Maternal Metabolic Dysfunction-Associated Steatotic Liver Disease During Pregnancy And Its Impact on Offspring Steatosis

Abstract

Background & Aim:

To investigate the yet unexplored links between maternal and offspring outcomes of pregnancy metabolic dysfunction-associated steatotic liver disease (MASLD).

Methods:

Prospective observational study using magnetic resonance imaging with proton density fat fraction (MRI-PDFF) to determine the presence of maternal and offspring liver steatosis during gestation (3^rd trimester) and in the first year postpartum. Women were deemed at high or low risk for MASLD if they had risk factors (pre-pregnancy obesity or MASLD, gestational diabetes, rapid weight gain during pregnancy). Multivariable regression and longitudinal mixed models assessed associations and changes in MRI-derived fat measures.

Results:

We recruited 57 pregnant women; 47 (82%) were high-risk for MASLD. MASLD was seen in 18% (n=10) of women (and 15% [n=7/47] of those high-risk), with a median PDFF of 5.9% (IQR: 5.6–12.0%). Maternal weight, BMI and HDL-C differed between groups. The fetal liver and placental PDFF of offspring of women with and without MASLD were not different. N=7 (12%) infants had clinically significant steatosis in utero, all offspring of mothers without gestational MASLD. Of the 30 mothers with repeat liver MRI-PDFF at 9 months postpartum, ten (33%)had MASLD. N=44 infants followed up at either 3 or 9 months of age, with only one still demonstrating significant steatosis 9 months. No infant exposed to gestational MASLD had significant steatosis in the first year.

Conclusions:

Gestational MASLD is generally mild, but hepatic fat fraction increases postpartum. Clinically significant hepatic steatosis occurs as early as in utero but is not associated with maternal MASLD.

Background

Maternal health during gestation has an impact on both maternal and offspring outcomes. Obesity is the most common condition affecting pregnant women, with a prevalence ranging from 19–36%, depending on racial/ethnic background(1). From a maternal perspective, obesity has been linked to the development of gestational diabetes (GDM), hypertension, need for C-section, and delivery complications, such as postpartum hemorrhage or infection(2). Obesity also modifies placental function and leads to placental inflammation and oxidative stress, ultimately altering nutrient crossing to the fetus(3, 4). Offspring of women with obesity are more likely to be born prematurely, have low or high birth weight, or even experience fetal demise(1, 5–7). They also have increased future risk of adiposity as well as neurodevelopmental and cardiovascular morbidity(8). Less is known about the impact of maternal metabolic dysfunction-associated steatotic liver disease (MASLD) during pregnancy on maternal and offspring outcomes.

MASLD is tightly linked to obesity and is estimated to affect 10% of women of childbearing age(9, 10) and 40% of pregnant women with obesity(11). The prevalence of gestational MASLD has tripled over the past decade and it is now the most common cause of cirrhosis during pregnancy(11, 12). Maternal MASLD is associated with GDM, pre-eclampsia, premature birth, as well as delivery of large-for-gestational-age infants(10, 13). The direct impact of gestational MASLD on maternal and offspring outcomes is less clear. For example, it is not known whether gestational MASLD is a predictor of future liver disease progression in these women. Furthermore, while maternal obesity has been associated with increased intrahepatic fat in neonates(14, 15), it remains to be determined whether there is a direct link between maternal and offspring MASLD development.

An unhealthy gestational milieu is as a risk factor for pediatric MASLD. Specifically, gestational obesity and diabetes have been associated with MASLD in adolescence(16). In addition, pediatric MASLD severity has been linked to high and low birth weights(17). While these data suggest a connection between gestational exposures and pediatric MASLD, it is not known whether pregnancy risk factors are associated with MASLD development early in life. In fact, to date, the age at which pediatric MASLD develops remains to be determined. As such, the optimal timing and approach to disease prevention cannot be ascertained. This is a significant limitation given the staggering prevalence of pediatric MASLD and the current lack of available pharmacotherapies (18).

Considering the limitations of the available literature, the objective of this study was to investigate the links between maternal MASLD and offspring hepatic steatosis, as well as the longitudinal outcomes of both women and offspring affected by gestational MASLD. We hypothesized that women with MASLD during pregnancy were more likely to have persistent steatosis postpartum and to carry offspring with higher intrahepatic fat in utero, and during the first year of life, indicating signs of early MASLD development.

Methods

This was a prospective, observational study performed at Cincinnati Children’s Hospital Medical Center (CCHMC), University of Cincinnati, and The Christ Hospital. Participants were enrolled between September 2020 and March 2023. Eligible women were approached for recruitment during their regular clinic visits with their obstetrician and via marketing emails. The Cincinnati Children’s Hospital Medical Center, University of Cincinnati and The Christ Hospital Institutional Review Boards had approved the study prior to the initiation of any study-related procedures.

Pregnant women were eligible for enrollment if they were 18 years of age or older, and with a gestational age of at least 28 weeks. Exclusion criteria were inability to undergo magnetic resonance imaging (MRI), as well as multiple gestation, active health issues during pregnancy requiring hospitalization or surgery (other than the cardiometabolic complications studied herein), known genetic conditions that predispose to hepatic steatosis (e.g., fatty acid oxidation disorder, lipodystrophy), use of drugs of abuse or alcohol at the time of enrollment, concurrent use of medications contributing to steatosis (e.g., corticosteroids), inability to provide informed consent, or fetus with known congenital anomalies.

