Nonalcoholic fatty liver disease (NAFLD) is a spectrum characterized by hepatic fat accumulation which ranges from simple steatosis to non-alcoholic steatohepatitis (NASH) and cirrhosis [1]. NAFLD is potentially one of the most common causes of liver disease worldwide both in adults and children [2-4]. Considering its ever increasing and epidemic proportions, its timely identification and optimum management should be a priority in the present times [5].

This prospective observational study was undertaken in the departments of Hepatology and Pediatric Hepatology in a tertiary care institute. The study was approved by the Institutional Review Board. The patients were enrolled after informed consent and assent was obtained from the parent(s) and patients. The duration of the study was from Ist October 2014 to 31st December 2016. Inclusion criteria included: (i) all overweight/obese children (aged 8-18 years, overweight defined as a body mass index or BMI ³85th　percentile to <95th　 percentile, and obesity defined as a BMI ³95th percentile, for children and teens of the same age and sex) and their parents, and (ii) adults (along with their spouses) suffering from diabetes mellitus (DM), obesity, dyslipidemia, metabolic syndrome or NAFLD having overweight/obese progeny (aged 8-18 years). Exclusion criteria included (i) history suggestive of acute hepatitis in last 6 months, (ii) abnormal thyroid profile, (iii) Wilson disease (³1 of following- low ceruloplasmin/increased urinary copper/Kayser Fleischer ring +ve), (iv) hepatitis B/C infection, (v) concomitant liver diseases, (vi) severe malnutrition, (vii) ongoing total parenteral nutrition/jejunal-ileal bypass, (viii) alcohol intake of more than 20g/week, (ix) syndromic obesity, (x) medication use like steroids, estrogens, Amiodarone, Methotrexate, Tamoxifen and antitubercular therapy.

Statistical analysis: The mean differences between the groups were tested by independent sample t-test. The chi square (or Fisher’s exact) test was used to compare differences between the groups for categorical variables. Variables affecting Family Clustering were analyzed using univariate and multivariate analysis. Data was analyzed by using SPSS 22 version. Probability of Y (Outcome, for e.g, Pediatric NAFLD) is predicted by: Probability (Y) = 1/1 + e – (b0 + b1X1 + b2X2...) where P(Y) is the probability of Y occurring, e is the base of natural logarithms, X1/X2/X3 etc are predictor variables, bo/b1/b2 etc are Beta Coefficients [16].

A total of 99 overweight and obese children were included in the study, with 69 subjects in NAFLD group (Fig. 1). In the NAFLD group, there were 59 boys (85.5%) with median age of 13.1 years, while in the non- NAFLD group, there were 21 boys (70%) with a median age of 11.1 years. There was high incidence of metabolic diseases in the families having children with NAFLD where more than 3/4th of the families had atleast one parent with either fatty liver (80%) or low HDL levels (84 %). Similarly there was high incidence (>2/3rd of families) of insulin resistance, hypertension and high triglycerides in atleast one parent in the NAFLD group. In the NAFLD group, homozygosity (GG status) and heterozygosity (CG status) for PNPLA3 polymorphism was seen in 24 (34.8%) and 23 (33.3%) overweight/obese children respectively. In the non-NAFLD group, only 1 subject had homozygous mutation, while heterozygous status was found in 8 (26.7%).

Fig. 1 Flowchart depicting patient selection in the study. (*NAFLD diagnosis based on ultrasonographic evidence of fatty liver).

Analysis of presence or absence of family history of metabolic risk factors (NAFLD, hypertension, insulin resistance/IR, type 2 diabetes mellitus, dyslipidemia and presence of metabolic syndrome) in the parents showed that presence of NAFLD in any one parent (OR 3.9, 95% CI 1.5 to 10.6; P=0.008) or both parents (OR 6.7, 95 % CI 1.4 to 30.6; P=0.009) and presence of insulin resistance in any parent (OR 3.6, 95 % CI 1.1 to 12.0; P= 0.009) or both parents (P=0.01) was significantly associated with occurrence of NAFLD in the progeny (Table I). Amongst the clinical parameters, presence of acanthosis and higher mean BMI significantly differentiated NAFLD from non- NAFLD group (Table II). In the laboratory features, presence of higher mean serum aspartate aminotransferase (AST) levels, higher mean serum alanine aminotransferase (ALT) levels, higher mean uric acid levels, higher mean cholesterol levels, high mean fasting insulin levels, higher Homeostasis model assessment of insulin resistance-1 (HOMA-1) index, higher HOMA-2 index and lower Quantitative insulin sensitivity check index (QUICKI), presence of insulin resistance and presence of homozygosity of PNPLA3 polymorphism predicted occurrence of fatty liver disease in children (Table II).

