Pediatrics

Heme oxygenase-1 gene promoter polymorphism and the risk of pediatric nonalcoholic fatty liver disease

Abstract

Background and objectives:

Oxidative stress and the insulin-resistant state are thought to be key components in the pathogenesis of pediatric nonalcoholic fatty liver disease (NAFLD). Heme oxygenase (HO) is important in the defense against oxidative stress. This study aimed to assess the association of HO-1 gene promoter polymorphism and insulin resistance with NAFLD among obese children.

Methods:

A total of 101 obese children aged 6–17 years were recruited. Anthropometric, serum biochemical variables and biomarkers for glucose and insulin metabolism were measured. We screened the allelic frequencies of (GT)n repeats in the HO-1 gene promoter among these obese children. NAFLD was determined through liver ultrasonography. Because the distribution of numbers of (GT)n repeats was bimodal, we divided the alleles into two classes: class S included shorter (27) repeats, and class L included longer (27) repeats. We assessed the effects of the length of (GT)n repeats in HO-1 gene promoter on pediatric NAFLD.

Results:

Of the 101 obese subjects, 27 (26.7%) had NAFLD. The alanine aminotransferase level was higher in patients carrying L alleles (L/L and L/S) than patients with S alleles (S/S) (46.2±49.3 IU|−1 versus 30.2±20.1 IU|−1; P=0.027). The significant risk factors for pediatric NAFLD were patients carrying L alleles (L/L and L/S) (odds ratio (OR)=18.84; 95% confidence interval (CI): 1.45–245.22; P=0.025), homeostasis model assessment of insulin resistance (OR=1.40; 95% CI: 1.07–1.83; P=0.014) and age (OR=1.24; 95% CI: 1.03–1.50; P=0.025).

Conclusion:

In this hospital-based study, the obese children with longer GT repeats in the HO-1 gene promoter and insulin resistance were susceptible to NAFLD.

Introduction

Obesity is strongly associated with the prevalence of nonalcoholic fatty liver disease (NAFLD) in pediatric populations. Concurrent with the epidemic of childhood obesity, pediatric NAFLD is a global health problem.1 NAFLD represents a spectrum of conditions ranging from simple hepatic steatosis to potentially fatal nonalcoholic steatohepatitis and cirrhosis.2 The pathogenesis of pediatric NAFLD is multifactorial. Not all obese children develop NAFLD, which suggests that environmental and genetic factors influence the susceptibility in each obese individual.3, 4, 5, 6 Oxidative stress and lipid peroxidation play an important role in the pathogenesis of NAFLD as their end products can induce hepatocellular injury and fibrogenesis.7, 8 Heme oxygenase-1 (HO-1) is a stress-responsive protein, which plays a fundamental role against the oxidative process.9 It cleaves pro-oxidant heme into equimolar amounts of carbon monoxide, biliverdin/bilirubin and free iron.10 These enzymatic reaction products have significant and useful biological properties, such as antioxidant, anti-inflammatory and anti-apoptotic activities.11, 12, 13, 14 The human HO-1 gene has been mapped to chromosome 22q12, and a (GT)n dinucleotide repeat has been identified in the proximal promoter region.15 The (GT)n repeat is highly polymorphic and modulates gene transcription by oxidant challenge.16 The previous studies have demonstrated that a longer (GT)n repeat exhibits lower transcriptional activity.17, 18 Data on the relationship between HO-1 gene promoter microsatellite polymorphism and pediatric NAFLD are limited. Insulin resistance is also reported to be the key pathogenic factor of NAFLD.8, 19 Insulin resistance leads to increased lipolysis and increased hepatic uptake of free fatty acids with increased hepatic triglyceride synthesis and accumulation. Herein we investigated the effects of the risk factors, including age, body mass index (BMI), waist circumference, insulin resistance and HO-1 gene promoter microsatellite polymorphism on pediatric NAFLD.

Materials and Methods

Study population

From February 2010 to September 2012, 101 obese children aged 6–17 years were enrolled from the pediatric clinic in the Far Eastern Memorial Hospital. Obesity was defined as BMI>95th percentile with age and gender adjustment according to the standard of the Department of Health in Taiwan.20 Those children with alcohol intake, hepatitis B, hepatitis C, autoimmune hepatitis, inborn errors of metabolism and Wilson’s disease were excluded. All the participants’ parents signed the informed consent. This study was approved by the research board of the Far Eastern Memorial Hospital, Taipei, Taiwan.

