Highlights
- •Maternal obesity has been linked to NAFLD in offspring.
- •All biopsy-proven NAFLD cases in Sweden aged 25 or younger were matched to controls.
- •Data on maternal body mass index and socio-economic confounders were recorded.
- •Maternal obesity was a risk factor for NAFLD in offspring.
- •Obesity might have inter-generational consequences.
Background & Aims
Maternal obesity has been linked to the development of cardiovascular disease and diabetes in offspring, but its relationship to non-alcoholic fatty liver disease (NAFLD) is unclear.
Methods
Through the nationwide ESPRESSO cohort study we identified all individuals ≤25 years of age in Sweden with biopsy-verified NAFLD diagnosed between 1992 and 2016 (n = 165). These were matched by age, sex, and calendar year with up to 5 controls (n = 717). Through linkage with the nationwide Swedish Medical Birth Register (MBR) we retrieved data on maternal early-pregnancy BMI, and possible confounders, in order to calculate adjusted odds ratios (aORs) for NAFLD in offspring.
Results
Maternal BMI was associated with NAFLD in offspring: underweight (aOR 0.84; 95% CI 0.14-5.15), normal weight (reference, aOR 1), overweight (aOR 1.51; 0.95-2.40), and obese (aOR 3.26; 1.72-6.19) women. Severe NAFLD (biopsy-proven fibrosis or cirrhosis) was also more common in offspring of overweight (aOR 1.94; 95% CI 0.96-3.90) and obese (aOR 3.67; 95% CI 1.61-8.38) mothers. Associations were similar after adjusting for maternal pre-eclampsia and gestational diabetes. Socio-economic parameters (smoking, mother born outside the Nordic countries and less than 10 years of basic education) were also associated with NAFLD in offspring but did not materially alter the effect size of maternal BMI in a multivariable model.
Conclusions
This nationwide study found a strong association between maternal overweight/obesity and future NAFLD in offspring. Adjusting for socio-economic and metabolic parameters in the mother did not affect this finding, suggesting that maternal obesity is an independent risk factor for NAFLD in offspring.
Lay summary
In a study of all young persons in Sweden with a liver biopsy consistent with fatty liver, the authors found that compared to matched controls, the risk of fatty liver was much higher in those with obese mothers. This was independent of available confounders and suggests that the high prevalence of obesity in younger persons might lead to a higher risk of fatty liver in their offspring.
Graphical abstract
Graphical Abstract
Keywords
Introduction
Changes in food quality and a more sedentary lifestyle have led to a high prevalence of obesity globally.
[1]
Trailing this epidemic is the rapid increase in the prevalence of non-alcoholic fatty liver disease (NAFLD), now the most common liver disease worldwide,[2]
which affects an estimated 25% of the global population. Obesity has also become increasingly common even early in life, including in women of reproductive age.[3]
This does not only have consequences for affected women, but maternal obesity is also a risk factor for obesity, type 1 diabetes and cardiometabolic disease in offspring.4
, 5
, 6
, 7
Additional human studies have shown that adolescents with ultrasound-diagnosed NAFLD often have obese mothers.
[8]
A recently formulated hypothesis based on preclinical data suggests that adaptions to maternal obesity in early life impact on the risk of metabolic disorders (e.g. NAFLD) in offspring, both in rodent models and in humans.9
, 10
, 11
, 12
, 13
It is unclear if the risk of severe NAFLD is also increased and if the effect of maternal obesity on liver-related disease in offspring can be confounded by other factors. An increase in risk could be explained partly by intrinsic maternal factors such as obesity, but possibly also by socio-economic determinants. Such a distinction might be important, since societal and individual changes including reduced food availability and intake, or increased education might improve obesity-associated diseases such as NAFLD if there are no “programmed” behaviours which are generally less susceptible to intervention.Herein, we hypothesised that increased maternal BMI in early pregnancy is a risk factor for biopsy-proven NAFLD, especially severe NAFLD, in offspring.
Patients and methods
Study population
We performed a population-based case-control study using the ESPRESSO (Epidemiology Strengthened by Histopathology Reports in Sweden) cohort.