Because gestational MASLD can only be detected through imaging after recruitment, we designed our recruitment strategy to enrich the study cohort with women with MASLD, which was the gestational exposure of interest. The study thus initially focused on identification and recruitment of high-risk women, such as those with pre-pregnancy obesity (defined as body mass index [BMI] >30 kg/m²), GDM, excessive weight gain during pregnancy, as defined by the Institute of Medicine criteria (19), or a pre-pregnancy diagnosis of MASLD. The study subsequently started enrolling participants without these risk factors, who were considered to be at no increased risk for MASLD. While a 5% hepatic steatosis threshold was used to determine MASLD in the mothers, per the accepted criteria for adults(20), a 3% hepatic steatosis threshold was used in the offspring. Given the paucity of data regarding the optimal steatosis threshold at that age, we chose to use a cutoff previously shown to correlate better with histopathology(21).

Baseline clinical and demographic information were collected from the participants and the electronic medical record at enrollment. Demographic data included age, sex, race and ethnicity. Clinical information included gestational age at enrollment, maternal medical history, medications, and pregnancy history. Information regarding the paternal medical history was also obtained from the participants.

Enrolled patients completed a research visit at Cincinnati Children’s Hospital Medical Center at or beyond the 32^nd week of gestation. This time point was chosen as the fetal size at that gestational age allowed for more accurate imaging of the fetal liver. Additionally, most of the offspring adipose tissue deposition occurs in the 3^rd trimester of pregnancy (22); therefore, imaging at that time allowed us to measure subcutaneous fat thickness as well (see below). During that visit, anthropometrics (weight, height, and BMI), laboratory studies, and an abdominal MRI were performed. Gestational age at the time of imaging was determined based on last menstrual period data. A fasting MRI-proton density fat fraction (PDFF) was obtained using a 3 Tesla research MRI scanner. Separate image acquisitions were used to image and quantify the fat content of the maternal and fetal liver, as well as the placenta. Maternal, fetal and placental PDFF measurements were manually obtained based on four measurements of each tissue by a single observer (board certified radiologist, AT). PDFF values were expressed as an area-weighted mean for analysis (23–26). Fetal abdominal wall subcutaneous fat thickness was obtained via a single measurement over the fetal liver. Fasting blood laboratory tests were obtained to measure high- and low-density lipoprotein-cholesterol (HDL-C, LDL-C), total cholesterol, triglycerides, alanine aminotransferase (ALT), and aspartate aminotransferase (AST). Three 24-hour dietary recalls were performed at and after the study visit. Trained interviewers employed the United States Department of Agriculture automated multiple-pass method, a standardized interview approach(27).

Following delivery, mothers were invited for a second research visit at 9 months postpartum and their offspring had two research visits, at 3 and 9 months of age. Maternal data collected at the 9-month postpartum visit were the same as in their first research visit described above. The following information was collected at birth from the offspring: birth history (gestational age at birth, mode of delivery), offspring demographics (sex, race, ethnicity) and anthropometrics (birth weight, length, head circumference, and their respective z-scores per the World Health Organization criteria). At the 3- and 9-month visits, offspring underwent abdominal MRI-PDFF as above, had their anthropometrics measured, and dietary history collected.

Statistical Analyses

Descriptive statistics were used to determine the mean and SD or median and interquartile range (IQR) of continuous variables, as indicated by distribution. Categorical variables were reported as absolute counts and percentages. Presence of maternal MASLD determined using MRI-PDFF was defined as hepatic fat fraction >5%. The primary outcome was fetal PDFF as a continuous variable. Placental PDFF and fetal abdominal wall subcutaneous fat thickness were also examined as continuous variables. Comparisons between maternal MASLD groups were made using Student’s t-test for normally distributed and Mann-Whitney U or Kruskal-Wallis rank test for non-normally distributed continuous data and Fisher’s exact test or χ² test for categorical variables. Spearman correlation coefficient was used to assess the relationship among MRI measurements. Multivariable generalized linear regression models were used to evaluate risk factors for maternal PDFF as a continuous variable, as well as maternal MASLD relationships with fetal hepatic PDFF, placenta PDFF, and fetal subcutaneous fat thickness, adjusting for key covariates. Covariates of interest included maternal age, race, gravida and parity history, laboratory measurements, gestational age and BMI at the time of MRI, enrollment risk status (high risk vs. low risk) and presence of each risk factor condition, where feasible (e.g., pre-pregnancy obesity, GDM, excess weight gain, pre-pregnancy MASLD). Factors were retained in the model if p <0.05 or inclusion of the variable improved model fit (change in Bayesian Information Criterion >4) with highly collinear variables (e.g., weight and BMI) tested in separate models.

