- Email: [email protected]
Prevalence, risk factors, and predictors of nonalcoholic fatty liver disease among schoolchildren: A hospital-based study in Alexandria, Egypt
Arab Journal of Gastroenterology 15 (2014) 76–81
Contents lists available at ScienceDirect
Arab Journal of Gastroenterology journal homepage: www.elsevier.com/locate/ajg
Original Article
Prevalence, risk factors, and predictors of nonalcoholic fatty liver disease among schoolchildren: A hospital-based study in Alexandria, Egypt Yasmine M. Alkassabany, Azza G. Farghaly, Engy M. El-Ghitany ⇑ Tropical Health Department, High Institute of Public Health, Alexandria University, Alexandria, Egypt
a r t i c l e
i n f o
Article history: Received 10 July 2013 Accepted 22 May 2014
Keywords: Fatty liver Children Risk factors Obesity Predictor Metabolic syndrome
a b s t r a c t Background and study aims: Nonalcoholic fatty liver disease (NAFLD) is an emerging problem in children and adolescents worldwide. This study was done to investigate the prevalence of NAFLD in children and adolescents as well as to determine the associated risk factors of fatty liver and to explore the ability of some obesity indices to predict and consequently be used as a screening method of fatty liver disease at certain cutoff points in schoolchildren. Patients and methods: A cross-sectional, nested case–control study was carried out. Cases and controls were randomly selected from outpatient schoolchildren aged 6–18 years attending the radiology clinic at Sporting Health Insurance Paediatric Hospital in Alexandria. They were subjected to ultrasonic examination as well as complete anthropometric and laboratory measurements including fasting plasma glucose (FPG) level, fasting insulin, alanine aminotransferase (ALT) level, and lipid profile. Results: Fatty liver was prevalent in schoolchildren (15.8%) and increased significantly with age (p = 0.004). Positive family history of diabetes mellitus (DM), hypertension (HTN), obesity, and liver disease were all statistically significant risk factors for fatty liver. Waist circumference (WC), body mass index (BMI) and its Z-score were significantly sensitive predictors. BMI was considered the best predictor of paediatric NAFLD at a cutoff = 22.9. NAFLD was significantly associated with high triglycerides (TGs), low high-density lipoprotein cholesterol (HDL), homoeostatic model assessment (HOMA) percentile, and the number of metabolic syndrome (MS) components. Conclusion: Paediatric NAFLD is a substantial problem in schoolchildren and has a close relationship with obesity, dyslipidaemia, insulin resistance (IR), and consequently MS. BMI and WC can be used as useful predictors and screening tools for NAFLD in schoolchildren. Ó 2014 Arab Journal of Gastroenterology. Published by Elsevier B.V. All rights reserved.
Introduction Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in developed countries and presents an emerging problem in children and adolescents worldwide [1,2]. It refers to a wide spectrum of liver diseases ranging from steatosis, when fat accumulates in the liver, nonalcoholic steatohepatitis (NASH), when fat in the liver causes liver inflammation, to cirrhosis, when chronic inflammation progresses to advanced scarring of the liver [3]. It is highly associated with obesity, which is becoming a serious public health problem globally, and in turn shifted the attention from steatosis in adults to children and adolescents in recent years [4]. It is also frequently associated with insulin ⇑ Corresponding author. Address: Tropical Health Department, High Institute of Public Health, Alexandria University 165 El-Horreya Avenue, Alexandria, Egypt. Mobile: +20 1001781333; fax: +20 3 5457037. E-mail address: [email protected] (E.M. El-Ghitany).
resistance (IR), type 2 diabetes mellitus (DM), and dyslipidaemia, which are the main components of metabolic syndrome (MS) [5–7]. Most of the children diagnosed to have NAFLD (83.3%) presented with at least one feature of MS, whereas overt MS (i.e., more than three features) was present in 28.8% of them [8]. Reports of paediatric fatty liver disease have increased over the past decades. However, the actual prevalence of NAFLD in children remains unknown because of the lack of population-based studies and the lack of reliable concrete screening methods [2]. Fatty liver prevalence differs according to the country, the method of screening, the age range, and the clinical characteristics of the target population. Some of the available population-based studies of paediatric NAFLD suggest a prevalence of 7.1%, 4.4%, and 3.2% in Iran, Japan, and Korea, respectively [9–11]. The problem of fatty liver in Egyptian children has not been sufficiently studied in terms of prevalence and risk factors. Studying the risk factors behind the development of fatty liver in children is crucial for the development of a primary prevention programme
http://dx.doi.org/10.1016/j.ajg.2014.05.002 1687-1979/Ó 2014 Arab Journal of Gastroenterology. Published by Elsevier B.V. All rights reserved.
