While non-alcoholic fatty liver disease (NAFLD) is associated with increased risk of impaired glucose tolerance and type 2 diabetes mellitus (DM) in non-pregnant patients, the clinical significance of NAFLD during pregnancy is still unclear. We hypothesized that sonographic findings of NAFLD during pregnancy would be associated with gestational diabetes mellitus (GDM) and predict abnormal postpartum glucose metabolism. NAFLD was assessed by ultrasound during and after pregnancy. Standard 2-hour 75 g oral glucose tolerance test (OGTT) was used during pregnancy and post partum to establish GDM and the diagnosis of normal, impaired fasting glucose, or DM. We also measured plasma insulin, C peptide, and free fatty acids (FFA) concentration during an OGTT to evaluate glucose tolerance, insulin secretion and insulin resistance. Of the 84 subjects, 12 had sonographic evidence of NAFLD (5 of whom had OGTT post partum). There was a non-significant trend toward higher mean weight and body mass index during and after gestation in the NAFLD group, but no statistically significant differences in mean age, ethnicity, prepregnancy and postpregnancy hemoglobin A1C values, and postpartum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), glucose, insulin, or FFA. We did not find an association between sonographic evidence of NAFLD during the third trimester of pregnancy and abnormal glucose metabolism during or after pregnancy. This study also suggests that while AST and ALT are not reliable diagnostic tools for NAFLD during the postpartum period, ultrasound is a reasonably safe, practical, and cost-effective modality to assess maternal hepatic fat during pregnancy.
- diabetes Mellitus
- fatty Liver
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Significance of this study
What is already known about this subject?
Non-alcoholic fatty liver disease (NAFLD) is associated with increased risk of impaired glucose tolerance, hyperinsulinemia, and type 2 diabetes mellitus in non-pregnant patients.
The prevalence of NAFLD is reported to be 14%–38% among women with history of gestational diabetes mellitus (GDM), which is higher than its prevalence among women without such history.
While NAFLD is associated with increased risk of cirrhosis and hepatocellular carcinoma, the prevalence of these conditions among women with GDM is currently unknown.
What are the new findings?
Sonographic evidence of NAFLD during the third trimester of pregnancy was not associated with abnormal glucose metabolism during or after pregnancy.
Alanine aminotransferase and aspartate aminotransferase are not reliable diagnostic tools for NAFLD in the postpartum period.
How might these results change the focus of research or clinical practice?
Ultrasound is a reasonably safe, practical, and cost-effective modality to assess maternal hepatic fat during pregnancy.
Early education, intervention, and disease-modifying lifestyle changes might be effective to delay and prevent not only diabetes mellitus, but also NAFLD, cirrhosis, and hepatocellular carcinoma in women with GDM.
