Clinical characteristics and risk factors of nonalcoholic fatty liver disease in children with obesity

This study was performed in Nanjing, China, at the Children’s Hospital of Nanjing Medical University. The study protocol was approved by the Medical Ethics Committee of the Children’s Hospital of Nanjing Medical University (201412004.1). Children from 6 to 16 years old were recruited to the study from September 2015 to April 2018. The study was conducted in obese children. Inclusion criterion was body mass index (BMI) ≥95th percentile according to China’s children’s BMI classification. Exclusion criteria were any other endocrine diseases, viral hepatitis, hereditary diseases and other infectious or chronic diseases.

Pubertal stage was evaluated on the basis of the Tanner scale. Children with obesity were further divided into three subgroups by combining ultrasound and ALT 2-fold elevation: simple obese (SOB) group with normal liver in ultrasound, simple steatosis (SS) group with fatty liver in ultrasound and ALT < 80 U/L, and NASH group with fatty liver in ultrasound and ALT ≥80 U/L, according to the two-fold lab reference standard for impaired liver function [15].

Study staff gave oral and written information to parents/caregivers. The parent(s) or caregiver(s) of all subjects provided written informed consent before inclusion in the study. Written assent was obtained for all children. Through interview with children and their parents, experienced interviewers collected the basic information about the children, including birth history, family history and lifestyles.

All participants had anthropometric assessments. Experienced researchers measured height and weight using standardized measuring methods [19]. A digital scale (graduation 100 g) was used for weight measurement, and a gauge (graduation 1 mm) (Seca 704, Germany) for height measurement. BMI was calculated by dividing body weight (in kilograms) by the square of height (in meters). BMI standard deviation score (BMI-SDS) was calculated according to WHO reference values. At the end of the expiratory period, the waist circumference was measured at the navel with a non-retractable soft ruler to the nearest 0.1 cm [20]. Waist circumference (in centimeters) was divided by height (in centimeters) to calculate the waist to height ratio (WHtR). Body fat and skeletal muscle were measured by InBody J20 (Biospace, Korea). Percentage of body fat and percentage of skeletal muscle were calculated. After a quiet rest for 10 min, blood pressure was measured three times by an electronic sphygmomanometer (Omron HBP-1300) and the mean value was taken.

All subjects were told to fast overnight for 12 h before phlebotomy the next morning. Fasting laboratory assays included glucose, insulin, total cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, alanine aminotransferase (ALT), and aspartate aminotransferase (AST).

All subjects were examined after 12 h of fasting overnight. Hepatic ultrasonography was performed by ultrasound technician to establish a diagnosis according to at least two of the following criteria: higher echogenicity higher in the liver than in the spleen or kidney; blurred hepatic vasculature; high attenuation of signals [21].

The diagnosis of metabolic syndrome was defined as having at least three criteria with the following cut points [22]: abdominal obesity (waist circumference ≥ 90th percentile of waist circumference of children of the same age and gender [23], elevated triglycerides (≥1.47 mmol/L [130 mg/dL]) [24], low HDL cholesterol (< 1.03 mmol/L [40 mg/dL]), elevated blood pressure [systolic blood pressure (SBP) ≥90th percentile of SBP or diastolic blood pressure (DBP) ≥90th percentile of DBP of children of the same age and height and gender], and impaired fasting glucose (≥5.6 mmol/L [100 mg/dL]). In addition to the factors of metabolic syndrome, insulin resistance was defined according to homeostasis model assessment of insulin resistance (HOMA-IR) which was calculated as fasting insulin (mU/mL) × [fasting glucose (mmol/L)/22.5] [25]. The insulin resistance threshold was defined as ≥3.16, as described in the literature [26]. Hyperuricemia was defined as uric acid value ≥357 umol /L [27].

A total of 428 children with obesity were recruited for the study, including 148 (34.6%) with SOB, 235 (54.9%) with SS and 45 (10.5%) with NASH. Anthropometric characteristics are shown in Table 1. There was no significant difference in age or Tanner stage classification among the three groups. Obesity indicators such as BMI, BMI-SDS, waist circumference, and Body fat in the NASH group were significantly higher than in the SS and SOB groups, but percentage of skeletal muscle in the NASH group was significantly lower than in the other groups (p < 0.001). Percentage of body fat and WHtR in the SS and NASH groups were significantly higher than in the SOB group, but there was no significant difference between the first two groups, indicating that these two anthropometric measurements could not well distinguish the spectrum of NAFLD.