[14]
The ESPRESSO cohort holds detailed data on liver histopathology from all 28 Swedish pathology departments (1965-2017), defining histopathology findings using Systematized Nomenclature of Medicine (SNOMED) coding, such as steatosis, non-cirrhotic fibrosis and cirrhosis.[15]
Reports also include the unique Swedish personal identity number,[16]
which we used to link ESPRESSO data to several national registers containing validated prospectively recorded data on demographics including socio-economic data and development of diseases with diagnoses made at hospital level (available since 1964 but nationwide from 1987), and since 2001 also on outpatient visits in specialised healthcare.[17]
Finally, data were linked to the Swedish Medical Birth Register (MBR) with data on BMI since 1992. The MBR contains data from the first antenatal visit until delivery and discharge from the delivery hospital.[18]
For this study, we identified all liver biopsy specimens with a SNOMED diagnosis of hepatic steatosis. The process of identifying NAFLD has been validated (positive predictive value 92%) and has been described elsewhere.
[19]
We excluded participants aged >25, since they were by definition born prior to 1992 (with no data on maternal BMI in the MBR).We then excluded individuals with a competing diagnosis that could potentially cause steatosis, such as alcohol-related liver disease, viral hepatitis or rare paediatric liver diseases that could cause steatosis (Table S1). We further identified the NAFLD population with more severe disease, defined as presence of liver fibrosis or cirrhosis based on SNOMED coding (M4900x or M4950x). As we expected the number of individuals with cirrhosis to be low, cases with fibrosis or cirrhosis were merged into 1 subgroup, defined as “severe NAFLD”.
We then matched each NAFLD case with up to 5 controls from the general population. Controls were systematically sampled by the central authority “Statistics Sweden” that hold detailed census-level data on all Swedish citizens through the Total Population Register.
[20]
Controls were matched at the time of first liver biopsy in the index individual. Matching criteria were sex, exact age, county and calendar year. Exclusion and inclusion criteria were identical for cases and controls (Table S1).Next, we identified mothers to cases and controls.
[21]
From the MBR, we obtained data on maternal and pregnancy-related parameters that a priori were considered important for the development of NAFLD in offspring (Table 1). Parameters obtained from the MBR included maternal age, calendar year, maternal country of birth, smoking status, and level of education. Maternal country of birth was categorised as born in a Nordic country (yes/no). Smoking status was divided into three categories (non-smoker, smoking 1-9, or 10 or more cigarettes per day, or missing data), and level of education into four categories (≤9 years, 10-12 years; ≥13 years, missing). Information on height and weight was collected at the first antenatal visit. Height was self-reported, while weight was measured by a midwife. BMI was calculated and categorised as underweight (<18.5 kg/m2), normal weight (18.5-24.9 kg/m2, used as the reference category), overweight (25.0-29.9 kg/m2) and obesity (≥30 kg/m2).Table 1Characteristics of patients with NAFLD and matched population comparators (birth year 1992–2016).
| Characteristic | NAFLD (n = 165) | Controls (n = 717) | p value |
|---|---|---|---|
| Sex, n (%) | |||
| Women | 65 (39.4%) | 285 (39.7%) | 0.93 |
| Men | 100 (60.6%) | 432 (60.3%) | |
| Age at index date (years) | |||
| Median (IQR) | 12.0 (4.4–16.9) | 11.7 (3.4–16.3) | |
| Categories, n (%) | |||
| <11y | 70 (42.4%) | 324 (45.2%) | 0.75 |
| 11-<18y | 70 (42.4%) | 297 (41.4%) | |
| 18-≤25 | 25 (15.2%) | 96 (13.4%) | |
| Birth year, n (%) | |||
| 1992–1999 | 102 (61.8%) | 435 (60.7%) | 0.96 |