Longitudinal outcomes of women and infants exposed/not exposed to pregnancy MASLD were evaluated using either longitudinal mixed models (for infants) or linear regression models of the differences in PDFF between 9 months postpartum and pregnancy (for mothers). The primary outcome was change in maternal or fetal hepatic PDFF, adjusting for key covariates such as enrollment risk status, demographic characteristics, baseline PDFF (for maternal models), anthropometry changes, dietary intake, and laboratory values. As with the cross-sectional analysis, factors were included in the model if p <0.05 or inclusion of the covariate improved model fit, with highly collinear variables (e.g., weight and BMI) tested in separate models.

Results

A total of 698 pregnant women were assessed for eligibility between September 2020, and March 2023. Of those, 619 were found to be ineligible for participation on pre-screening or declined to participate. Seventy-nine participants were enrolled, and 22 were withdrawn prior to study completion, leaving 57 participants who completed the baseline study (Figure 1). A total of 44 infants were reassessed at 3 and/or 9 months of age, while 30 mothers were re-evaluated at 9 months postpartum.

Figure 1:

Participant flow diagram

The baseline demographic and clinical characteristics of the mothers that participated in this study are shown in Table 1. Most were White and non-Hispanic with a median age of 31 years (IQR 28–34). Their past medical history included hypertension (25%), polycystic ovarian syndrome (12%), obstructive sleep apnea (5%) and dyslipidemia (4%). The majority were at high risk for MASLD based on our pre-determined criteria (47 of 57; 82%), which was expected given the enrollment priority of these groups. The majority had pre-pregnancy obesity (56%) and 28% had been diagnosed with GDM in this pregnancy. Paternal risk factors, such as MASLD, obesity and diabetes were seen in <10% of pregnancies (data not shown).

Table 1.

Demographics and clinical characteristics of mothers at baseline

Characteristic	No Gestational MASLD (n = 47)	Gestational MASLD (n = 10)	p-value
Maternal age at MRI, median (IQR)	31 (28, 35)	28.5 (28, 31)	0.11
Race, n (%) Black	9 (19%)	1 (10%)	1.00
Hispanic ethnicity, n (%)	1 (2%)	1 (10%)	0.32
Number of pregnancies (Gravida)	2 (1, 5)	3 (1, 5)	0.80
Number of prior births (Parity)	1 (0, 2)	1 (0, 1)	1.00
Pre-selected MASLD risk factors, n (%) *
Pre-pregnancy diagnosis of MASLD	3 (6%)	4 (40%)	0.01
Pre-pregnancy BMI >30 kg/m2	27 (57%)	5 (50%)	0.73
Pre-pregnancy T2D	6 (13%)	0 (0%)	0.58
Excessive weight gain 1st trimester	9 (19%)	1 (10%)	0.48
Excessive weight gain 2nd trimester	15 (32%)	1 (10%)	0.27
Excessive weight gain before MRI	19 (40%)	2 (20%)	0.39
GDM in current pregnancy	13 (28%)	3 (30%)	1.00
Risk Status based on pre-selected MASLD risk factors
Low Risk (no risk factors)	7 (15%)	3 (30%)	0.36
High Risk (any risk factor)	40 (85%)	7 (70%)
GA at MRI in weeks, median (IQR)	34 (32, 35)	33 (32, 35)	0.69
Maternal anthropometrics at MRI, median (IQR)
Weight (kg)	95 (79, 113)	98 (80, 123)	0.52
Height (cm)	165 (160, 170)	167 (164, 169)	0.34
BMI (kg/m2)	34.2 (30.2, 41.4)	35.8 (29.6, 45.4)	0.59
Laboratory measurements at MRI, median (IQR) †
ALT (U/L)	13 (10, 21)	15.5 (11, 24)	0.41
AST (U/L)	20 (16, 22)	20.5 (16, 23)	1.00
Triglycerides (mg/dL)	185 (152, 238)	221 (196, 253)	0.09
Total cholesterol (mg/dL)	217 (193, 256)	229 (216, 267)	0.61
LDL-C (mg/dL)	115 (101, 145)	149 (114, 193)	0.06
HDL-C (mg/dL)	64 (56, 77)	54 (52, 58)	0.02
Direct bilirubin	0.1 (0.1, 0.1)	0.1 (0.1, 0.1)	0.23
Total bilirubin	0.3 (0.2, 0.4)	0.3 (0.3, 0.3)	0.88
Albumin (g/dL)	2.8 (2.7, 2.9)	2.8 (2.6, 3.0)	0.79
Maternal dietary intake, median (IQR) ††
Energy, total kcal/day	2186 (1490, 2499)	2088 (1802, 2465)	0.68
Total protein, g/day	83.1 (69.3, 102)	88.0 (70.9, 103)	0.62
Total fat, g/day	93.5 (70.2, 114.5)	86.1 (71.7, 111.5)	0.96
Omega-3 fatty acid, g/day	1.9 (1.3, 2.4)	2.4 (1.5, 2.4)	0.52
Omega-6 fatty acid, g/day	17.1 (10.9, 24.1)	16.4 (12.6, 20.6)	0.73
Total carbohydrate, g/day	235 (164, 299)	217 (187, 375)	0.59
Total sugar, g/day	85.5 (48.7, 128.1)	108.6 (59.7, 183.5)	0.26
Added sugar, g/day	45.4 (20.0, 84.3)	75.5 (26.3, 103.1)	0.34
HEI total score (range 0–100)	47.7 (42.6, 58.1)	44.6 (40.8, 54.4)	0.40

^*Number (%) in risk factor groups do not sum to 100% as women may have more than one risk factor.