The American Diabetes Association (ADA) defines gestational diabetes mellitus (GDM) as diabetes diagnosed in the second or third trimester of pregnancy in women who did not have clearly overt pregravid diabetes.1 Women with history of GDM are at increased risk of subsequent pre-diabetes and type 2 diabetes mellitus (DM), independent of pregravid metabolic risk factors.2–6 In the setting of rising prevalence and disease burden of DM, identifying women with GDM at higher risk of development of DM is important. Non-alcoholic fatty liver disease (NAFLD) is defined as hepatic steatosis in the absence of significant alcohol consumption and other secondary causes of steatosis (eg, drugs, autoimmune liver disease, viral infection, hemochromatosis).7 Both NAFLD and GDM are mediated by insulin resistance.1 7 NAFLD is associated with increased risk of impaired glucose tolerance, hyperinsulinemia, and type 2 DM in non-pregnant patients.8 9 Previous observational studies have reported NAFLD among 14%–38% of women with history of GDM and demonstrated a higher prevalence of NAFLD among these women, compared with those without such history.10 11 While NAFLD is a risk factor for both cirrhosis and hepatocellular carcinoma, the prevalence of these consequences among women with GDM is currently unknown.12 Existing studies are limited by being retrospective analyses of middle-aged and postmenopausal women, as well as lack of consistent adjusting for potential confounders. Using linked administrative data sets in Sweden, coded diagnosis of NAFLD prior to pregnancy was associated with a higher risk of GDM, compared with absence of a coded diagnosis of NAFLD.13
Only few studies to date have studied NAFLD in pregnancy. In one, hepatic fat was identified in a sample of five pregnant women with high body mass index (BMI).14 In a study of 57 women, the semiquantitative measure of NAFLD by liver ultrasound in the first or third trimester of pregnancy was similar in obese and normal BMI groups.15 A study of 476 women in Canada found a positive association between ultrasound evidence of hepatic fat in early pregnancy and abnormal glucose metabolism and GDM in mid-pregnancy.16 Among 608 Korean women, NAFLD (defined by either sonographic imaging or non-invasive steatosis indices) at 10–14 weeks of gestation was associated with subsequent GDM development.17 Moreover, the risk of developing GDM correlated with severity of NAFLD.17 Among 400 pregnant women in Egypt, women with ultrasound evidence of NAFLD during the first trimester of pregnancy were at greater risk of GDM development, compared with women without NAFLD.18 There are no published studies in the USA of NAFLD during pregnancy that have simultaneously assessed for GDM and abnormal postpartum glucose metabolism. This study examines the relationship between NAFLD during pregnancy and GDM as well as postpartum glucose metabolism. We hypothesized that NAFLD during pregnancy would be associated with GDM as well as abnormal fasting or postprandial glucose levels post partum, after adjusting for other metabolic risk factors.
Materials and methods
Recruitment was done at the Obstetrics and Gynecology clinic at University of Florida in Gainesville, Florida from 2013 to 2015. All women between 18 and 50 years of age who were in their third trimester of pregnancy were eligible to participate. Exclusion criteria included history of pregravid DM (type 1 or 2), presence of any chronic liver disease other than NAFLD, inability to understand and speak English, inability to provide consent, and plans to relocate out of the area within 6 months post partum. A questionnaire was used at the time of recruitment to determine baseline maternal characteristics of eligible patients who provided written informed consent, including maternal age, ethnicity, obstetric history, and family history. Subjects’ charts were reviewed to determine other variables, such as presence of GDM. We used the ADA criteria to define GDM: fasting plasma glucose >92 mg/dL, 1-hour glucose >180 mg/dL, or 2-hour glucose >153 mg/dL during a standard 2-hour 75 g oral glucose tolerance test (OGTT) done as part of routine clinical care during pregnancy.1 Consented subjects underwent a liver ultrasound during the last trimester, using E8 Voluson (GE Healthcare) and a 2–5 MHz abdominal probe. Fatty infiltration of the liver was diagnosed when the appearance of the liver had greater echogenicity than that of the maternal right kidney parenchyma.19 Using ultrasound to detect hepatic steatosis is reported to have a within-observer and between-observer reliability of 0.95, a sensitivity of 60%–95%, and a specificity of 84%–100%.20 We decided to perform the ultrasound during the third trimester of pregnancy, as insulin resistance worsens during pregnancy and such screening for NAFLD would be more sensitive during this period.
At 4–12 weeks post partum, subjects underwent another standard OGTT. We classified OGTT glucose <140 mg/dL as normal, 140–199 as impaired fasting glucose, and ≥200 as DM.1 Fasting and postprandial concentrations of plasma glucose, insulin, C peptide, and free fatty acids (FFA) were measured at −15, 0, 30, 60, 90, and 120 min. We also collected the following clinical and metabolic parameters: pregnancy outcomes, infant feeding modality, sociodemographic and behavioral risk factors (eg, employment, tobacco/alcohol use, exercise), height, weight, BMI, blood pressure, hemoglobin A1C (HbA1C), lipid panel, liver function tests, and adiponectin. At this postpartum visit, hepatic fat content was measured by a right upper quadrant ultrasound. One author (RE) interpreted all hepatobiliary imaging. All other testing was done without the knowledge of ultrasound results.
