Higher BMI Predicts Liver Fibrosis Among Obese Adolescents with NAFLD - an Interventional Pilot Study

Background Non-alcoholic fatty liver disease (NAFLD) can range from simple steatosis to steatohepatitis with or without brosis. The predictors for liver brosis, so as the effect of nutritional intervention on hepatic brosis in pediatric population are not well established. We aimed to investigate the predictors for liver brosis and the effects of short-term nutritional intervention on steatosis and brosis among obese adolescents with NAFLD. Methods Cross-sectional study among Obese adolescents. Sociodemographic and clinical data were collected. Liver brosis was estimated by Shearwave elastography. All participants were recommended to consume a low carbohydrate diet and were followed biweekly. Blood tests and elastography were performed upon admission and repeated after 3 months.


Conclusion
Pediatric patients with high BMIs and severe liver steatosis are at risk for severe liver brosis. Nutritional intervention with minimal weight loss may improves hepatic brosis among pediatric population.

Background
Non-alcoholic fatty liver disease (NAFLD) is the most common etiology of chronic liver disease in adults and children in the developed world [1], with prevalence of 3-10% in the general pediatric population and up to > 70% in obese children [2,3]. NAFLD encompasses a wide spectrum of histological and clinical manifestations, ranging from simple steatosis with debatable clinical signi cance to non-alcoholic steatohepatitis (NASH), with or without brosis that may develop into cirrhosis and liver failure, even in young children [3]. Therefore, in view of the increase prevalence of obesity among children and adolescents, it is of upmost importance to identify young patients at risk for advanced brosis who may develop cirrhosis and liver failure. Only few studies have aimed to nd predictors for advanced brosis in pediatric NAFLD patients [4][5][6]. Moreover, the current mainstay of treatment for both adult and pediatric Loading [MathJax]/jax/output/CommonHTML/jax.js NAFLD is weight loss, but, the effect of dietary intervention on hepatic brosis in pediatric population is not well established [7][8].
The aims of the present study were to investigate predictors for liver brosis in obese pediatric patients, and to assess the effects of short-term dietary intervention on steatosis and brosis.

Patient Population
We prospectively recruited all children and adolescents (age 7-18 years) with obesity who were admitted to the Obesity clinic at Dana-Dwek Children's Hospital of the Tel Aviv Medical Center between December 1, 2018, and December 1, 2019. All children with a BMI > 95 percentile for age were included in the study.
Patients with a diagnosed primary liver disease (autoimmune liver disease, metabolic liver disease, Wilson`s disease, alpha 1 antitrypsin de ciency), patients treated with medications known to induce steatosis (such as valproate, amiodarone or prednisone), and patients with hepatic virus infections or history of parenteral nutrition were excluded from the study.

Study Design and measurements
This study is part of a clinical trial that asses the effect of bariatric surgery and dietary intervention on hepatic brosis in obese pediatric population with NAFLD (Clinical Trial Registration-NCT04561804). At the initial visit, data were collected on socioeconomic parameters, lifestyle, birth details, and medical, family, and social histories. All patients underwent anthropometric measures (height, weight and BMI) and a physical examination focused on obesity-related conditions. Laboratory evaluation included liver enzyme pro le, lipid pro le, glucose, insulin, and HbA1C. Liver brosis was estimated by Shearwave elastography (Supersonic) and categorized into 4 levels, F0-F4, according to liver stiffness (measured by kPa), as recently demonstrated elsewhere [9]. Liver steatosis was calculated by a hepatorenal index (HRI), as described by Webb et al [10] and divided into 3 levels of severity. All measurements were taken by a single experienced radiologist (MW), who was blinded to the results of other parameters of the patients. A multidisciplinary team included a gastroenterologist, hepatologists, a registered dietitian (RD) and a psychologist. All participants received nutritional recommendations (see below) and general recommendation for a healthy lifestyle (regular engagement in daily physical activity and reduction of screen time). Compliance with the dietary guideline was reviewed by a RD on a biweekly basis with a 3day food questionnaire (2 weekdays and 1 day of the weekend). Blood tests and elastography were repeated after 3 months of intervention.
Determination of the BMI percentiles for age and gender was based on WHO growth charts. "Obesity" was de ned as a BMI > 95th percentile [11]. Abnormal glucose metabolism included taking a hypoglycemic medication or having an elevated homeostatic model assessment index of insulin resistance, glucose, or HBA1c. Hypertriglyceridemia and hypercholesterolemia were de ned as a serum level > 95 percentile for age and sex [12]. Hypertension was diagnosed as systolic and/or diastolic blood pressure ≥ 95 percentile Loading [MathJax]/jax/output/CommonHTML/jax.js for age and sex [13]. Patients with clinical suspicion of obstructive sleep apnea (OSA) were diagnosed by polysomnography that was conducted by the hospital ss ≤ epspecialists. Socioeconomicstatuswasdef ∈ eda or d ∈ g → parents years of education.