Abbreviations: ALT: alanine aminotransferase; AST: aspartate aminotransferase; BMI: body mass index; GA: gestational age; GDM: gestational diabetes mellitus. HDL-C: high density lipoprotein cholesterol; HEI: healthy eating index; IQR: interquartile range; LDL-C: low density lipoprotein cholesterol; MASLD: metabolic dysfunction associated steatotic liver disease; MRI: magnetic resonance imaging; T2DM: type 2 diabetes mellitus

Data presented as median (IQR). P-values comparing no maternal MASLD with maternal MASLD by Fisher’s exact test or Kruskal-Wallis test for categorical and continuous variables, respectively.

^†One participant without MASLD did not have blood work obtained.

^††Diet intake data is average daily intake over three days. Diet data were unavailable for one subject in each group.

At an approximate 34 weeks gestation, gestational MASLD was identified in 10 (18%) women, with prevalence not differing by pre-determined risk status (15% [7/47] among high risk, and 30% [3/10] among low risk, p=0.36, Table 1). Those with MASLD had a median liver PDFF of 5.9% (IQR: 5.6–12.0%), while those without MASLD had a median liver PDFF of 2.7% (IQR: 1.8–3.3%). Of the laboratory tests conducted, only HDL-C differed among those with versus without gestational MASLD (54 mg/dl [52, 58] versus 64 mg/dl [56, 77], respectively, p=0.02), while LDL-C trended higher in those with MASLD (149 mg/dl [114, 193] versus 115 mg/dL [101, 145], p=0.06). Overall, maternal PDFF was positively correlated with concurrent weight and BMI (ρ=0.26, p=0.046 and ρ=0.28, p=0.03, respectively), and marginally negatively associated with maternal age (ρ=−0.26, p=0.055) and HDL-C (ρ=−0.26, p=0.051). No other demographic, baseline clinical, anthropometric, dietary, or metabolic laboratory measures were associated with maternal PDFF.

The fetal liver PDFF of offspring of women with and without MASLD was not significantly different (1.16% [0.70–1.78%] vs. 1.85% [1.09–2.55%], respectively; p=0.11). Using a threshold of ≥3% to define MASLD in the offspring, 7 (12%) infants were classified as having MASLD in utero, all offspring of mothers without gestational MASLD. Overall, in the entire cohort, fetal PDFF was positively associated with maternal total cholesterol (ρ=0.28, p=0.03) and maternal AST (ρ=0.34, p=0.01) and marginally negatively associated with maternal weight at MRI (ρ=−0.24, p=0.07). Fetal PDFF also marginally negatively correlated with maternal caloric and total fat intake (ρ=−0.26, p=0.054 and ρ=−0.25, p=0.06, respectively), and marginally positively correlated with maternal total Healthy Eating Index score (ρ=0.24, p=0.08). Other clinical and dietary factors were not correlated with fetal PDFF.

No differences were noted in the placental fat fraction between offspring with (1.45% [0.95–1.93%]) versus without MASLD (1.38% [0.83–2.05%], p=0.83) or in the fetal abdominal subcutaneous fat measurements between the groups (3.25% [3.00–3.80] vs. 3.50% [3.30–3.90], respectively; p=0.28). Overall, placental fat fraction was positively correlated with higher maternal parity (ρ=0.26, p=0.049). Fetal abdominal subcutaneous fat was positively correlated with maternal BMI (ρ=0.34, p=0.01). No other variables of interest were associated with the MRI measurements during gestation. In addition, maternal PDFF, fetal PDFF, placental fat fraction and fetal subcutaneous fat measurements were not significantly correlated with each other.

At birth, the infant sex and gestational age were not different between women with and without MASLD (Table 2). No differences were seen in the delivery method between the groups. Offspring of women with gestational MASLD had similar birth weight z-scores but higher length z-scores than their counterparts born to women without MASLD.

Table 2:

Offspring characteristics at birth, by maternal MASLD status

Characteristic	No Gestational MASLD (n = 47)	Gestational MASLD (n = 10)	P-value
Infant sex, male (n, %)	23 (49%)	3 (30%)	0.32
GA at birth, weeks	38 (37, 39)	38 (37, 39)	0.63
Delivery method, vaginal (n, %)	26 (55%)	7 (70%)	0.49
Weight, g	3186 (2900, 3481)	3242 (3105, 3480)	0.39
Weight z score	−0.31 (−0.75, 0.36)	0.01 (−0.28, 0.46)	0.33
SGA	4 (9%)	0 (0%)	1.00
AGA	40 (85%)	10 (100%)
LGA	3 (6%)	0 (0%)
Length, cm	50.0 (48.3, 50.8)	50.8 (50.2, 50.8)	0.14
Length z score	0.46 (−0.84, 0.85)	0.73 (0.55, 0.89)	0.03
Head circumference, cm	34.0 (32.5, 35)	33.0 (33.0, 34.5)	0.82
Head circumference z score	−0.32 (−1.16, 0.52)	−0.74 (−0.74, 0.52)	0.97

AGA: appropriate for gestational age; GA: Gestational age; LGA: large for gestational age; MASLD: metabolic dysfunction associated steatotic liver disease; SGA: small for gestational age.