The biochemical parameters are shown in Table 2. Children in the NASH group had higher hepatic enzymes including ALT and AST, indicating liver injury. The triglyceride levels were significantly higher in the NASH and SS groups compared with the SOB group. The uric acid concentration and HOMA-IR levels involved in the pathogenesis of NAFLD were ranked as follows: SOB < SS < NASH.

The occurrence rates of metabolic syndrome in the NASH and SS groups were significantly higher than in the SOB group (53.3, 49.8% vs 19.6%, p < 0.001). Children with NASH had significantly elevated blood pressure and triglycerides compared with children in the SOB group, but there was no difference in the frequency of abdominal obesity, low HDL and impaired fasting glucose when dichotomous cut points were used among groups (Table 3).

Further investigation indicated that the prevalence of insulin resistance (38.5% / 46.8% / 68.9%) and hyperuricemia (45.9% / 68.5% /86.7%) were steadily higher from the SOB group to the SS group to the NASH group, respectively (p < 0.001). The progression of NAFLD was positively associated with cardiovascular risk factors and metabolic syndrome in these children with obesity. As shown in Fig. 1, all children in the SS and NASH groups had at least one metabolic syndrome factor. Moreover, the distribution of metabolic syndrome features in children in the NASH and SS groups was significantly (p < 0.001) shifted to the right, with more features present than in children with obesity in the SOB group.

After adjusting for all other variables among the children with obesity, the odds of having NASH were significantly higher in those with severe obesity (BMI-SDS ≥ 3) (OR 2.56; 95% CI 1.06–6.17) compared to those with mild obesity, males (OR 4.94; 95% CI 1.41–17.35) compared to females, those with hyperuricemia (OR 2.98; 95% CI 1.14–7.76) compared to those without hyperuricemia, and those with insulin resistance (OR 2.71; 95% CI 1.29–5.69) compared to those without insulin resistance (see Table 4).

In the context of the obesity epidemic among children in China, the rate of NAFLD has become increasingly serious in the past 20 years. In this study of 428 children with obesity, 54.9% had NAFLD, and 10.5% NASH. Moreover, children with NAFLD had a higher risk of cardiovascular and metabolic syndrome than those without it. Severe obesity, gender, hyperuricemia and insulin resistance are risk factors for NASH in children with obesity. Clinical evidence indicates that children with NAFLD experience an increased incidence and mortality of cardiovascular disease in adulthood [9]. Early identification of the clinical characteristics is therefore particularly important for Chinese children with NAFLD and NASH in light of the current obesity epidemic.

Due to differences in diagnostic criteria, selected population and ethnicity, the reported prevalence of NAFLD varies greatly in the literature [3]. A meta-analysis of epidemiological studies in children showed that an average prevalence of NAFLD was 7.6% in general population studies and 34.2% in studies of children with obesity [3]. Due to the limitations of the invasive diagnostic methods for NASH, there is currently a lack of population-based studies on NASH. A recent community-based study reported that 193 subjects (55%) were diagnosed with NAFLD, and 105 subjects (30%) were diagnosed with NASH by biopsy [16]. In our study, the incidence of NAFLD in children with obesity was 65.4%, which was higher than other reports, and the incidence of NASH was 10.5%. The risk of NASH in severely obese (BMI-SDS ≥ 3) children was 2.56 times higher than in mildly obese (2 ≤ BMI-SDS < 3) children. These results highlight that obesity is still an important trigger of NAFLD, although NAFLD and NASH can be found in the normal BMI population.

At the same time, our results demonstrated that the incidence of NASH in males was significantly higher than in females (14% vs. 2.4%), and gender was still a risk factor for NASH after adjusting for age, obesity and other components of the metabolic syndrome (OR = 4.94, 95% CI (1.41–17.35)). The gender difference in the distribution of NASH is similar to the results in adults. However, the reasons for this are not clear so far. It is possible that the fat distribution is more likely to accumulate in the viscera in males, which is more likely to cause metabolic disorders compared with subcutaneous fat, and that estrogen can improve the insulin sensitivity of adipose tissue in females [28].