Data Collection

Demographic data

At the time of clinic visit, the data of each child’s age, gender, BMI, waist circumference and blood pressure were obtained. BMI was calculated as weight (kg) divided by height (m) squared. Standardized BMI z score values were not available for Taiwanese children. To adjust for the intrinsic effects of age and sex on BMI, we derived a variable, which we call 'adjusted BMI', by subtracting the age- and sex-specific median BMI value for Taiwanese children from each subject’s BMI value. Waist circumference was measured at the navel level at the end of a normal expiration.

Laboratory investigations

A blood sample was collected in the morning after fasting for 8 h. Biochemical measurements were performed, including total serum bilirubin, alanine aminotransferase (ALT), aspartate aminotransferase, γ-glutamyltransferase, fasting plasma glucose, triglyceride, total cholesterol and high-density lipoprotein cholesterol. ALT and aspartate aminotransferase were measured by Kinetic UV test on Hitachi Modular (Santa Clara, CA, USA). High-sensitivity C-reactive protein levels were measured with the use of an immunonephelometry assay (Dade Behring Inc., Newark, NJ, USA). Insulin levels were measured using chemiluminescence immunoassay (DPC, Los Angeles, CA, USA). Insulin resistance was measured by homeostasis model assessment of insulin resistance (HOMA-IR) and was calculated as follows: HOMA-IR=(fasting insulin [μUml−1])(fasting glucose [mmoll−1]/22.5).21 An oral glucose-tolerance test was performed with the administration of 1.75 g of glucose per kilogram of body weight (maximal dose, 75 g). A plasma glucose level was measured again 2 h after oral glucose-tolerance test (2-h PG).

Determination of length polymorphism of (GT)n repeats in HO-1 gene promoter

Total genomic DNA was isolated from whole blood cells using the blood DNA isolation kit (Maxim Biotech Inc., San Francisco, CA, USA).The 5′-flanking region containing (GT)n repeats of the HO-1 gene was amplified by PCR with a FAM-labeled sense primer, 5′-IndexTermAGAGCCTGCAGCT TCTCAGA-3′, and an antisense primer, 5′-IndexTermACAAAGTCTGGCCATAGGAC-3′, as described previously.17The PCR products were mixed together with a GenoType TAMRA DNA ladder (size range 50–500 bp; GibcoBRL, Grand Island, NY, USA) and analyzed with an automated DNA sequencer (ABI Prism 377). Each size of the (GT)n repeat was calculated using GeneScan Analysis software (PE Applied Biosystems, Foster City, CA, USA).

Liver ultrasonography

Ultrasonography is widely accepted as a surrogate method of liver biopsy and used as a tool for grading NAFLD.22, 23 All the study subjects underwent an ultrasonographic study of the liver performed by one operator. The machine (EnVisor C, PHILIPS, Bothell, WA, USA) was equipped with an 8–3.5 MHz curved linear array probe (C5040, PHILIPS). NAFLD was defined as the presence of an ultrasonographic pattern consistent with the following criteria: liver-kidney echo discrepancy, attenuated echo penetration, visibility of diaphragm and obscure hepatic vessel structures. The above-mentioned ultrasonographic pattern was scored as described by Chan et al.24

Because ultrasonography is much more accurate at detecting moderate-severe steatosis than mild steatosis,25 we set a score 4 as the diagnostic criterion for NAFLD to avoid possible false-positive results.3, 4, 5, 6

Statistical analysis

We analyzed the data with IBM SPSS statistics software, version 19.0 (Chicago, IL, USA). In a univariate analysis, the qualitative variables were expressed as proportions and compared using the Chi-squared test, whereas continuous variables were expressed as mean±s.d. and compared using the independent two-sample t test. P<0.05 or a 95% confidence interval (CI) for odds ratio (OR)1.0 was defined as statistically significant. Next, a multivariate analysis was conducted to identify predictive factors of pediatric NAFLD. To ensure the quality of the analytic results, we used standard model-fitting techniques for variable selection, goodness-of-fit assessment and regression diagnostics in our regression analyses. The best final regression model was identified manually by reducing the significance levels to 0.05 corresponding to the chosen a level.