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
aiming at modification of lifestyle in children at risk. Moreover, early detection of steatosis in children is mandatory for early intervention and prevention of further complications. The aim of the present study was to estimate the prevalence and determine the associated risk factors of fatty liver in the studied schoolchildren and to explore the ability of some obesity indices to predict fatty liver disease and consequently be used as a screening method for fatty liver disease at certain cutoff points in children.
Patients and methods This study was a hospital-based cross-sectional, nested case– control study. The study participants were selected by simple random sampling from outpatient schoolchildren aged 6–18 years attending the radiology clinic for screening of fatty liver at the Sporting Health Insurance Paediatric Hospital in Alexandria. Children who were diabetics, had malignancies, on immunosuppressive therapy, and/or had a known liver disease or hepatitis C virus (HCV) infection by history were excluded from the study. The study comprised 800 students who were screened by ultrasonography (US) for the presence of NAFLD after getting their informed assent and their parents’ written informed consent. Although liver biopsy is the gold standard for fatty liver diagnosis, its relatively high expense and invasiveness make it unsuitable for screening. Therefore, we used US for screening because of its low cost, noninvasiveness, safety, and absence of radiation exposure. We used a convex-type transducer of an ultrasound device with 3.5–5-MHz frequency. NAFLD was diagnosed according to the following features: [12,13] (a) the echo level of the liver in contrast to that of the kidney, (b) the clarity of the hepatic vessels, and (c) posterior attenuation and visibility of the diaphragm. Fatty liver was graded as follows. Grade 1 (mild): a slight diffuse increase in fine echoes in the hepatic parenchyma with normal visualisation of the diaphragm and intrahepatic vessel borders. Grade 2 (moderate): a moderate diffuse increase in fine echoes with slightly impaired visualisation of the intrahepatic vessels and diaphragm. Grade 3 (severe): a marked increase in fine echoes with poor or no visualisation of the intrahepatic vessel borders, diaphragm, and posterior portion of the right lobe of the liver. Age and sex were the only recorded data from the children during the initial screening. Among the US confirmed cases, only 75 children who agreed to continue in the study were further studied together with a similar number of age- and sex-matched controls who were free from fatty liver. Among the 75 fatty liver cases, 47 were mild and 28 were moderate but none were severe. A questionnaire was used to obtain information on sociodemographics, medical history, physical activities, and nutritional habits. Cases and controls were also subjected to complete anthropometric measurements and laboratory investigations. Weight (kg) and height (cm) were measured and body mass index (BMI) and Z-score were calculated using the Child and Teen BMI Calculator [14]. BMI was calculated using the following formula (BMI = weight (kg)/(height (m))2) [15]. The BMI-for-age percentile was used to interpret the BMI value which is age- and sex-specific for children and teens. The Centers for Disease Control and Prevention (CDC) BMI-for-age growth charts for girls and boys allow translation of a BMI number into a percentile for a child’s or teen’s sex and age [16]. The weight status category for the calculated BMI-for-age percentile was defined as follows: underweight (<5th percentile), normal weight (5th percentile to <85th percentile), overweight (85th to <95th percentile), and obese (P95th percentile). Waist circumference (WC; cm) was also measured at the umbilical level. Blood samples were collected from the 150 children in the morning after 12 hours of fasting using the universal sterile
77
precautions. Using the enzymatic colorimetric method, and measurement by spectrophotometry, plasma levels of fasting plasma glucose (FPG; Linear chemicals, Barcelona, Spain), alanine aminotransferase enzyme (ALT; Human, Wiesbaden, Germany), triglycerides (TG; Biosystems, Barcelona, Spain), high-density lipoprotein cholesterol (HDL; Biosystems, Barcelona, Spain), and cholesterol (Biosystems, Barcelona, Spain) were measured. Low-density lipoprotein cholesterol (LDL) level was calculated using the Friedewald (1972) formula [17]. Serum fasting insulin was measured using insulin ELISA Kit (LDN, Nordhorn, Germany). IR was calculated using the homeostasis model of insulin resistance (HOMA-IR) according to the following formula: [18] fasting Glucose (mg/dl) fasting insulin (lU/ml)/405. Children were diagnosed as having MS if they met at least three of the following criteria: (1) WC P90th percentile, (2) elevated blood pressures (BPs; systolic BP (SBP) and/or diastolic blood pressure (DBP) P90th percentile), (3) low HDL (<40 mg/dl) and/or high fasting serum TG (P120 mg/dl), and (4) high FPG (P100 mg/dl). Ethical consideration The study procedure strictly followed the international ethical consideration of the Helsinki Declaration. The study protocol was approved by the ethics committee of HIPH and students medical insurance sector in Alexandria. Statistical analysis Data entry and analyses were done using SPSS version 16. All statistical analyses were done using two-tailed tests and alpha error of 0.05 p value. Descriptive statistics were presented using frequencies as well as means and standard deviation to describe the categorical