TABLE I  Comparison of Demographic Features of the NAFLD versus Non-NAFLD Group 

TABLE II	Comparison of Clinical and Laboratory Features of NAFLD Versus Non-NAFLD Group

Only 11 of the total NAFLD group gave consent for liver biopsy. The results showed simple steatosis in 5 and presence of NASH (with NAS score ³5) in 6 children. In the 6 patients with NASH, 1 patient had stage 3 fibrosis, 3 patients had stage 2 fibrosis while 2 patients showed stage 1 fibrosis. Due to limited number of subjects, no statistical analysis could be performed.

On comparing transient elastography features (LSM and CAP), higher controlled attenuation parameter (CAP) values significantly differentiated NAFLD from non NAFLD group while there was no significant difference on LSM between two groups (Table II). At a cut-off of 259.5 dB/m, CAP could predict presence of NAFLD in children with 88.4 % sensitivity and 100 % specificity and AUROC of 0.965 (95 % CI 0.931 to 1.000) (Fig. 2).

Fig. 2 Receiver operating characteristic (ROC) curve for Controlled Attenuation Parameter (by transient elastography) for NAFLD prediction.

Multivariate binary logistic regression analysis showed that family history of NAFLD (in any parent), higher ALT levels and higher total cholesterol levels may independently predict presence of NAFLD (Table III). Only ALT levels reached significance on calculating area under receiver operating characteristic (AUROC) curve, where in an overweight/obese child, ALT levels > 31.5 IU/L predicted presence of NAFLD with 80.6% sensitivity and 60.9% specificity with an AUROC of 0.825 (95 % CI 0.743 to 0.908).

TABLE III	Results of Multivariate Analysis (NAFLD versus Non NAFLD) 

Parameter	P	Adj OR (95% CI)
*Family history of NAFLD	0.004	18.81 (2.55, 138.94)
Alanine aminotransferase	0.014	1.08 (1.02, 1.15)
Total cholesterol	0.012	1.05 (1.01, 1.08)
*(Any parent); NAFLD: Non-alcoholic fatty liver disease; HDL-C: High Density Lipoprotein Cholesterol.

Based on the logistic regression model of Pediatric NAFLD in overweight/obese children, probability of developing Pediatric NAFLD was derived by the equation: P (Y) = 1/1 + e – [-7.088 + (2.935 × Family History of NAFLD in any parent) + (0.075 × ALT) + (0.045 × Cholesterol)] where ‘Family History of NAFLD in any parent’ is 0 or 1 when absent or present, respectively. The above equation leads us to a probability of developing NAFLD in overweight/obese children as 92.4 % based on the sample data.

Higher ALT levels and higher total cholesterol levels also independently predicted pediatric NAFLD in the present study. In the present study, each 10 unit increase in ALT (in IU/L) and each 20 unit increase in total cholesterol (in mg/dL) increased the risk of pediatric NAFLD by approximately 1.5 times and 2 times, respectively. In previous studies, serum ALT had been used as a screening tool for NAFLD in children [17-20]. In the present study, 28.9 % children with NAFLD had normal ALT values. We had used the adult cut off values (³40 IU/L) for defining normal ALT values to allow comparison with available literature. If we use the proposed normal ‘pediatric’ ALT values (i.e 25.8 U/L in boys and 22.1 U/L in girls), frequency of abnormal ALT increased from 71% to 88.4% (61 out of 69 children) [21]. Thus, though ALT independently predicted NAFLD, this limitation highlights the importance of not depending upon ALT alone to diagnose NAFLD since we may miss upto 12-29 % children. Similarly, Schwimmer, et al. [22] had also found that children with NAFLD had significantly higher serum total cholesterol, fasting glucose, insulin, low density lipoprotein (LDL), and triglycerides (TG) levels and significantly lower HDL than those without NAFLD. Huang, et al. [23] had also shown that higher body mass index (BMI) and ALT levels were significant independent predictors of pediatric NAFLD.

1. Lazo M, Clark JM. The epidemiology of nonalcoholic fatty liver disease: A global perspective. Semin Liver Dis. 2008;28:339-50.

2. Day CP. Non-alcoholic fatty liver disease: a massive problem. Clin Med. 2011;11:176-8.

3. Matthiessen J, Velsing Groth M, Fagt S, Biltoft-Jensen A,　Stockmarr A, Andersen JS, et al. Prevalence and trends in overweight and obesity among children and adolescents in Denmark. Scand J Public Health. 2008;36:153-60.

4. Ji CY, Cooperative Study on Childhood Obesity: Working Group on Obesity in China (WGOC): The prevalence of childhood overweight/obesity and the epidemic changes in 1985-2000 for Chinese school-age children and adolescents. Obes Rev. 2008;9:78-81.

5. Feldstein AE,　Charatcharoenwitthaya P,　Treeprasertsuk S,　Benson JT,　Enders FB,　Angulo P. The natural history of non-alcoholic fatty liver disease in children: a follow-up study for up to 20 years. Gut. 2009;58:1538-44.