| 2000–2009 | 60 (36.4%) | 268 (37.4%) | |
| 2010–2016 | 3 (1.8%) | 14 (2.0%) | |
| Year of index date, n (%) | |||
| 1992–2000 | 20 (12.1%) | 93 (13.0%) | 0.85 |
| 2001–2010 | 56 (33.9%) | 255 (35.6%) | |
| 2011–2016 | 89 (53.9%) | 369 (51.5%) | |
| Comorbidities before index date, n (%) | |||
| Cardiovascular disease | 13 (7.9%) | 13 (1.8%) | <0.001 |
| Diabetes | 9 (5.5%) | 3 (0.4%) | <0.001 |
| Hypertension | 4 (2.4%) | 0 | <0.001 |
| Dyslipidaemia | 0 | 0 | — |
| NAFLD severity, n (%) | |||
| Fibrosis | 71 (43.0%) | — | |
| Cirrhosis | 5 (3.0%) | — | |
| Cirrhosis or fibrosis | 76 (46.1%) | — | |
| Maternal and delivery characteristics | |||
| Maternal BMI at first visit (kg/m2) | |||
| Median (IQR) | 25.0 (22.0–29.0) | 23.3 (21.3–26.6) | |
| Categories, n (%) | |||
| <18.5 | 5 (2.8%) | 27 (3.8%) | <0.001 |
| 18.5–<25 | 77 (46.8%) | 436 (60.8%) | |
| 25–<30 | 51 (31.2%) | 194 (27.0%) | |
| ≥30 | 32 (19.3%) | 60 (8.4%) | |
| Gestational age (days) | |||
| Median (IQR) | 279 (272–285) | 281 (273–287) | |
| Categories (weeks), n (%) | |||
| <37 | 15 (9.1%) | 34 (4.7%) | 0.05 |
| 37–41 | 142 (86.1%) | 630 (87.9%) | |
| ≥42 | 8 (4.8%) | 53 (7.4%) | |
| Birth weight (grams) | |||
| Median (IQR) | 3,350 (3,090–3,773) | 3,590 (3,265–3,945) | <0.001 |
| Maternal smoking in early pregnancy | |||
| Non-smoking | 130 (78.7%) | 606 (84.5%) | 0.04 |
| 1–9 cig/day | 16 (9.7%) | 69 (9.7%) | |
| ≥10 cig/day | 19 (11.6%) | 42 (5.8%) | |
| Birth order - Parity | |||
| 1 | 65 (39.4%) | 312 (43.5%) | 0.33 |
| 2 | 57 (34.5%) | 255 (35.6%) | |
| ≥3 | 43 (26.1%) | 150 (20.9%) | |
| Caesarean section, n (%) | 22 (13.3%) | 102 (14.2%) | 0.77 |
| Gestational diabetes, n (%) | 4 (2.4%) | 7 (1.0%) | 0.13 |
| Pre-eclampsia, n (%) | 20 (12.1%) | 48 (6.7%) | 0.02 |
| Maternal diabetes, n (%) | 1 (0.6%) | 2 (0.3%) | 0.52 |
| Maternal age at birth (years) | |||
| Median (IQR) | 28.7 (25.0–33.5) | 29.7 (26.0–33.0) | |
| Highest level of education in parents | |||
| ≤9 years | 11 (6.7%) | 15 (2.1%) | 0.003 |
| 10–12 years | 77 (46.7%) | 306 (42.7%) | |
| ≥13 years | 77 (46.7%) | 396 (55.2%) | |
| Country of birth in mother, n (%) | |||
| Nordic | 120 (72.7%) | 618 (86.2%) | <0.001 |
| Other | 45 (27.3%) | 99 (13.8%) | |
| Living with partner | |||
| Yes | 140 (84.8%) | 641 (89.4%) | 0.24 |
| No/missing | 25 (15.2%) | 76 (10.6%) | |
NAFLD, non-alcoholic fatty liver disease.
∗ Student’s t test, Chi-squared test, or Wilcoxon-Mann-Whitney test were used as appropriate.
We also retrieved data on diagnoses of commonly occurring comorbidities of NAFLD in the offspring from the National Patient Register.
[17]
These included cardiovascular disease, diabetes type 1 or 2, hypertension and hyperlipidaemia (definitions in Table S2).For a subsequent sensitivity analysis, we used the Total Population Register
[20]
to identify all full siblings to patients with a NAFLD biopsy, who then served as the control population, consistent with our prior work.[19]
,[22]
Sibling analyses have the advantage of addressing potential intrafamilial confounding due to shared genetics and early environmental factors.Statistical analysis
We estimated adjusted ORs for the risk of NAFLD in offspring based on BMI categories using conditional logistic regression. As the causal pathway between maternal factors such as a high BMI and offspring NAFLD is unknown, we considered two statistical models a priori. First, we constructed a model adjusted only for the matching factors (age at NAFLD diagnosis, sex and municipality). Next, we used a model adjusted for the matching factors plus the following maternal factors: age, country of birth, education, parity and smoking at the time of first entry in the MBR.
Missing data for BMI (18.2% in cases and 18.3% in controls) and smoking (7.3% in cases and 5.7% in controls) were imputed using a multiple imputation regression model by fully conditional specification methods with 5 iterations.