Results

Of the 101 obese subjects, 27 (26.7%) had NAFLD. There were no significant differences between subjects with and without NAFLD in terms of BMI, blood pressures, serum cholesterol and triglycerides (Table 1). No subjects had splenomegaly. Values of aspartate aminotransferase, ALT, fasting plasma glucose, 2- h plasma glucose level and HOMAR-IR increased significantly with NAFLD (P<0.05) (Table 1). The GT repeat numbers ranged from 16 to 37, and (GT)23 and (GT)30 are the two most common alleles in our study population (Figure 1). The proportion of allele frequencies below or above 27 GT repeats was around 50%, so we divided the alleles into two classes according to the number of (GT)n repeats: class S included alleles with repeat number <27; class L included alleles with 27 GT repeats. The patients were then classified to be S/S, S/L or L/L genotype according to their alleles. Table 2 listed the proportions of allele and genotypic frequencies of the study population. The proportion of L allele frequency was significantly higher in patients with NAFLD (65.0%) than in those without NAFLD (49.0%) (chi-square test, P=0.03). The distribution of genotypic frequencies of patients without NAFLD was similar to that of subjects with NAFLD. However, NAFLD was found to have a significant interaction with genotypes: the proportions of S/S, S/L and L/L genotypes were 4%, 59% and 37%, respectively (Table 2). Patients with NAFLD who had the L/L and L/S genotype were significantly higher than those in carriers of the S allele (P=0.014). The ALT levels were higher in subjects carrying L alleles (L/L and L/S) than those with S alleles (S/S) (46.2±49.3 IU|−1 versus 30.2±20.1 IU|−1, P=0.027) (Table 3). With regard to the genotypic distribution, the OR of NAFLD was 8.98 (95% CI 1.14–70.77, P=0.037) for subjects carrying L alleles (L/L and L/S) by univariate analysis. Multivariate analysis of correlates of NAFLD was performed with the variables of age, gender, adjusted BMI, HOMA-IR and carrying L alleles (L/L and L/S); the presence of L allele (OR=18.84; 95% CI:1.45-245.22; P=0.025), HOMA-IR (OR=1.40; 95% CI:1.07-1.83; P=0.014) and age (OR=1.24; 95% CI:1.03-1.50; P=0.025) were independently associated with an increased risk for NAFLD (Table 4).

Table 1 Baseline anthropometric and metabolic characteristics of the subjects
Figure 1
figure1

Frequency distribution of number of (GT)n repeats in the patients without NAFLD (a, n=74, non-NAFLD patients), and in the patients with NAFLD (b, n=27, NAFLD patients).

Table 2 Distribution of HO-1 promoter genotypes and allele frequencies of all study population with or without NAFLD
Table 3 General characteristics of the study population stratified by HO-1 promoter genotypes
Table 4 Multiple logistic regression analysis of predictive factors for pediatric NAFLD

Discussion

In this study, the length polymorphism of GT repeat in HO-1 gene promoter is an independent risk factor for pediatric NAFLD. The (GT)n repeat in HO-1 gene is highly polymorphic that could modulate the gene transcription.26 The previous study demonstrated that the more (GT)n repeats in promoter region and the less transcription of HO-1 gene in rat aortic smooth muscle cells.17 Therefore, we thought that the large (GT)n repeat in the HO-1 gene promoter may reduce the induction of HO-1, thereby resulting in the development of NAFLD. An inverse relationship between the HO-1 activity and development of fatty liver was demonstrated in animal models.27 HO-1 is at low or undetectable level in hepatocytes and expresses mainly in Kupffer cells in general conditions, but it responds to external stimuli with a rapid transcriptional activation and expresses in both Kupffer cells and hepatocytes.28 HO-1 induction is considered to be an adaptive cellular response to survive exposure to environmental stresses. HO-1 may play vital roles in many aspects, such as suppression of oxidative stress, inflammation and cell proliferation, microcirculation improvement and regulation of cytokine expression and so on.29, 30, 31 The protective effects of HO-1 may be due to its antioxidative product biliverdin/bilirubin, which can scavenge peroxide radicals and inhibit lipid peroxidation.32 The induction of HO-1, an antioxidant defense enzyme, interrupted the progression of steatohepatitis by inducing an antioxidant pathway and suppressing inflammatory cytokines TNF-α, IL-6 and SOCS1.32 HO-1 prevents nonalcoholic steatohepatitis through suppressing hepatocyte apoptosis in mice.