and numeric data, respectively. For comparing numeric parametric data, t-test was used. Mann–Whitney test was used to compare ranks (medians) for two independent groups of cases. Pearson’s chi-square test was used to test for the association between the categories of two independent samples. Monte Carlo exact and Fisher’s exact tests were used if there were many small expected values. Odds ratio (OR) and their confidence intervals (CIs) were used to assess the risk of NAFLD if a certain factor is present. Receiver Operating Characteristic (ROC) curve was used to test for the effect of some obesity indices (BMI, Z-score, and WC) on fatty liver and to have the optimal cutoff point (discriminating between groups) based on sensitivity and specificity for each selected point. This was created by plotting the fraction of true positives out of the positives (true positive rate, TPR) versus the fraction of false positives out of the negatives (FPR, false positive rate), at various threshold settings. Results Among the initially screened 800 students, 349 (43.6%) were boys and 451 (56.4%) were girls. NAFLD was diagnosed in 126 children giving a prevalence of 15.8%, 95% CI 13.2–18.3. Mild, moderate, and severe fatty liver represented 65.9%, 31.7%, and 2.4%, respectively of fatty liver cases. Prevalence of fatty liver increased significantly with age (p = 0.004) ranging from 8.7% for ages 6–10 years up to 20.4% for ages 15–18 years. Although there was no significant relation between gender and fatty liver (p = 0.561) in the overall population, a significant difference was noted in the oldest age group (15–18 years) where NAFLD was significantly associated with female gender (Fig. 1). Table 1 shows the basic demographic characteristics, medical as well as family history of the 150 participants in the case–control study. It is well shown that both groups were matched with regard
78
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
50.0
Male
45.0
Female
40.0 35.0
25.2
%
30.0 19.3
25.0 20.0
14.3 11.4
7.9
9.6
15.0 10.0 5.0 0.0 6-
10-
15- 18
Age
Fig. 1. The adjusted prevalence for age and sex among the studied sample. v2 test for difference within males = 5.5; p = 0.061. v2 test for difference within females = 14.2; p = 0.001.
to age, gender, and residence. No relation was found between birth weight or being full term and the development of NAFLD. Only two cases had DM versus none in the controls. A history of chronic disease including bronchial asthma and autoimmune disease was associated with a 27-fold more risk to develop fatty liver. Five cases versus none of the controls reported history of current cortisone therapy which increased the risk of NAFLD development (OR = 11.8, CI = 0.89–20.2). Positive family history of DM, hypertension (HTN), obesity, and liver disease were all statistically significant risk factors. Family history of obesity carried the highest risk (OR = 5.2). Obesity and overweight were distributed nearly similarly in both genders in children <15 years old. In older children, normalweight boys constituted 71.4% compared with 54.8% in girls, while obesity and overweight were higher in girls (23.8% and 21.4%, respectively) compared with 14.3% in each category in boys. Mean BMI and mean WC were not only significantly associated with fatty liver but also with its grade (Table 2). ROC curve analysis
of NAFLD according to the three obesity indices (BMI, Z-score, and WC) revealed that all of them were significant predictors at certain cutoff points with considerable areas under the curve, sensitivities, and specificities (Fig. 2). BMI was the most sensitive indicator at a cutoff of 22.9. The best discriminating (cutoff) points for other predictors were 80 for WC and 0.89 for Z-score. The relationships between fatty liver and laboratory findings are shown in Table 3. Fatty liver cases had significantly higher levels of lipid profile components. HDL levels were significantly of lower values in the cases. HOMA-IR value ranged in cases from 1.0 to 10.3 with a median of 2.8 compared with a significantly lower range of 0.01–8.1 and a median of 2.4 in the controls (p = 0.005). However, there was no significant difference between cases and controls regarding mean levels of FPG. Fig. 3 explores the relationship between IR and paediatric NAFLD. About half of the cases were above the 75th percentile compared to about one-third of the controls. Children with a HOMA level above the 75th percentile were significantly more at risk to develop NAFLD (OR = 2.9, CI = 1.0–9.0). The relationship between the clustering of MS components and the risk of NAFLD is illustrated in Table 4. All the cases have at least one component of the MS. The ORs for NAFLD increased with the clustering of the number of components (the linear trend was significant). The presence of three or more components indicated highest risk, and almost half of the cases had typical MS compared to 5.3% of the controls. Discussion The estimated prevalence of NAFLD in this study (15.8%) is higher than that reported by other studies, which also used US for screening because the patients are a selected group of hospital referrals. A similar research has been conducted in Iran resulting in 7.1% prevalence categorised as mild (84.1%), moderate (14.3%), and severe (1.6%) [9]. This difference in prevalence may be attributed to the fact that only 15.9% were obese in this study compared with
Table 1 Distribution of the fatty liver cases and controls according to sociodemographic characteristics, medical, and family history. Variable
Fatty liver cases n = 75
v2 (p)
Controls n = 75
OR (95% CI)
No.
