healthcare services. ³⁵ An elevated rate is expected in high‐risk groups given the presence of factors such as obesity and GDM. Not surprisingly, we found LGA outcomes to be positively associated with GWG and HbA1C levels in multivariate analyses, as well as with multiparity in univariate analyses.

In multivariate analyses, SGA was not associated with any predictor variables. Still, we found a higher rate of SGA of 12.9% in the study group, versus 9% in the background population. Known risk factors for SGA include hypertensive disorders of pregnancy, poorly controlled chronic diseases, malnutrition, and placental dysfunction. Half of the participants reported having at least one comorbidity, which illustrates the health challenges of pregnant women with obesity. Six percent of the mothers reported smoking during pregnancy; we observed no clear impact on infant weight or growth.

As illustrated in Figure 1 our BPD measurements appeared to be slightly above average, especially between 30 and 38 weeks' gestation, in line with reports from Zhang et al., ³⁶ where head circumference was measured instead of BPD. In addition, Zhang et al. found that infants of women with obesity had significantly longer femur and humerus than fetuses of non‐obese women already from 21 weeks' gestation, while we did not measure femur and humerus lengths. Our crude analyses suggested that BPD, rather than MAD, was negatively associated with female fetal sex. This finding could question the appropriateness of using undifferentiated fetal BPD growth charts for both sexes. The BPD seems to be negatively associated with adiponectin from early pregnancy, which may reflect the role of adiponectin in regulating metabolism and reducing insulin resistance. In general, adiponectin levels are typically decreased in pregnancies with maternal obesity compared to maternal normal‐weight pregnancies, and increased adiponectin is reported to be associated with better metabolic health. ¹⁹ Some studies have found an inverse association between birthweight and adiponectin levels, but the studies lack longitudinal growth measurements. ³⁷ We did not find any substantial association between adiponectin or leptin and fetal growth.

MAD1 was consistently and negatively associated with nulliparity in all analyses, in accordance with prior knowledge. No such association was found later in pregnancy, but we found that LGA was also negatively associated with nulliparity in crude analyses. MAD appears to be primarily influenced by nulliparity during the early stages of pregnancy; however, in the later stages, factors such as nutrition and metabolic conditions seem to take on more significance, diminishing the impact of nulliparity, and these factors may not differ significantly between the sexes.

The strength of our study lies in the longitudinal measurements of fetal growth, along with maternal weight and metabolic factors derived from blood samples drawn throughout pregnancy. We observed a wide diversity in our patients' places of residence, cultural backgrounds, and patients originating from a diverse range of urban and rural referral offices, increasing the geographic and sociodemographic representativeness of the sample. This diversity enhances the generalizability of findings, helps to reduce selection bias, and improves patient care. All patients were followed according to the same national recommendations. Because we did not have a control group of women with BMI <35 kg/m², we used data from the Medical Birth Registry of Norway and the local registry of births from Drammen Hospital for comparison of selected outcomes. Given the differences in both antenatal care programs and the background populations across different countries, our study results are particularly relevant for Nordic countries.

A limitation of our study was that the interobserver variability for the ultrasound measurements was not evaluated, and our limited sample size of 163 participants. There was a change in Norwegian GDM criteria during the inclusion period, which resulted in different diagnostic thresholds for fasting and 2‐h glucose levels for patients included before and after the change. Both criteria yielded a high prevalence of GDM, but more women would have been diagnosed if the new criteria were applied for all participants. Another limitation is the use of two GDM criteria, as fasting hyperglycemia represents a different population of women compared to those with elevated postprandial levels, and these groups may also have different etiologies. HbA1C levels may be impacted by several factors, such as anemia, which can lead to reduced red blood cell lifespan and affected glycation rates, as well as altered red blood cell turnover during pregnancy, which may skew the results, shorten the period it reflects, and potentially misrepresent true glycemic control. ¹⁶ , ³⁰

5. CONCLUSION

In our population of pregnant women with pBMI ≥35 kg/m², we found that weekly GWG and HbA1C measured around GW 36 were positively associated with birthweight (z‐score) and LGA, while fetal growth measured as BPD and MAD was positively associated with weekly GWG alone. These results remained after adjustment for GDM status and other covariates in multiple analyses. After adjusting for multiple comparisons using a table‐wise Bonferroni correction, weekly GWG was the only exposure that remained statistically significantly associated with MAD2, BPD4, and LGA. Results that were not statistically significant after the Bonferroni correction should therefore be interpreted with caution and regarded primarily as hypothesis‐generating. There were no associations between pBMI and any outcome variables.

Taken together, our findings suggest that GWG is a modifiable exposure with potential to influence fetal growth and birthweight. Notably, HbA1c in late pregnancy was associated with fetal birthweight in multivariate analyses adjusted for GDM. The role of HbA1c measured in late pregnancy in monitoring fetal growth merits further investigation.

AUTHOR CONTRIBUTIONS

MLB contributed to data collection, data analysis, and manuscript writing and editing. LTN, EQ, and GH contributed to manuscript writing and editing. WRPD contributed to editing and data analysis. MCPR contributed to protocol/project development, data collection, data analysis, and manuscript writing and editing. All authors approved the final draft.

FUNDING INFORMATION

The study was funded by Vestre Viken Hospital Trust 16C002.

CONFLICT OF INTEREST STATEMENT

None.

ETHICS STATEMENT

The study was approved by the Regional Committee for Medical Research Ethics Southeast Norway on June 30, 2015 (reference number: 16496) and performed according to the ethical principles of the Declaration of Helsinki. All participants signed a written informed consent.

Supporting information

Table S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.