Data presented as median (IQR) or n(%). P-values from Kruskal-Wallis test or Fisher’s Exact test, as appropriate.

Longitudinal changes in maternal PDFF

At the 9-month postpartum visit, 30 mothers had repeat MRI scans for hepatic PDFF, with a median PDFF of 4.5% [IQR: 2.6%, 8.8%] (Figure 2). Ten of these women had MASLD (33%). There were no clinically significant differences at baseline between mothers who were or were not followed to 9 months, except a later GA at MRI (p=0.0004) and higher placental PDFF (p=0.006; Supplemental Table 1). Women with gestational MASLD had lower HDL-C and albumin, as well as lower intake of both omega-3 and omega-6 fatty acids, at 9 months postpartum (Table 3). Overall, higher maternal PDFF at 9 months was significantly correlated with higher BMI at both 9 months postpartum (ρ=0.62, p=0.0002) and during gestation (ρ=0.64, p=0.0002), and lower HDL-C (ρ=−0.41, p=0.03), lower albumin (ρ=−0.47, p=0.008) and higher ALT (ρ=0.37, p=0.049) at 9 months postpartum. Remaining maternal dietary and other clinical characteristics at 9 months were not significantly associated with maternal 9-month PDFF.

Figure 2: Change in maternal hepatic fat fraction over time.

A) Women with gestational MASLD; B) Women without gestational MASLD. The red line indicates the 5% PDFF level designating maternal MASLD. Black symbols and lines in both panels represent women whose MASLD status changed between pregnancy and 9 months postpartum; Gray symbols and lines in both panels represent women who either had or did not have MASLD at both visits.

Table 3.

Clinical characteristics at 9 months postpartum in women with or without gestational MASLD.

Characteristic	No Gestational MASLD (n = 24)	Gestational MASLD (n = 6)	P-value
Months since delivery, median (IQR)	9.5 (9.2, 9.8)	9.5 (9.2, 11.4)	0.57
Maternal anthropometrics, median (IQR)
BMI (kg/m2) at 9 months	31.9 (25.4, 36.8)	31.8 (26.7, 36.4)	0.98
Change in BMI from pregnancy to 9 months postpartum	−1.6 (−3.2, 0.5)	−2.5 (−3.1, −1.2)	0.42
Laboratory measurements, median (IQR)
ALT (U/L)	17 (14, 26)	14.5 (11, 19)	0.33
AST (U/L)	24.5 (22, 28.5)	20.5 (19, 24)	0.24
Triglycerides (mg/dL)	89 (68, 110)	85 (55, 157)	0.96
Total cholesterol (mg/dL)	176 (152, 197)	166 (162, 167)	0.16
LDL-C (mg/dL)	107 (90, 117)	104 (95, 114)	0.65
HDL-C (mg/dL)	53 (43, 58)	39 (35, 44)	0.04
Albumin (g/dL)	4.1 (3.9, 4.2)	3.8 (3.7, 4.0)	0.047
Maternal dietary intake, median (IQR) †
Energy, total kcal/day	2124 (1687, 2591)	1841 (1689, 2021)	0.25
Total protein, g/day	87.2 (66.8, 99.4)	88.8 (61.2, 90.0)	0.56
Total fat, g/day	88.5 (77.9, 115.0)	76.4 (55.1, 83.1)	0.18
Omega-3 fatty acid, g/day	2.3 (1.4, 2.9)	0.9 (0.8, 1.4)	0.007
Omega-6 fatty acid, g/day	18.8 (13.2, 23.7)	8.1 (7.1, 8.5)	0.008
Total carbohydrate, g/day	243 (186, 301)	195 (172, 239)	0.25
Total sugar, g/day	86.7 (61.3, 128.5)	79.9 (74.6, 94.8)	1.00
Added sugar, g/day	61.9 (26.4, 98.4)	55.5 (36.8, 79.9)	0.77
HEI total score (range 0–100)	48.6 (39.2, 61.3)	51.6 (51.4, 56.4)	0.69
PDFF, median (IQR)
PDFF at 9 months (%)	4.2 (2.5, 5.2)	9.5 (4,8, 11.5)	0.09
Change in PDFF from pregnancy to 9 months postpartum (absolute %)	1.3 (0.6, 3.5)	1.6 (−1.4, 6.3)	0.76
MASLD at 9 months (%)
Yes	6 (25%)	4 (66%)	0.16
No	18 (75%)	2 (33%)

ALT: alanine aminotransferase; AST: aspartate aminotransferase; LDL-C: low density lipoprotein cholesterol; HDL-C: high density lipoprotein cholesterol; HEI: healthy eating index.

Data presented as median (IQR) or n(%). P-values from Kruskal-Wallis test or McNemar test of symmetry for MASLD status at 9 months

^†Diet intake data is average daily intake over three days. Diet data were unavailable for one subject in the gestational MASLD group.