[23]
Regression estimates from each of the 5 sets of data were combined using the MIANALYZE procedure in SAS (v9.4).In sensitivity analyses, we added pre-eclampsia and gestational diabetes to the second model. We did so to explore whether any association between maternal obesity and NAFLD in offspring might be mediated through pregnancy-related metabolic factors.
We chose not to adjust for parameters in the offspring such as diabetes, as such parameters were considered likely to be part of the causal pathway, which could introduce bias.
[24]
A second sensitivity analysis was restricted to complete-case data (no imputations), replicating the above analyses. In a third sensitivity analysis we compared NAFLD cases with full sibling comparators. A fourth analysis stratified the results by offspring sex. Finally, we explored univariate associations with NAFLD in offspring in the regression analyses.
Ethical considerations
The study was approved by the Stockholm Ethics Review Board on August 27, 2014 (No.2014/1287-31/4). Informed consent was waived as the study was registry based.
[25]
Results
We identified 718 cases with a liver biopsy-based diagnosis of NAFLD below the age of 25 years and without any differential diagnosis. From these, we excluded 530 cases with a birthdate prior to 1992, and therefore no data on maternal BMI. Further, we excluded 20 cases who were not born in Sweden, and 3 cases who were not singleton births. Thus, our sample consisted of 165 cases of NAFLD. These were matched with 717 controls, subject to the same exclusion criteria. A flowchart for study inclusion is presented in Table S1.
Most NAFLD cases were diagnosed after 2010. They had a median age of 12.0 years (IQR 4.4–16.9), and 60.6% were male (Table 1). Interestingly, offspring with later NAFLD had a lower birth weight compared to matched controls (median 3.35 vs. 3.59 kg, p <0.001).
General population analyses
Maternal BMI was higher in cases with NAFLD compared with controls (Table 1). Logistic regression revealed a higher prevalence of maternal obesity in offspring with NAFLD (19.3%) compared with controls (8.4%), with evidence of a dose-response effect of maternal BMI (ptrend=0.006). Compared to mothers with a normal BMI, the risk of NAFLD in offspring was 3-fold higher (aOR 3.26; 95% CI 1.72–6.19) in mothers with a BMI ≥30 kg/m2. This risk was not statistically significant and was numerically lower in overweight mothers (aOR 1.51; 95% CI 0.95–2.40), and was also not seen in underweight mothers (aOR 0.84; 95% CI 0.14–5.15).
Further, 76 cases (46%) with NAFLD fulfilled our criteria for severe NAFLD (presence of fibrosis: n = 71 or cirrhosis: n = 5). The risk of severe NAFLD in offspring was increased in both obese (aOR 3.67; 95% CI 1.61–8.38) and overweight (aOR 1.94; 95% CI 0.96–3.90) mothers. The estimates for NAFLD in offspring are presented in Table 2, and for severe NAFLD in Table 3.
Table 2Odds ratio for biopsy-proven NAFLD in offspring by maternal BMI category using conditional logistic regression.
| Maternal BMI | NAFLD (n = 165) | Controls (n = 717) | OR (95% CI) | OR (95% CI) |
|---|---|---|---|---|
| <18.5 | 5 (2.8%) | 27 (3.8%) | 0.86 (0.16–4.81) | 0.84 (0.14–5.15) |
| 18.5–<25 (reference) | 77 (46.8%) | 436 (60.8%) | 1.00 | 1.00 |
| 25–<30 | 51 (31.2%) | 194 (27.0%) | 1.50 (0.97–2.30) | 1.51 (0.95–2.40) |
| ≥30 | 32 (19.3%) | 60 (8.4%) | 3.07 (1.72–5.50) | 3.26 (1.72–6.19) |
NAFLD, non-alcoholic fatty liver disease; OR, odds ratio.
∗ Conditioned on matching set (age, sex, county and calendar year).
∗∗ Conditioned on matching set and further adjusted for maternal age, maternal country of birth, parity, highest level education in parents, and smoking in early pregnancy.
Table 3Odds ratio for biopsy-proven severe NAFLD (cirrhosis or fibrosis) in offspring by maternal BMI category using conditional logistic regression.
| Maternal BMI | NAFLD (n = 76) | Controls (n = 334) | OR (95% CI) | OR (95% CI) |
|---|---|---|---|---|
| <18.5 | 1 (1.8%) | 12 (3.6%) | 0.71 (0.08–5.98) | 0.72 (0.08–6.32) |
| 18.5–<25 (reference) | 31 (40.3%) | 198 (59.3%) | 1.00 | 1.00 |
| 25–<30 | 27 (35.8%) | 92 (27.5%) | 1.89 (0.94–3.77) | 1.94 (0.96–3.90) |
| ≥30 | 17 (22.1%) | 32 (9.6%) | 3.48 (1.61–7.52) | 3.67 (1.61–8.38) |
NAFLD, non-alcoholic fatty liver disease; OR, odds ratio.