We expressed continuous data as mean (95% CI) and categorical data as percentages, unless specified otherwise. Comparisons between two groups were performed by Student’s t-test or Kruskal-Wallis (depending on distribution) for numeric variables and by χ2 or Fisher’s exact test for categorical variables. A two-tailed p value <0.05 was regarded as statistically significant. Analyses were performed with Stata V.15.0.
While 126 women initially consented to the study, 42 subjects were lost to follow-up, withdrew from the study, or did not have interpretable hepatic ultrasound images prior to completion of the study. Table 1 summarizes the characteristics of the 84 study subjects at the time of recruitment/study entry. Twelve subjects (14.3%) had sonographic evidence of NAFLD and 49 (58.3%) were diagnosed with GDM. While there were not any statistically significant differences between the group with NAFLD and the group without NAFLD in terms of mean age, ethnicity, blood pressure, obstetric history, GDM during current pregnancy, family history of DM, and prepregnancy HbA1C, there was a trend toward higher weight and BMI in the NAFLD group. In addition, more women in the NAFLD group reported a history of GDM in previous pregnancies and family history of DM.
Seventy-four subjects had postpartum data available for analysis (table 2). The NAFLD group continued to have a statistically non-significant higher average weight and BMI, compared with the group without NAFLD. These two groups did not have statistically significant differences in terms of postpartum blood pressure, alanine aminotransferase (ALT), aspartate aminotransferase (AST), fasting plasma glucose, or HbA1C.
Twenty-seven subjects underwent postpartum OGTT. There were not any significant differences in glucose, insulin, or FFA values during the OGTT between the group with and without NAFLD (figure 1). Similarly, the two groups did not demonstrate statistically significant differences in postpartum parameters for insulin secretion or insulin resistance (table 3).
Based on previous medical literature, we had hypothesized that NAFLD would be associated with GDM as well as postpartum glycemic dysregulation, and therefore would help identify women who would benefit from early testing for GDM as well as interventions to prevent future DM. However, we did not find a statistically significant association between NAFLD during the third trimester of pregnancy and GDM or abnormal postpartum glucose metabolism. The prevalence of NAFLD in our study (14.3%) was similar to the Canadian (17.6%) and Korean (18.4%) studies of NAFLD during pregnancy,16 17 but lower than the NAFLD study among 57 pregnant volunteers (50%).15
The non-significant trend toward higher weight and BMI during and after pregnancy in the NAFLD group is consistent with the well-established link between NAFLD and obesity in the general population.7 However, studies to date examining the association of NAFLD during pregnancy and obesity have yielded conflicting results. One study detected hepatic fat via ultrasound or liver biopsy in a sample of five pregnant or recently pregnant women with elevated BMI, without GDM, and with otherwise unexplained elevated ALT and AST.14 On the other hand, Castracane et al 15 found no difference in the distribution of NAFLD in lean and obese pregnant women. Similarly, Mousa et al 18 reported similar mean BMI in women with NAFLD and women without NAFLD. Of note, while NAFLD is associated with increased concentrations of ALT in cross-sectional studies in the general population,21 two of the studies of NAFLD during pregnancy failed to find an association between plasma ALT elevation and NAFLD or GDM during pregnancy.15 16 Possible explanations for the lack of an association were the relatively young age of subjects and/or the effect of pregnancy on ALT levels, including hemodilution.16 Interestingly, another study reported an association of NAFLD during the first trimester of pregnancy with elevated AST, but not ALT.18 In our study, ALT and AST were measured post partum, thereby minimizing the influence of pregnancy on their concentrations. However, there were no significant differences between postpartum values of plasma AST and ALT between the groups with versus without NAFLD. This finding adds to previous literature in that AST and ALT are unreliable diagnostic tools for NAFLD during pregnancy or after delivery.