The dietary intervention
The participants received nutritional recommendations for a low carbohydrate, low glycemic load, and isocaloric diet. The diet was composed of carbohydrates (CHO;30-40%), fats (35-50%), and proteins (20-25%), and was tailored to individual preferences and calorie requirements. The number of CHO, protein and fat servings was determined based on the recommended total energy requirements for age, calculated on the basis of dietary reference intake (DRI). Participants were not instructed to restrict calories, but to reduce carbohydrate based on their glycemic load. High glycemic index (GI) carbohydrate intake (re ned grains, potatoes, sweet and salted snacks and sugar sweetened beverages) was completely restricted, low GI carbohydrates (non-starchy vegetables, legumes, nuts,) were allowed and some low /moderate GI carbohydrate were allowed but limited such as fruits and whole grain bread.
The subjects were instructed about appropriate food choices, and each participant was provided a diet information booklet containing food list, sample menus and recipes

Statistical Analyses
Descriptive statistics were examined for all variables. Continuous variables were expressed as median with interquartile range (IQR) when they were not normally distributed and as mean ± standard deviation (SD) for normally distributed variables. Categorical variables were presented as number and percentage. The Pearson correlation and simple linear regression analysis were performed to examine bivariate associations between brosis and metabolic and nutritional parameters. The Wilcoxon signed rank test was applied to compare the difference between steatosis, brosis, and metabolic parameters between the 2 time points (baseline and 3-month follow-up). A P level < 0.05 was considered statistically signi cant. All statistical tests were 2-sided. The statistical analysis was performed with SPSS (IBM SPSS statistics,

Ethical Considerations
The study protocol was approved by the institutional review board of the medical center (TLV-0097-17). Signed informed consent was obtained from the parents of all the participants. The study was design in accordance to the CONSORT guidelines.

Description of Overall Study Sample
Loading [MathJax]/jax/output/CommonHTML/jax.js Ninety-ve consecutive children with obesity were recruited. Fifteen patients were excluded for refusal to undergo elastography examination, 13 patients were excluded for invalid elastography examinations and 10 patients were subsequently excluded due to missing data. The nal cohort consisted of 57 patients [35 (61%) males, at a mean age of 13.5 ± 2.9 years and a mean BMI of 38.8 ± 9.7 (Table 1). Baseline blood tests demonstrated impaired fasting glucose in 22 subjects (39%), elevated triglycerides in 26 (45%), and hypercholesterolemia in 14 (25%). Hypertension and OSA were documented in 5 patients (9% each) ( Table 1). Fifty-three (92%) subjects were diagnosed with liver steatosis upon admission to the clinic. A total of 34 (60%) patients had liver brosis which was moderate/severe (F ≥ 2) in 24 (70%) of them prevalent in subjects with brosis levels of ≥ F2 compared to subjects with minimal or no brosis (43.8 ± 9.5 vs 34.9 ± 8, respectively, P < 0.001) (   (Table 4, P = 0.008). After the nutritional intervention a signi cant decrease in the incidence of liver brosis and steatosis was noted (Fig. 1). Before the dietary intervention, 3 patients (18%) had F4 brosis, 5 (35%) had F3 brosis, 3 (18%) had F2 brosis, and 4 (23%) had F1 brosis. After the dietary intervention, none of the patients had F4 brosis. The level of brosis decreased to F2 in 1 patient with F4 and to F2 in 2 patients. The brosis decreased from F3 to F2 in 1 patient and from F3 to F1 in 2 patients. Among the 3 patients with F2 brosis, the level of brosis decreased to F1 in 2 and to F0 in one. There were 4 patients with F1 who had complete normalization of the brosis after the dietary intervention. (P = 0.001, Fig. 1a-b). A similar improvement was also noted in the liver fat content as measured by the HRI (Fig. 1, d-c). These changes were also accompanied by a signi cant decrease in ALT and triglyceride serum levels (from 61 ± 34 mg/dl before intervention to 42 ± 26.4 mg/dl after intervention and from 147.6 ± 68 mg/dl before intervention to 102.2 ± 44.4 mg/dl after intervention, respectively), with no signi cant difference in LDL or total cholesterol levels (   Values are given ± standard deviation.