∗ Conditioned on matching set (age, sex, county and calendar year).
∗∗ Conditioned on matching set and further adjusted for maternal age, maternal country of birth, parity, highest level education in parents, and smoking in early pregnancy.
Adjusting for maternal pre-eclampsia and gestational diabetes yielded similar results (Table S3), while slightly higher risk estimates were seen in our complete case analysis (Table S4) than in our main analysis.
Besides maternal BMI, socio-economic factors were significantly linked to NAFLD in offspring (Table 4). Compared to women born outside the Nordic countries, the offspring of women born in the Nordic countries had a significantly lower risk of NAFLD (aOR 0.35; 95% CI 0.22–0.57). Smoking ≥10 cigarettes per day was associated with an increased risk of NAFLD in offspring (aOR 2.13; 95% CI 1.07–4.25), as was less than 10 years of completed education, albeit this association was not statistically significant (aOR 2.22; 95% CI 0.94–5.26).
Table 4Univariable and multivariable conditional logistic regression for biopsy-proven NAFLD based on parameters other than BMI.
| Parameter | Model 1 | Model 2 | ||
|---|---|---|---|---|
| OR (95% CI) | p value | OR (95% CI) | p value | |
| Maternal age (continuous) | 0.99 (0.96–1.02) | 0.47 | 0.99 (0.95–1.02) | 0.46 |
| Nordic country of birth | 0.39 (0.25–0.60) | <0.001 | 0.35 (0.22–0.57) | <0.001 |
| Nulliparous (yes/no) | 0.84 (0.60–1.19) | 0.34 | 0.96 (0.65–1.43) | 0.85 |
| Maternal smoking in early pregnancy | ||||
| Non-smoking (reference) | 1.09 (0.59–2.02) | 0.77 | 0.89 (0.48–1.66) | 0.71 |
| 1–9 cig/day | 1.00 | — | 1.00 | — |
| ≥10 cig/day | 2.16 (1.20–3.91) | 0.01 | 2.13 (1.07–4.25) | 0.03 |
| Highest level of education in parents | ||||
| ≤9 years | 2.74 (1.23–6.09) | 0.01 | 2.22 (0.94–5.26) | 0.07 |
| 10–12 years (reference) | 1.00 | — | 1.00 | — |
| ≥13 years | 0.76 (0.53–1.09) | 0.14 | 1.02 (0.69–1.52) | 0.92 |
Model 1 presents univariable associations, that is the association between the parameter and NAFLD in the offspring. Model 2 presents the multivariable-adjusted associations for parameters other than BMI, which is presented in Table 3.
NAFLD, non-alcoholic fatty liver disease; OR, odds ratio.
∗ Conditioned on matching set (age, sex, county and calendar year).
∗∗ Conditioned on matching variables, and further adjusted maternal age, maternal country of birth, parity, highest level education in parents, and smoking in early pregnancy.
Sibling analyses
In the sibling analysis, we compared 108 cases with NAFLD with their 156 siblings. Maternal BMI was similar between cases with NAFLD and their sibling controls (median 25.2 vs. 25.3) (Table S5). After multivariable adjustment, we found no association between maternal obesity and NAFLD in offspring (aOR 1.38; 95% CI 0.35–5.39) (Table S6).
Stratification by sex
There were 65 females and 100 males among the offspring. The odds of NAFLD in the offspring of obese mothers were comparable for male (aOR 4.22; 95% CI 1.68–10.59) and female (aOR 2.87; 95% CI 1.22–6.79) offspring (Table S7A,B).
Discussion
In this national, population-based case-control study, we demonstrate an increased risk of biopsy-proven NAFLD in offspring born to mothers with a high early-pregnancy BMI. This excess risk seems to be independent of several important socio-economic factors, as well as of smoking and gestational diabetes, that were otherwise linked to future risk of NAFLD. In fact, adjusting for the available socio-economic parameters increased the ORs for maternal BMI somewhat (from 3.07 to 3.26), suggesting that the association between maternal BMI and offspring NAFLD is unlikely to be fully explained by such factors.