Similar to our study, Castracane et al 15 did not find an association between sonographic evidence of NAFLD during pregnancy and GDM. However, more recent larger studies of 476 pregnant women in Canada, 608 pregnant women in Korea, and 400 pregnant women in Egypt reported a positive association between maternal hepatic fat on ultrasound in early pregnancy and development of GDM.16–18 It should be noted that in addition to being much larger than our study and being performed outside of the USA, the subjects underwent hepatic sonographic imaging earlier in pregnancy in these studies.16–18 It is possible that steatosis, a dynamic process, is altered during different trimesters due to altered metabolism during pregnancy. A strength of our study is that it is the first published study to date that has prospectively examined the relationship between NAFLD and glucose metabolism during pregnancy as well as in the postpartum period. Lack of a statistically significant association between NAFLD during the third trimester of pregnancy and postpartum glucose metabolism abnormality could have resulted from the small size of the study, the timing of the sonographic evaluation during pregnancy, the small number of subjects with NAFLD as well as subjects who underwent postpartum OGTT. Another biologically plausible explanation, since metabolic syndrome, GDM, and NAFLD are all driven by insulin resistance, is that young women with GDM might go on to develop NAFLD±DM later in life, as demonstrated by observational studies among middle-aged and postmenopausal women.10 11 In other words, while we hypothesized that NAFLD would be an early marker for developing GDM as well as postpartum fasting or postprandial hyperglycemia, developing GDM might be the earliest marker for the development of metabolic syndrome and NAFLD later in life. Interestingly, more women in the NAFLD group in this study reported a history of GDM in their previous pregnancies, compared with the group without NAFLD. Early education, intervention, and disease-modifying lifestyle changes might be effective to delay and prevent not only DM, but also NAFLD, cirrhosis, and hepatocellular carcinoma in women with GDM.7
Another strength of this study is that the ultrasound used for assessment of hepatic fat is a simple and standardized imaging modality that is safe in pregnancy and can be conveniently done at the same time as one of the fetal ultrasounds. As a limitation, ultrasound is a semiquantitative measure of hepatic fat, cannot distinguish various stages of NAFLD, and might not detect small amounts of hepatic steatosis.22 In fact, ultrasound might yield negative results unless hepatic fat content exceeds 30%.23 Since few patients in this study had hepatic steatosis, larger studies will be needed to confirm the current findings. Further limitations of this study are its low retention rate as a large number of patients dropped out and did not complete the planned studies, including postpartum OGTT. Larger studies, with longer postpartum follow-up and assessing novel imaging modalities (ie, controlled attenuation parameter; FibroScan) during pregnancy, will be needed to further elucidate the prevalence of NAFLD in pregnant women, as well as the relationship between NAFLD during pregnancy and glucose metabolism during pregnancy and after delivery.
Sonographic evidence of NAFLD during the third trimester of pregnancy was not associated with GDM or abnormal postpartum glucose metabolism in this study. However, the prevalence of NAFLD in this study population was only 14.3%. Larger studies with longer postpartum follow-up are needed to further clarify the prevalence of NAFLD during pregnancy and examine the relationship between NAFLD during pregnancy and abnormal glucose metabolism during and after pregnancy. Our study demonstrates that using ultrasound during fetal assessment is a safe, practical, and cost-effective modality to assess maternal hepatic fat during pregnancy, when invasive procedures, such as liver biopsy, are not recommended.
Contributors All the listed authors have contributed to design, carrying out the study and analyses, and writing of the manuscript.
Funding This work was supported by the University of Florida Gatorade Trust.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval The institutional review board at the University of Florida approved this observational study.
Provenance and peer review Not commissioned; internally peer reviewed.
Data availability statement Data are available upon reasonable request.