In the present study, the NAFLD children showed higher cardiovascular risk and more metabolic syndromes than non-NAFLD children with obesity. This result is consistent to that reported in a previous case-control analysis [11]. The NAFLD children were more likely to develop abnormal glucose, insulin, triglycerides and HDL cholesterol than non-NAFLD children with matched age, sex and BMI [11]. A study covering 400 obese children also revealed that abnormal hepatic echo was associated with systolic and diastolic dysfunction [29]. Metabolic syndrome is a collection of cardiovascular risk factors that can predict diabetes and cardiovascular disease better than any single component. More and more evidence has shown that a large proportion of children with NAFLD meet the diagnostic standard of metabolic syndrome, suggesting that these children will develop diabetes and cardiovascular diseases in the future. This belief was supported by a 14-year-long Swedish cohort study of NAFLD patients diagnosed by biopsies. There was a 9% prevalence of diabetes at baseline, and most patients (78%) had developed impaired glucose tolerance or diabetes by the end of the 14-year period [18]. In the same study, the survival of NAFLD adults diagnosed by biopsy was lower than that of a matched control group, largely due to higher cardiovascular mortality [18]. Therefore, health care workers should pay more attention to the early identification of cardiovascular risk factors in the NAFLD population, which is helpful to facilitate targeted prevention and treatment strategies.

We found that the risk of insulin resistance in the NASH group was significantly higher than that in the SS and SOB groups, and insulin resistance was a risk factor for NASH after adjusting for confounding factors. Many clinical studies suggest that insulin resistance is always associated with NAFLD in adults and children [30‐33]. A biopsy-proven NAFLD study showed that 95% of children had insulin resistance [34]. In an experimental study, Bugianesi et al. found that adipose tissue is an important site for the development of insulin resistance, which can contribute to NAFLD [35]. However, the association between insulin resistance and NASH remains unclear. The pathogenesis of NAFLD is hepatocyte steatosis and hepatocyte injury, in which insulin resistance plays an important role in the development of NAFLD. As peripheral insulin resistance arouses in NAFLD patients, serum-free fatty acids, synthesized through the lipolysis of visceral fatty tissue and dietary fat, serve as the origin of hepatic triglycerides. Insulin resistance acts in metabolic disorders induced by free fatty acids and mitochondria. Lipid overload can also repress oxidative performances. Meanwhile, with the production of reactive oxygen species, cytokines are induced, inflammatory cells chemoattracted, and hepatic stellate cells activated, all leading to damage to hepatocytes [36]. Therefore, insulin resistance is involved in the occurrence and development of NAFLD and NASH. On account of the central role of insulin resistance in glucose metabolism, youth diagnosed with fatty liver need further evaluation to exclude diabetes or impaired glucose tolerance due to insulin resistance.

A growing number of clinical studies have shown that hyperuricemia is related to NAFLD in both adults and children [37‐39]. In a recent study of adolescents, hyperuricemia independently predicted [OR 2.5, 95% CI (1.87–2.83)] the presence of NASH after adjustment for potential confounders [40]. A biopsy-proven NAFLD study also found that the higher the serum uric acid, the greater the steatosis grade and lobular inflammation [41]. Similar to previous studies, we confirmed that obese patients with hyperuricemia had a higher risk of NASH occurrence [OR 2.98, 95% CI (1.14–7.76)]. Experimental studies observed that incubation of HepG2 cells with uric acid can increase intracellular triglycerides. The potential mechanisms may be that uric acid activates mitochondrial stress, subsequently activates endoplasmic reticulum stress and activates sterol regulatory element-binding protein 1C, while uric acid can stimulate the NOD-like receptor family pyrin domain containing 3 inflammasomes to regulate hepatic steatosis and inflammatory response in NAFLD patients [42, 43].

Firstly, we cannot confirm whether the presumed NASH patients do have abnormal hepatology, so we chose the combination of ultrasound and twice the upper limit of ALT to define NASH, in order to reduce the diagnostic error. Secondly, a total of 428 obese children were included in this study, nearly 70% being males. Thirdly, because of its cross-sectional nature, the causal relationship between NAFLD and cardiovascular risk factors was not fully explored. In addition, other potential factors such as dietary, lifestyle, and family genetic factors will be taken into account in future studies.