Our results are largely consistent with those from preclinical rodent models.
9
, 10
, 11
The prevalence of paediatric NAFLD in the US doubled between the end of the 1980s and 2010, when it was estimated at around 10%.[26]
Moreover, a secondary analysis of the Western Australian Pregnancy Cohort study found that out of 1,170 17-year-old adolescents, 15.2% had ultrasound-defined NAFLD, and maternal obesity was a risk factor for NAFLD in offspring, with an OR of 3.16, which is very similar to our point estimate.[8]
Our results supplement these findings by confirming the presence and severity of NAFLD by means of liver biopsy data, and further demonstrating that maternal BMI is linked to disease severity, even after adjustment for important maternal clinical factors and socio-economic parameters. Further, we show that adjusting for gestational diabetes and pre-eclampsia, as proxies for more severe metabolic disease, did not affect the risk of NAFLD in offspring.In the sibling-comparison, maternal obesity was not associated with NAFLD in offspring. We cannot rule out that the observed association between maternal BMI and risk of NAFLD in offspring is mediated by other factors like foetal growth, comorbidities, diet or exercise that could differ between siblings, or that our sibling analysis was underpowered (precluding any sub-analyses). Another explanation could be that there is undiagnosed NAFLD in the siblings. In Sweden and elsewhere, siblings to individuals with biopsy-verified NAFLD do not routinely undergo liver biopsy for screening purposes, and this is not recommended in international guidelines.
[27]
The NAFLD diagnosis was most often made in older rather than younger children, suggesting that NAFLD is less common in younger children. Metabolic comorbidities such as diabetes and hypertension were more common in patients with NAFLD compared to controls, suggesting a shared dysmetabolic milieu. Moreover, we found that key socio-economic parameters such as being born to immigrant mothers, smoking and lower education were risk factors for NAFLD in offspring, suggesting that groups with such characteristics are at a heightened risk of NAFLD and could be considered for focused public health interventions. We speculate that such factors are likely to be markers of a more sedentary lifestyle and a less healthy diet, which to a large extent is then adopted by the offspring, leading to a higher risk of NAFLD. Indeed, previous studies have shown that individuals with lower education have higher BMI and more type 2 diabetes than the general population.
[28]
,[29]
Given the known association between obesity and NAFLD, it is likely safe to assume that such factors also affect the prevalence of NAFLD and could help identify at-risk populations for public health interventions.With the increasing prevalence of overweight and obesity in the population,
[1]
including in pregnant women,[3]
our results suggest that the future prevalence of NAFLD in the paediatric and adolescent populations will increase, most likely continuing into later life. It has previously been shown that a high BMI early in life is associated with development of severe liver disease,30
, 31
, 32
and these results suggest that being exposed to obesity while at a reproductive age might also have cross-generational consequences. This further highlights the importance of obtaining a healthy lifestyle and a normal BMI prior to pregnancy, as part of family planning.Women at reproductive age with an elevated BMI should receive active advice and education on how to reduce the risk of obesity-related conditions in themselves and their offspring, such as improvement in diet and physical exercise.
The current study has several strengths. First, it was nationwide and population-based, allowing us to identify all individuals with a biopsy-based diagnosis of NAFLD in Sweden during the study period. Second, maternal BMI and data on confounders were derived from national registries with prospectively collected and validated data, which reduces the risk of recall bias (often a threat to the internal validity of case-control studies). Third, we had data on liver biopsy which is the gold standard for both diagnosing and staging NAFLD and we have recently shown that the positive predictive value for NAFLD in this cohort is 92%.
[19]
Finally, we were able to exclude cases and controls with competing liver diseases.Limitations include a risk of selection bias in that by nature of the study design, biopsy was mandated. Thus, we might have only captured the most severe cases of NAFLD, supported by the high prevalence of fibrosis in our study. However, our results should be generalisable to countries similar to Sweden. Additionally, we lack information on the indication for liver biopsy, but the high prevalence of fibrosis and cirrhosis suggest that suspicion of advanced NAFLD was a prominent reason for biopsy, which is in accordance with guidelines for when to perform a biopsy in paediatric NAFLD.