Between gestation and 9 months postpartum, maternal PDFF increased overall (1.3% [0.1%, 3.6%], p=0.005, Figure 1) even while BMI decreased significantly (−1.95 [−3.10, −0.20] kg/m², p=0.007). The increase in maternal PDFF was similar between women who had gestational MASLD and those who did not (1.6% [−1.4%, 6.3%] and 1.3% [0.6%, 3.5%], respectively, p=0.76). Overall, the change in maternal PDFF was significantly positively correlated with both concurrent BMI (ρ=0.56, p=0.001), BMI during pregnancy (ρ=0.55, p=0.002), and the change in BMI between pregnancy and 9 months postpartum (ρ=0.41, p=0.02). Change in maternal PDFF was also correlated with maternal ALT at 9 months postpartum (ρ=0.54, p=0.003), but none of the other laboratory assays or any aspects of maternal diet.

Longitudinal changes in infant PDFF

Of this birth cohort, 44 infants were followed up at either 3 or 9 months of age. Infants who were followed up had a larger head circumference z-score (p=0.047) and lower placental PDFF (p=0.05) but otherwise did not have clinically significant differences from those who did not (Supplemental Table 2). Overall, the median offspring liver PDFF decreased between in utero and 9-month of age (Table 4, Figure 3). Of the 7 infants classified as having MASLD in utero, one had a liver PDFF >3% at 9 months of age (Figure 3). None of the infants exposed to gestational MASLD had significant hepatic steatosis at 3 or 9 months of age, but those exposed to gestational MASLD had an earlier introduction to solid foods (p=0.02). Longitudinal mixed models of infant PDFF across fetal, 3-month and 9-month time points did not reveal any significant predictors, however, including exposure to maternal MASLD or any anthropometric or dietary factors.

Table 4.

Clinical characteristics at 3 and 9 months of age in infants with or without exposure to gestational MASLD.

Clinical Characteristic	No Gestational MASLD	Gestational MASLD	P-value
N at 3 months †	32	9
N at 9 months †	24	5
Infant anthropometrics, median (IQR)
BMI (kg/m2) at 3 months	16.4 (15.6, 16.9)	16.8 (14.6, 17.1)	0.91
BMI z-score at 3 months	−0.37 (−1.14, 0.08)	−0.21 (−1.38, 0.16)	0.94
BMI (kg/m2) at 9 months	17.7 (17.1, 18.4)	17.4 (15.6, 18.5)	0.56
BMI z-score at 9 months	0.66 (0.21, 1.06)	0.42 (−0.75, 0.99)	0.64
Accelerated weight gain (change in weight z-score>0.67)
Birth to 3 months, n(%)	5/32 (16%)	2/9 (22%)	0.64
3 to 9 months, n(%)	10/21 (48%)	2/5 (40%)	1.00
Birth to 9 months, n(%)	8/24 (33%)	2/5 (40%)	1.00
High risk for obesity (BMI>85th percentile)
At 3 months, n(%)	2/31 (6%)	1/9 (11%)	0.55
At 9 months, n (%)	7/24 (29%)	1/5 (20%)	1.00
Breastfeeding, n (%)
Never	6 (13%)	1 (10%)	1.00
Ever	40 (87%)	9 (90%)
<3 months	21 (46%)	3 (30%)	0.41
3 to 9 months	10 (22%)	2 (20%)
>9 months	9 (20%)	4 (40%)
Age at Solid Food Introduction, months, median (IQR)	6.0 (3.0, 6.0)	4.0 (1.0, 4.0)	0.02
Infant 9-month dietary intake, median (IQR) ††
Energy, total kcal/day	777 (571, 1027)	724 (703, 926)	0.73
Kcal/kg/day	86 (62, 109)	94 (93, 104)	0.36
Total fat, g/day	39.2 (30.0, 45.0)	30.8 (24.2, 34.8)	0.20
Fat (g)/kg/day	4.2 (3.3, 5.1)	4.0 (3.1, 5.2)	0.77
% Kcal from total fat	44.1 (40.4, 47.6)	38.6 (29.8, 43.4)	0.15
Total protein, g/day	15.7 (11.7, 25.3)	17.0 (12.8, 18.6)	0.98
Protein (g)/kg/day	1.8 (1.3. 2.6)	1.7 (1.7, 2.4)	0.49
% Kcal from total protein	8.5 (7.5, 9.6)	8.1 (6.9, 10.1)	0.89
Total carbohydrate, g/day	98 (71, 121)	112 (91, 183)	0.39
Carbohydrate (g)/kg/day	10.6 (6.7, 13.5)	14.4 (13.5, 18.7)	0.09
% Kcal from total carbohydrate	46.4 (43.4, 50.3)	53.3 (49.5, 59.9)	0.11
Omega-3 fatty acid, g/day	0.73 (0.38, 0.98)	0.47 (0.36, 0.48)	0.53
Omega-3 (g)/kg/day	0.08 (0.04, 0.11)	0.05 (0.05, 0.06)	0.60
Omega-6 fatty acid, g/day	6.1 (3.0, 8.2)	3.4 (2.8, 3.5)	0.42
Omega-6 (g)/kg/day	0.66 (0.35, 1.00)	0.41 (0.35, 0.45)	0.60
Total sugar, g/day	59.2 (46.0, 87.1)	52.6 (47.2, 70.2)	0.42
Total sugar (g)/kg/day	6.2 (5.2, 9.4)	6.8 (5.1, 7.1)	0.60
Added sugar by total sugar, g/day	23.2 (2.8, 46.0)	2.8 (1.3, 11.8)	0.42
Added sugar (g)/kg/day	2.7 (0.3, 4.7)	0.4 (0.1, 1.5)	0.45
HEI total score (range 0–100)	37.8 (33.7, 45.4)	43.7 (39.6, 53.3)	0.14
PDFF, median (IQR)
PDFF in utero (%)	1.9 (1.1, 2.5)	1.2 (0.7, 1.8)	0.11
PDFF at 3 months (%)	1.4 (1.0, 2.0)	0.7 (0.7, 1.4)	0.10
PDFF at 9 months (%)	0.6 (0.4, 1.0)	0.7 (0.6, 1.2)	0.31