6. Struben VM, Hespenheide EE, Caldwell SH. Nonalcoholic steatohepatitis and cryptogenic cirrhosis within kindreds. Am J Med. 2000;108:9-13.

7. Willner IR,　Waters B,　Patil SR,　Reuben A,　Morelli J,　Riely CA. Ninety patients with non- alcoholic steatohepatitis: insulin resistance, familial tendency, and severity of disease. Am J Gastroenterol. 2001;96:2957-61.

8. Abdelmalek MF,　Liu C,　Shuster J,　Nelson DR,　Asal NR. Familial aggregation of insulin resistance in first-degree relatives of patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol. 2006;4:1162-9.

9. Schwimmer JB, Celedon MA, Lavine JE, Salem R, Campbell N, Schork NJ, et al. Heritability of non-alcoholic fatty liver disease. Gastroenterology. 2009;136:1585-92.

10. Loomba R, Abraham M, Unalp A, Wilson L,　Lavine J,　Doo E, et al. and the Nonalcoholic Steatohepatitis Clinical Research Network. Association between diabetes, family history of diabetes, and risk of nonalcoholic steatohepatitis and fibrosis. Hepatology. 2012; 56:943-51.

11. Krawczyk M, Portincasa P, Lammert F. PNPLA3-Associated Steatohepatitis: Toward a Gene-Based Classification of Fatty Liver Disease. Semin Liver Dis. 2013;33:369-79.

12. Bhatt SP,　Nigam P,　Misra A,　Guleria R,　Pandey RM,　Pasha MA, et al. Genetic variation in the patatin-like phospholipase domain-containing protein-3 (PNPLA-3) gene in Asian Indians with nonalcoholic fatty liver disease. Metab Syndr Relat Disord. 2013;11:329-35.

13. Valenti L, Alisi A, Galmozzi E, Bartuli A, Del Menico B, Alterio A, et al. I148M patatin-like phospholipase domain-containing 3 gene variant and severity of pediatric non-alcoholic fatty liver disease. Hepatology. 2010;4:1274-80.

14. Lin YC,　Chang PF,　Hu FC,　Yang WS,　Chang MH,　Ni YH. A common variant in the PNPLA3 gene is a risk factor for non-alcoholic fatty liver disease in obese Taiwanese children. J Pediatr. 2011;158:740-4.

15. Misra A, Khurana L. The metabolic syndrome in South Asians: epidemiology, determinants, and prevention. Metab Syndr Relat Disord. 2009;7:497-514.

16. Field, Andy. Discovering Statistics Using SPSS, 3rd ed. London: SAGE Publications; 2009.

17. Fraser A, Longnecker MP, Lawlor DA. Prevalence of elevated alanine aminotransferase among US adolescents and associated factors: NHANES 1999-2004. Gastroenter-ology. 2007;133:1814-20.

18. Wiegand S,　Keller KM,　Röbl M,　L’Allemand D,　Reinehr T,　Widhalm K, et al; APV-Study Group and the German Competence Network Adipositas. Obese boys at increased risk for nonalcoholic liver disease: evaluation of 16,390 overweight or obese children and adolescents. Int J Obes. 2010;34:1468-74.

19. Welsh JA,　Karpen S,　Vos MB. Increasing prevalence of nonalcoholic fatty liver disease among United States adolescents, 1988-1994 to 2007-2010. J Pediatr. 2013;162):496-500.

20. Yang HR,　Yi DY,　Choi HS. Comparison between pediatric health promotion center and pediatric obesity clinic in detecting metabolic syndrome and nonalcoholic fatty liver disease in children. J Korean Med Sci. 2014;29:1672-7.

21. Schwimmer JB, Dunn W, Norman GJ, Pardee PE, Middleton MS,　Kerkar N, et al. SAFETY study: Alanine amino transferase cutoff values are set too high for reliable detection of pediatric chronic liver disease. Gastroenter-ology 2010;138:1357-64.

22. Schwimmer JB, Zepeda A, Newton KP, Xanthakos SA, Behling C, Hallinan EK, et al; Nonalcoholic Steatohepatitis Clinical Research Network. Longitudinal assessment of high blood pressure in children with nonalcoholic fatty liver disease. PLoS One.　2014; 9;e112569.

23. Huang SC, Yang YJ. Serum retinol-binding protein 4 is independently associated with pediatric NAFLD and fasting triglyceride level. J Pediatr Gastroenterol Nutr. 2013; 56:145-50.

24. Chan WK,　Nik Mustapha NR,　Mahadeva S. Controlled attenuation parameter for the detection and quantification of hepatic steatosis in nonalcoholic fatty liver disease. J Gastroenterol Hepatol.　2014;29:1470-6.

25. Desai NK, Harney S, Raza R, Al-Ibraheemi A, Shillingford N, Mitchell PD, et al. Comparison of Controlled Attenuation Parameter and Liver Biopsy to Assess Hepatic Steatosis in Pediatric Patients. J Pediatr.　2016;173:160-4.