HOMA-IR was another independent risk factor of NAFLD in this study. Insulin resistance is believed to represent a common pathogenic factor for NAFLD.33 A number of studies conducted in NAFLD patients have shown an impaired ability of insulin to suppress endogenous glucose production indicating the presence of hepatic insulin resistance.34, 35, 36 Otherwise, insulin resistance is thought to lead to excess fatty acid accumulation in the liver. NAFLD patients show a reduced insulin-mediated inhibition of lipolysis that results in increased flux of free fatty acids to the liver and in a blunted inhibition of fatty acid oxidation.37, 38 Free fatty acids are associated with the development of liver steatosis. Moreover, this promotes both a local proinflammatory state leading to progressive liver injury and the release of proinflammatory cytokines, which further exacerbates insulin resistance.39

Several limitations existed in this study. First, our study was limited to Taiwanese children and the sample size is small. It makes generalization to the other ethnic groups uncertain. Future studies should examine HO-1 genotypes among other ethnic groups to confirm the finding. Second, the diagnostic modality used in this study was liver ultrasonography, which is unable to stage hepatic fibrosis in NAFLD. Third, the present study design was cross-sectional, and we cannot conclude the causality. The reduced levels of antioxidants, such as HO-1, could reflect either consumption of antioxidants or a primary cause predisposing to oxidative stress. Our previous studies have reported that antioxidants such as UGT1A1 in counteracting oxidative stress are associated with NAFLD and genetic variants in GCKR and PNPLA3 confer susceptibility to NAFLD in obese individuals.3, 4, 5, 6 These studies provide a biological basis for NAFLD. It indicates that there may be many factors at work in developing pediatric NAFLD. In the future studies, we need to weigh all the factors and explore their interactions in the pathogenesis of NAFLD.

In summary, we have demonstrated that the polymorphism in the promoter of HO-1 gene and insulin resistance might be associated with the development of NAFLD among obese children and adolescents.

References

  1. 1

    Janssen I, Katzmarzyk PT, Boyce WF, Vereecken C, Mulvihill C, Roberts C et al. Comparison of overweight and obesity prevalence in school-aged youth from 34 countries and their relationships with physical activity and dietary patterns. Obes Rev 2005; 6: 123–132.

  2. 2

    Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ . Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116: 1413–1419.

  3. 3

    Lin YC, Chang PF, Hu FC, Chang MH, Ni YH . Variants in the UGT1A1 gene and the risk of pediatric non-alcoholic fatty liver disease. Pediatrics 2009; 124: 1221–1227.

  4. 4

    Lin YC, Chang PF, Hu FC, Yang WS, Chang MH, Ni YH . A common variant in the PNPLA3 gene is a risk factor of non-alcoholic fatty liver disease in Taiwanese obese children. J Pediatr 2011; 158: 740–744.

  5. 5

    Lin YC, Chang PF, Chang MH, Ni YH . A common variant in the peroxisome proliferator-activated receptor-γ coactivator-1α gene is associated with nonalcoholic fatty liver disease in obese children. Am J Clin Nutr 2013; 97: 326–331.

  6. 6

    Lin YC, Chang PF, Chang MH, Ni YH . Genetic variants in GCKR and PNPLA3 confer susceptibility to nonalcoholic fatty liver disease in obese individuals. Am J Clin Nutr 2014; 99: 869–874.

  7. 7

    Ludwig J, McGill DB, Lindor KD . Nonalcoholic steatohepatitis. J Gastroenterol Hepatol 1997; 12: 398–403.

  8. 8

    Sanyal AJ, Campbell-Sargent C, Mirshahi F, Rizzo WB, Contos MJ, Sterling RK et al. Non alcoholic steatohepatitis: association of insulin resistance and mitochondrial abnormalities. Gastroenterology 2001; 120: 1183–1192.