[27]
The disease severity staging was derived from administrative coding that did not allow for more granular staging of fibrosis or the presence of non-alcoholic steatohepatitis (NASH), such as defined by the NASH clinical research network system.[33]
We did not have data on ethnicity but used country of birth as a proxy. NAFLD status was not systematically ascertained in siblings, so the results of full sibling comparisons should be interpreted with caution. We lacked detailed data on breastfeeding, which has been suggested to protect against NAFLD in offspring,[34]
,[35]
but we also lacked data on diet, moderate alcohol consumption and maternal exercise habits.Even if we had access to prospective data on potential socio-economic and medical confounders, we cannot rule out residual confounding, especially in diet and physical activity patterns in mothers. As such, maternal obesity might not be a causal factor in that it induces specific changes in the foetal metabolism leading to a higher tendency of NAFLD or other metabolic diseases in offspring. An alternative hypothesis, partly supported by these data, is that mothers with NAFLD are more exposed to socio-economic determinants of poor health. Nevertheless, these data certainly suggest that maternal obesity is a marker and risk factor for NAFLD in offspring. Finally, we did not have granular data on lifestyle factors in offspring. While such factors cannot impact on maternal BMI and are hence not confounders, they might have helped to explain the association with NAFLD in offspring seen in this paper.
In this population-based case-control study, we show that maternal BMI early in pregnancy is an independent risk factor for the diagnosis and severity of NAFLD in their offspring. As obesity is increasing, this has implications for the future prevalence of NAFLD. Mothers with an elevated BMI should receive active counselling on how to reduce the risk of NAFLD in their offspring.
Abbreviations
aOR, adjusted odds ratio; ESPRESSO, Epidemiology Strengthened by Histopathology Reports in Sweden; MBR, Swedish Medical Birth Register; NAFLD, non-alcoholic fatty liver disease. NASH, non-alcoholic steatohepatitis. OR, odds ratio; SNOMED, Systematized Nomenclature of Medicine.
Financial support
Dr Hagström is supported by grants from Region Stockholm (clinical postdoctoral appointment). Dr. Simon is supported by grants from the National Institutes of Health (NIH) (K23 DK122104), the Dana-Farber/Harvard Cancer Center (DF/HCC) Gastrointestinal (GI) Specialized Program in Research Excellence (SPORE).
Authors’ contributions
Study conception and design: All. Acquisition of data: JFL. Statistical analysis: JS, BR. Analysis and interpretation of data: All. Drafting of manuscript: HH, TGS, JFL. Critical revision: All. Guarantor of article: JFL. All authors approved the final version of the article, including the authorship list.
Data availability statement
Due to the confidentiality of data, the data which support the findings of this study are generally not available due to current regulations. However, requests for additional analyses might be considered upon reasonable request.
Conflict of interest
Dr Hagström’s institution has received independent research grants from Intercept, Gilead, Astra Zeneca, EchoSens, MSD and Pfizer. Board advisory for Gilead and Bristol-Myers Squibb. These concern works unrelated to this project. Dr. Simon has received grants to the institution from Amgen and has served as a consultant to Aetion for work unrelated to this project. Dr Ludvigsson coordinates a study on behalf of the Swedish IBD quality register (SWIBREG). This study has received funding from Janssen corporation.
Please refer to the accompanying ICMJE disclosure forms for further details.