Discussion
The results of the present study indicate that higher BMI levels and moderate/severe liver steatosis are predictors for severe liver brosis in children and adolescents. Three months of nutritional intervention based on a low carbohydrate diet improved hepatic steatosis and brosis in a pediatric population with NAFLD.
Our data corroborate those of others [3,7,14,15] by demonstrating a high rate of NAFLD with a signi cant percent of moderate to severe brosis in morbidly obese young individuals, reaching approximately 70% of our patients. Only few attempts have been made to stratify the risk for advanced brosis in this unique population [6][7][8]. Moreover, recent data have suggested that adult scores may not be accurate to predict advanced brosis in children [7,[16][17], thus establishing a clear need to evaluate noninvasive approaches in children as well. The pediatric NAFLD brosis index is based on age, waist circumference, and triglycerides, and it has been described by Nobili et al as a possible tool to predict liver brosis in children [4]. It is, however, limited by not including children with moderate/severe brosis. The recent pediatric NAFLD brosis score which included ALT, alkaline phosphatase, platelet counts, and gamma glutamyl transferase was reported to predict the presence of signi cant brosis, but it lacks external validation [6].
Our current results demonstrated that moderate/severe brosis correlated with higher BMI levels and moderate/severe steatosis. This reinforces previous ndings which demonstrated that children and adolescent with severe obesity (BMI ≥ 120% of the 95th percentile or an absolute BMI ≥ 35 kg/m²) are more prone to severe complications, such as cardiovascular disease, dyslipidemia and in ammation [18,19] compared to children and adolescents with obesity and lower BMI levels. This highlights the need for early dietary intervention, even among youngsters, before further complications develop and the severity increases.
The only metabolic parameter that was signi cantly related to moderate/severe brosis was lower LDL. Moreover, we found a trend for higher triglyceride levels among patients with lower brosis levels. These results may re ect the recent NASH Clinical Research Network data which demonstrated that zone 1 steatosis, while rare in adult populations, was highly prevalent in children with NAFLD, and that it represents a distinct sub-phenotype with unique metabolic and histologic parameters. Children with zone 1 steatosis had lower fasting triglyceride levels and lower fasting insulin according to the NASH report. However, zone 1 steatosis was found to have more brosis of any grade (81% vs 51) and more advanced brosis (13% vs 5%) compare to children with zone 3 steatosis [20]. Our ndings did not include biopsy data, but these unique differences in metabolic parameters between subjects with moderate/severe brosis to patients with minimal or no brosis may also serve to emphasize the need for early intervention in NAFLD patients even if no other metabolic disorder is present.
Seventeen of our patients completed 3 months of follow-up with dietary interventions and repeated blood tests and elastography. There was a signi cant decrease in the BMI Z score with a signi cant decrease in liver brosis and steatosis scores at the end of follow-up. Moreover, ALT and triglycerides serum levels decreased signi cantly as well. There are several possible explanations for the signi cant restitution of liver brosis that was documented in our study after only 3 months. First, it may be due to the weight loss itself that was documented in our cohort. Reduction of visceral fat depots after weight loss protects against the over ow of fatty acids to the liver [21][22]. Increased availability of fatty acid, in turn, is pivotal to the pathogenesis of fatty liver, causing mitochondrial dysfunction and lipotoxicity [22]. Second, it may be due to the speci c dietary intervention. The change in liver fat in our study occurred without major weight loss. This was also described in other studies of adults and children [21][22][23], suggesting the possibility of clinical bene t solely with low carbohydrate dietary modi cation, since a lower glycemic response causes less hepatic glucose absorption [24][25][26]. Several clinical trials demonstrated that a reduction of fructose or sugar consumption resulted in lower intrahepatic fat, lipogenesis, in ammation, and insulin resistance [24][25][26]. Moreover, because this diet does not restrict either fat or protein, it may also be more behaviorally sustainable and can therefore result in better adherence over time [28]. Lastly, it could be that the rapid and signi cant reversal in liver histology, compared to the adult population, stems from the differences in histologic distribution among the 2 populations in terms of in ammation and hepatocellular damage [16,20,29].
The main limitation of our study is the lack of liver biopsies for assessing NAFLD, which is still considered the gold standard for NAFLD diagnosis. However, the well-known limitations of liver biopsy Loading [MathJax]/jax/output/CommonHTML/jax.js and the fact that liver biopsy cannot be applied to all patients suspected of having NAFLD have led to the development of noninvasive methods for the assessment of liver brosis. Shear-wave elastography was recently shown to be an accurate and reproducible noninvasive technique that e ciently depicts the presence of liver brosis in the pediatric population with NAFLD [9,17], with high levels of repeatability and reproducibility and a high intra-observer (ICC = 0.89-0.90) and inter-observer (ICC = 0.81-0.85) coe cients [30][31]. Other limitations of our study are the lack of a control group and the 3-month followup period that may not have been long enough to observe the full extent of in uence of macronutrient contents on NAFLD and brosis. Nevertheless, the prospective nature of this study and the fact that each patient serves as his/her own control enables us to draw important conclusions about the need for early intervention in the obese pediatric population with NAFLD, and be encouraged by the results that testify to the ability of histological improvement if appropriate treatment is offered in time.

Conclusion
Our study ndings reveal that a higher BMI carries a greater risk for advanced liver brosis in the pediatric population. A low carbohydrate, low glycemic index diet may improve hepatic steatosis and brosis already after a 3-month period. Longitudinal and larger cohorts are needed to compare the effectiveness of a low carbohydrate diet with that of other dietary interventions for preventing the progression of NAFLD toward more severe forms of liver derangements early in its natural history.

Declarations
Ethics approval and consent to participate The study protocol was approved by the "Helsinki" institutional review board of the medical center. reference number -TLV-0097-17. Signed informed consent was obtained from the parents of all the participants.

Consent for publication not applicable
Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.