  9. 9

    Kawakami T, Takahashi T, Shimizu H, Nakahira K, Takeuchi M, Katayama H et al. Highly liver-specific heme oxygenase-1 induction by interleukin-11 prevents carbon tetrachlorideinduced hepatotoxicity. Int J Mol Med 2006; 18: 537–546.

  10. 10

    Tsui TY, Lau CK, Ma J, Wu X, Wang YQ, Farkas S et al. rAAV-mediated stable expression of heme oxygenase-1 in stellate cells: a new approach to attenuate liver fibrosis in rats. Hepatology 2005; 42: 335–342.

  11. 11

    Takahashi T, Shimizu H, Morimatsu H, Inoue K, Akagi R, Morita K et al. Heme oxygenase- 1 a fundamental guardian against oxidative tissue injuries in acute inflammation. Mini Rev Med Chem 2007; 7: 745–753.

  12. 12

    Ryter SW, Alam J, Choi AM . Heme oxygenase-1/carbon monoxide: from basic science to therapeutic applications. Physiol Rev 2006; 86: 583–650.

  13. 13

    Ryter SW, Kim HP, Nakahira K, Zuckerbraun BS, Morse D, Choi AM . Protective functions of heme oxygenase-1 and carbon monoxide in the respiratory system. Antioxid Redox Signal 2007; 9: 2157–2173.

  14. 14

    Ryter SW, Morse D, Choi AM . Carbon monoxide and bilirubin: potential therapies for pulmonary/vascular injury and disease. Am J Respir Cell Mol Biol 2007; 36: 175–182.

  15. 15

    Lavrovsky Y, Schwartzman MC, Levere RD, Kappas A, Abraham NG . Identification of binding sites for transcription factors NF-B and AP-2 in the promoter region of the human heme oxygenase-1 gene. Proc Natl Acad Sci USA 1994; 91: 5987–5991.

  16. 16

    Yamada N, Yamaya M, Okinaga S, Nakayama K, Sekizawa K, Shibahara S et al. Microsatellite polymorphism in the heme oxygenase-1 gene promoter is associated with susceptibility to emphysema. Am J Hum Genet 2000; 66: 187–195.

  17. 17

    Chen YH, Lin SJ, Lin MW, Tsai HL, Kuo SS, Chen JW et al. Microsatellite polymorphism in promoter of heme oxygenase-1 gene is associated with susceptibility to coronary artery disease in type 2 diabetic patients. Hum Genet 2002; 111: 1–8.

  18. 18

    Kaneda H, Ohno M, Taguchi J, Togo M, Hashimoto H, Ogasawara K et al. Heme oxygenase-1 gene promoter polymorphism is associated with coronary artery disease in Japanese patients with coronary risk factors. Arterioscler Thromb Vasc Biol 2002; 22: 1680–1685.

  19. 19

    Pagano G, Pacini G, Musso G, Gambino R, Mecca F, Depetris N et al. Nonalcoholic steatohepatitis, insulin resistance, and metabolic syndrome: further evidence for an etiologic association. Hepatology 2002; 35: 367–372.

  20. 20

    Chen W, Chang MH . New growth charts for Taiwanese children and adolescents based on World Health Organization standards and health-related physical fitness. Pediatr Neonatol 2010; 51: 69–79.

  21. 21

    Chitturi S, Abeygunasekera S, Farrell GC, Holmes-Walker J, Hui JM, Fung C et al. NASH and insulin resistance: insulin hypersecretion and specific association with the insulin resistance syndrome. Hepatology 2002; 35: 373–379.

  22. 22

    Manco M, Marcellini M, Devito R, Comparcola D, Sartorelli MR, Nobili V . Metabolic syndrome and liver histology in paediatric non-alcoholic steatohepatitis. Int J Obes 2008; 32: 381–387.

  23. 23

    Fu JF, Shi HB, Liu LR, Jiang P, Liang L, Wang CL et al. Non-alcoholic fatty liver disease: An early mediator predicting metabolic syndrome in obese children? World J Gastroenterol 2011; 17: 735–742.

  24. 24

    Chan DF, Li AM, Chu WC, Chan MH, Wong EM, Liu EK et al. Hepatic steatosis in obese Chinese children. Int J Obes Relat Metab Disord 2004; 28: 1257–1263.