Supplementary data
The following are the supplementary data to this article:
- Multimedia component 1
- Multimedia component 2
- Multimedia component 3
References
- Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: a pooled analysis of 2416 population-based measurement studies in 128.9 million children, adolescents, and adults.Lancet. 2017; 390: 2627-2642
- Global epidemiology of nonalcoholic fatty liver disease-Meta-analytic assessment of prevalence, incidence, and outcomes.Hepatology (Baltimore, Md). 2016; 64: 73-84
- Estimated global overweight and obesity burden in pregnant women based on panel data model.PloS one. 2018; 13e0202183
- Maternal body mass index, gestational weight gain, and the risk of overweight and obesity across childhood: an individual participant data meta-analysis.PloS Med. 2019; 16e1002744
- Maternal overweight and obesity and risk of congenital heart defects.J Am Coll Cardiol. 2019; 73: 44-53
- Maternal obesity as a risk factor for early childhood type 1 diabetes: a nationwide, prospective, population-based case-control study.Diabetologia. 2018; 61: 130-137
- Maternal obesity and offspring's cardiovascular morbidity - results from a population based cohort study.Early Hum Dev. 2020; 151: 105221
- Infant nutrition and maternal obesity influence the risk of non-alcoholic fatty liver disease in adolescents.J Hepatol. 2017; 67: 568-576
- Maternal obesity during pregnancy and lactation programs the development of offspring non-alcoholic fatty liver disease in mice.J Hepatol. 2010; 52: 913-920
- Maternal obesity programs offspring nonalcoholic fatty liver disease by innate immune dysfunction in mice.Hepatology (Baltimore, Md). 2013; 58: 128-138
- Developmental programming of peripheral diseases in offspring exposed to maternal obesity during pregnancy.Am J Physiol Regul Integr Comp Physiol. 2020; 319 (R507-r16)
- Early life origins of metabolic disease: developmental programming of hypothalamic pathways controlling energy homeostasis.Front Neuroendocrinol. 2015; 39: 3-16
- Multigenerational maternal obesity increases the incidence of HCC in offspring via miR-27a-3p.J Hepatol. 2020; 73: 603-615
- Cohort profile: ESPRESSO (epidemiology strengthened by histoPathology Reports in Sweden).Clin Epidemiol. 2019; 11: 101-114
- Progress in medical information management. Systematized nomenclature of medicine (SNOMED).Jama. 1980; 243: 756-762
- The Swedish personal identity number: possibilities and pitfalls in healthcare and medical research.Eur J Epidemiol. 2009; 24: 659-667
- External review and validation of the Swedish national inpatient register.BMC Publ health. 2011; 11: 450
- A quality study of a medical birth registry.Scand J Soc Med. 1990; 18: 143-148
- Mortality in biopsy-confirmed nonalcoholic fatty liver disease: results from a nationwide cohort.Gut. 2020;
- Registers of the Swedish total population and their use in medical research.Eur J Epidemiol. 2016; 31: 125-136
- The Swedish Pregnancy Register - for quality of care improvement and research.Acta Obstet Gynecol Scand. 2018; 97: 466-476
- Association between celiac disease and mortality risk in a Swedish population.Jama. 2020; 323: 1277-1285
- Multiple imputation by fully conditional specification for dealing with missing data in a large epidemiologic study.Int J Stat Med Res. 2015; 4: 287-295
- Overadjustment bias and unnecessary adjustment in epidemiologic studies.Epidemiology (Cambridge, Mass). 2009; 20: 488-495
- Ethical aspects of registry-based research in the Nordic countries.Clin Epidemiol. 2015; 7: 491-508
- Increasing prevalence of nonalcoholic fatty liver disease among United States adolescents, 1988-1994 to 2007-2010.J Pediatr. 2013; 162: 496-500.e1
- Liver biopsy in children: position paper of the ESPGHAN Hepatology Committee.J Pediatr Gastroenterol Nutr. 2015; 60: 408-420
- Global region of birth is an independent risk factor for type 2 diabetes in Stockholm, Sweden.Eur J Publ Health. 2017; 27: 447-453
- Education and diabetes in a racially and ethnically diverse population.Am J Publ Health. 2006; 96: 1637-1642
- High BMI in late adolescence predicts future severe liver disease and hepatocellular carcinoma: a national, population-based cohort study in 1.2 million men.Gut. 2018; 67: 1536-1542
- Overweight in late adolescence predicts development of severe liver disease later in life: a 39years follow-up study.J Hepatol. 2016; 65: 363-368
- Body mass index in early pregnancy and future risk of severe liver disease: a population-based cohort study.Aliment Pharmacol Ther. 2019; 49: 789-796
- Design and validation of a histological scoring system for nonalcoholic fatty liver disease.Hepatology. 2005; 41: 1313-1321
- Longer lactation duration is associated with decreased prevalence of non-alcoholic fatty liver disease in women.J Hepatol. 2019; 70: 126-132
- A protective effect of breastfeeding on the progression of non-alcoholic fatty liver disease.Arch Dis Child. 2009; 94: 801-805
Article info
Publication history
Published online: July 18, 2021
Accepted:
June 29,
2021
Received in revised form:
June 22,
2021
Received:
March 22,
2021
Identification
Copyright
© 2021 The Author(s). Published by Elsevier B.V. on behalf of European Association for the Study of the Liver.
User license
Creative Commons Attribution (CC BY 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution (CC BY 4.0)
Permitted
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article
- Reuse portions or extracts from the article in other works
- Sell or re-use for commercial purposes
Elsevier's open access license policy
01899-7&id=ga1.jpg){kind=link}