HEI: healthy eating index.

Data presented as median (IQR) or n(%). P-values from Kruskal-Wallis test or Fisher’s Exact test, as appropriate

^†All 57 infants had PDFF in utero; 41 had PDFF at 3 months; 29 had PDFF at 9 months. N=26 had PDFF at all 3 time points.

^††Diet intake data is average daily intake over three days. Diet data were unavailable for one infant in the gestational MASLD group.

Figure 3: Change in offspring hepatic fat fraction over time.

The red line indicates the 3% PDFF level designating fetal/infant MASLD. Black triangles with gray dashed lines represent infants of mothers without gestational MASLD. Open circles with solid black lines represent infants whose mothers had gestational MASLD. Seven infants exceeded the infant MASLD level in utero, of which one retained MASLD at 3 months postpartum; in addition, one infant developed MASLD by 9 months postpartum.

Discussion

In this cohort of pregnant women, we used MRI in the third trimester of pregnancy to quantify and link maternal and offspring hepatic steatosis. First, we found that MASLD is not highly prevalent or severe during pregnancy, even among women with significant cardiometabolic risk factors. In fact, only 18% of study participants had MASLD and, among those, the median liver PDFF was just above the 5% threshold. Second, maternal risk factors, such pre-pregnancy obesity, GDM and dietary intake, were not different between those with and without MASLD. Thirdly, despite the previously reported links between maternal MASLD and offspring steatosis later in childhood, the in utero hepatic steatosis severity was not different between the offspring of women with and without MASLD. However, 12% of offspring were found to have clinically significant steatosis in utero. The birth characteristics, including offspring anthropometrics and mode of delivery, were not different between women with and without gestational MASLD. At follow-up, 30% of imaged women had MASLD, with the maternal liver PDFF increasing across the board. In contrast, the liver PDFF of the offspring decreased throughout the first year of life. One child (1/7; 14%) of those with significant steatosis in utero still had significant hepatic steatosis at 9 months of life.

MASLD is thought to affect approximately 15% of pregnant women, which is similar to the findings of our study (28). However, our study cohort was enriched in women at risk for MASLD (47 of the 57 participants met predetermined high-risk criteria) and despite that, only 18% of the total cohort, and 15% of those at risk, had MASLD. Interestingly, a much higher proportion of women (30%) had MASLD at 9 months postpartum, further suggesting that changes in pregnancy physiology may protect even high-risk women against MASLD during gestation. One of these changes may be lipid handling during pregnancy.

Insulin resistance, accompanied by hypertriglyceridemia, is at the core of MASLD pathogenesis (29, 30). Pregnancy is characterized by increased insulin resistance, which is amplified by maternal obesity. Triglyceride levels of pregnant women with obesity are 40–50% higher compared to their leaner counterparts(31). This increased circulating lipid load could worsen maternal steatosis. Conversely, estrogen levels are also higher in pregnancy, leading to increased very low-density lipoprotein synthesis(32). Very low-density lipoprotein synthesis and export from the maternal liver could attenuate hepatic steatosis and, in part, balance the negative effects of insulin resistance. Furthermore, in pregnancy, maternal lipids are used to nourish the fetus, ultimately decreasing lipid return to the maternal liver(4). The exact changes in maternal lipid handling of women with MASLD during gestation remain to be investigated. Additionally, given the alterations in hepatic physiology during pregnancy, further studies are needed to establish pregnancy-specific thresholds for clinically significant hepatic steatosis. In the present study, we applied the 5% PDFF threshold commonly used in the adult MASLD literature(20).