  25. 25

    Hernaez R, Lazo M, Bonekamp S, Kamel I, Brancati FL, Guallar E et al. Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 2011; 54: 1082–1090.

  26. 26

    Naylor LH, Clark EM . d(TG)n.d(CA)n sequences upstream of the rat prolactin gene from Z-DNA and inhibit gene transcription. Nucleic Acids Res 1990; 18: 1595–1601.

  27. 27

    Hinds TD Jr, Sodhi K, Meadows C, Fedorova L, Puri N, Kim DH et al. Increased HO-1 levels ameliorate fatty liver development through a reduction of heme and recruitment of FGF21. Obesity 2014; 22: 705–712.

  28. 28

    Farombi EO, Surh YJ . Heme-oxygenase-1 as a potential therapeutic target for hepatoprotection. J Biochem Mol Biol 2006; 39: 479–491.

  29. 29

    Camara NO, Soares MP . Heme oxygenase-1 (HO-1), a protective gene that prevents chronic graft dysfunction. Free Radic Biol Med 2005; 38: 426–435.

  30. 30

    Neto JS, Nakao A, Kimizuka K, Romanosky AJ, Stolz DB, Uchiyama T et al. Protection of transplant-induced renal ischemia-reperfusion injury with carbon monoxide. Am J Physiol Renal Physiol 2004; 287: F979–789.

  31. 31

    Lickteig AJ, Fisher CD, Augustine LM, Cherrington NJ . Genes of the antioxidant response undergo upregulation in a rodent model of nonalcoholic steatohepatitis. J Biochem Mol Toxicol 2007; 21: 216–220.

  32. 32

    Yu J, Chu ES, Wang R, Wang S, Wu CW, Wong VW et al. Heme oxygenase-1 protects against steatohepatitis in both cultured hepatocytes and mice. Gastroenterology 2010; 138: 694–704.

  33. 33

    Utzschneider KM, Kahn SE . Review: The role of insulin resistance in nonalcoholic fatty liver disease. J Clin Endocrinol Metab 2006; 91: 4753–4761.

  34. 34

    Marchesini G, Brizi M, Bianchi G, Tomassetti S, Bugianesi E, Lenzi M et al. Nonalcoholic fatty liver disease: A feature of the metabolic syndrome. Diabetes 2001; 50: 1844–1850.

  35. 35

    Bugianesi E, Gastaldelli A, Vanni E, Gambino R, Cassader M, Baldi S et al. Insulin resistance in non-diabetic patients with non-alcoholic fatty liver disease: Sites and mechanisms. Diabetologia 2005; 48: 634–642.

  36. 36

    Seppälä-Lindroos A, Vehkavaara S, Häkkinen AM, Goto T, Westerbacka J, Sovijärvi A et al. Fat accumulation in the liver is associated with defects in insulin suppression of glucose production and serum free fatty acids independent of obesity in normal men. J Clin Endocrinol Metab 2002; 87: 3023–3028.

  37. 37

    Flegal KM, Carroll MD, Ogden CL, Johnson CL . Prevalence and trends in obesity among US adults, 1999–2000. JAMA 2002; 288: 1723–1727.

  38. 38

    Baldridge AD, Perez-Atayde AR, Graeme-Cook F, Higgins L, Lavine JE . Idiopathic steatohepatitis in childhood: A multicenter retrospective study. J Pediatr 1995; 127: 700–704.

  39. 39

    Ueki K, Kondo T, Tseng YH, Kahn CR . Central role of suppressors of cytokine signaling proteins in hepatic steatosis, insulin resistance, and the metabolic syndrome in the mouse. Proc Natl Acad Sci USA 2004; 101: 10422–10427.

Download references

Acknowledgements

We are indebted to Jyh-Feng Lu for laboratory work, and to Chien-Hao Chen for assistance with the statistical analysis. This work was supported by research grants from the Far Eastern Memorial Hospital (FEMH- 2011- C - 30 and FEMH- 2012- C- 040).

Author information

Correspondence to Y-H Ni.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chang, P., Lin, Y., Liu, K. et al. Heme oxygenase-1 gene promoter polymorphism and the risk of pediatric nonalcoholic fatty liver disease. Int J Obes 39, 1236–1240 (2015). https://doi.org/10.1038/ijo.2015.46

Download citation

Further reading