The optimal threshold to determine hepatic steatosis in utero remains to be determined and validated. However, an increasing number of studies show that fat fraction thresholds of 2–3% may better differentiate those with and without impaired metabolism and true histologic evidence of clinically significant steatosis(21, 33). Furthermore, from an embryological standpoint, hepatic steatosis is anticipated to emerge in the later stages of gestation given the increase in maternal adipose tissue lipolysis (34), raising the possibility that the standard 5% PDFF threshold used in pediatric and adult populations may overestimate clinically relevant steatosis in the fetal liver. Therefore, applying a 3% threshold, we found that 12% of offspring in this cohort had clinically significant steatosis in utero. In contrast to our hypotheses, this was seen only in offspring of women who did not have MASLD during pregnancy. The latter may be secondary to differences in placental function or in the circulating lipidome of women with MASLD, which could ultimately lead to altered lipid crossing to the offspring(34, 35). For example, we found that the omega-3 intake of women without MASLD was higher. Interestingly, these lipids are known to cross the placenta preferentially and therefore could be predisposing the offspring to hepatic steatosis in utero(36). Other confounding factors such as paternal genetics or unmeasured dietary factors may have contributed to our results. Further, clinically significant offspring steatosis in utero was associated with maternal weight but inversely associated with maternal calorie and fat intake. As these two variables (weight and calorie intake) are typically aligned, it is possible that the discrepancy has to do with the known limitations of dietary recalls(37). Importantly, maternal BMI was also associated with fetal subcutaneous fat measurements in our study, suggesting another avenue through which maternal obesity may predispose the offspring to MASLD later in life. These findings underscore the importance of maintaining a healthy maternal weight in the peripartum period.

The placenta exerts protective effects to the fetus, and from a metabolic perspective determines the type and amount of nutrients to be crossed to the fetal circulation. Obesity is characterized by increased transplacental lipid transport, with accompanying increases in lipid esterification (leading to fat storage within the placenta) and peroxisomal lipid oxidation (eliminating some of the excess fat)(4). These are thought to be physiologic responses to the increased placental lipid delivery, ultimately aimed at protecting the fetus. In our study, we did not see a difference in the severity of placental steatosis between women with and without MASLD. This finding, if validated in future larger studies, may suggest that the degree of lipid esterification in the placenta is not altered in women with MASLD, beyond what is seen in those with obesity alone. To confirm this notion, possible alterations in placental fatty acid oxidation and esterification in the context of maternal MASLD should be investigated. It is also possible that, in the setting of maternal obesity, the hepatic fat accumulation associated with MASLD represents only a small fraction of the overall excess adiposity. As such, the presence of MASLD alone may not significantly exacerbate the degree of placental steatosis when comparing cohorts with excess adiposity.

In contrast to maternal liver PDFF, the infants had an overall drop in their hepatic fat measurements over their first 9 months of life. Of the 7 infants who had clinically significant hepatic steatosis in utero, only one (14%) had persistence of hepatic steatosis at 9 months of age. Additionally, no offspring with normal hepatic fat fraction in utero developed steatosis the first months of life. While the numbers are small, this study shows that significant steatosis does occur early in life and can persist. Careful longitudinal assessments of early-onset steatosis are needed to determine natural history and drivers of this process.

The strengths of this study include the use of MRI-PDFF for the concurrent detection and quantification of maternal and offspring hepatic steatosis, as well as placental steatosis. Most of the literature on maternal MASLD during pregnancy has relied on ultrasonography or ICD-9/10 coding data(11, 28). Furthermore, beyond autopsy studies, there are no previous reports on offspring hepatic steatosis in utero. As a growing number of studies suggest that pediatric MASLD can occur in preschool age children (38, 39), it is necessary to determine the disease onset, to be able to design effective preventive strategies.

Limitations of the study include the relatively small sample size of women with MASLD that may have introduced type 2 error. Specifically, the small number of women with MASLD during pregnancy does not provide adequate power to detect associations between maternal and offspring outcomes and caution should be exerted in interpreting negative findings (e.g. lack of differences in placental fat fraction of women with and without MASLD). Additionally, as these women sought medical care once pregnant, pre-gestational clinical, laboratory and imaging data that could assist us in determining the presence of preexisting MASLD were not available. Furthermore, the study was not designed to study placental physiology in the context of maternal MASLD, which could have furthered our understanding in the pathophysiology of this condition during gestation. However, placental imaging did provide novel preliminary information regarding placental lipid handling in the context of MASLD. Dietary recalls were used to determine the impact of diet on key outcomes of interest in our study. These data can be affected by reporting bias, particularly postpartum. Therefore, future studies should consider including objective biomarkers, such as erythrocyte fatty acid profiles to confirm differences in essential fatty acid intakes noted in our study, to complement dietary investigations. Lastly, our cohort was reflective of the population followed in our region and was enriched in White, non-Hispanic women. As such, our findings cannot be extrapolated to women of all races or ethnicities, and this study should be replicated in other, more diverse cohorts. The latter is particularly true for populations like Hispanics, who are not only predisposed to MASLD, but are also more likely to have more severe disease(40).

To conclude, we show that maternal MASLD during pregnancy is generally mild, but hepatic fat fraction increases postpartum. In addition, clinically significant hepatic steatosis occurs in offspring as early as in utero, but maternal MASLD is not associated with this finding. Therefore, women found to have MASLD during pregnancy should be monitored for progression in the postpartum period. Future studies should also investigate whether clinically significant steatosis seen in utero warrants close monitoring for the detection of cardiometabolic comorbidities in later childhood. The natural history and pathophysiologic links between maternal and offspring hepatic steatosis need to be studied in large cohort studies.