%
No.
%
Age in years 6 to <10 10 to <15 15–18 Mean ± SD
14 37 24 12.9 ± 3.1
18.7 49.3 32.0
14 36 25 12.6 ± 3.1
18.7 48.0 33.3
Gender Male Female
21 54
28.0 72.0
20 55
26.7 73.3
0.05 (0.855)
1.1 (0.52–2.2)
Residence Urban Rural
71 4
94.7 5.3
68 7
90.7 9.3
0.88 (0.374)
1.8 (0.51–6.5)
Any other chronic disease Yes No
20 55
26.7 73.3
1 74
1.3 98.7
7.2 (0.005)
26.9 (3.5–120.6)2)
45
60.0
18
24.0
19.9 (0.000)
4.8 (2.4–9.6)
36
48.0
18
24.0
9.4 (0.002)
2.9 (1.5–5.9)
39
52.0
13
17.3
19.6 (0.000)
5.2 (2.4–10.9)
31
41.3
14
18.7
9.2 (0.007)
1.0 1.1 (0.53–2.2) 1.0 (0.42–2.6) t = 0.45 (0.654)
Family history of DM Family history of hypertension
Family history of obesity Family history of liver disease
*
p value based on Monte Carlo exact probability. p < 0.05 (significant).
*
3.1 (1.6–5.7)
79
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81 Table 2 Relation between anthropometric measurements and NAFLD. Anthropometric measurements Cases n = 75
Weight category Underweight Overweight Obese Normal BMI Mean ± SD Z-score Mean ± SD Waist circumference Mean ± SD
Controls n = 75
Statistical test (p) OR (95% CI)
No.
%
No.
%
0 18 46 11
0.0 24.0 61.3 14.7
2 2 2 69
2.7 v2 = 97.2 (0.000)* 2.7 2.7 92.0
Fatty liver grade
Statistical test (p)
Mild (n = 47)
Moderate (n = 28)
No.
%
No.
%
2.1 (0.14–4.2) 56.5 (11.5–110.3)* 144.3 (30.6–425.1)* 1.0
0 12 26 9
0.0 66.7 56.5 81.8
0 6 20 2
0.0 33.3 43.5 18.2
v2 = 2.6 (0.274)
27.6 ± 5.9
19.2 ± 3.4
t = 10.7 (0.000)*
1.6 (1.4–1.9)*
25.9 ± 4.6
30.4 ± 6.8
t = 3.1 (0.003*)
1.7 ± 0.7
1.8 ± 9.4
Z = 8.5 (0.000)*
0.99 (0.95–1.0)*
1.6 ± 0.8
1.9 ± 0.6
t = 2.2 (0.034*)
90.7 ± 15.6
68.6 ± 4.9
t = 8.9 (0.000)*
1.2 (1.1–1.3)*
87.3 ± 13.7
96.4 ± 17.0
t = 2.5 (0.014)*
Z: Mann–Whitney test. p < 0.05 (significant).
*
ROC Curve 1.0
Sensitivity
0.8
0.6
0.4
Source of the Curve BMI 0.2
Z-score Waist circumference Reference Line
0.0 0.0
0.2
0.6
0.4
0.8
1.0
1 - Specificity Diagonal segments are produced by ties.
Obesity index BMI Z-score WC
Area under the curve 0.911 0.901 0.879
Cutoff
Sensitivity
Specificity
22.9 0.89 80.0
82.7 89.3 80.0
92.0 88.0 86.7
Fig. 2. Receiver Operating Characteristic curves of BMI, Z-score, and WC.
33.5% in our study. Similarly, mild fatty liver constituted the majority of cases (65.9%), and severe fatty liver were of low frequency (2.4%), but we reported moderate fatty liver more frequently (31.7%), which may also be attributed to the higher obesity prevalence which we also found to be closely related to fatty liver grade. In Japan, Tominaga et al. [10] have reported that the NAFLD prevalence in 846 children aged 6–15 years old was 4.4%. Most of the studies that showed higher prevalence of NAFLD by US than that detected in this study were in overweight and/or obese children. In Germany, a higher percentage (28%) was reported among 532 obese children aged 8–19 years old [19]. A higher percentage (38%) was also reported by El-Karaksy et al. [20] among 76 obese Egyptian children aged 2–15 years. Several studies demonstrated that the prevalence of fatty liver is higher in adolescents than in younger children and that fatty
liver prevalence significantly increases with age [9,21,22]. This was also confirmed in the present study. Reported factors explaining the higher rate of NAFLD in adolescents include hormonal changes surrounding puberty, which may potentiate accumulation of fat in the liver [23], their increased tendency for unhealthy food consumption, and lower physical activity with age [24]. Most of the published reports indicated that NAFLD is more common in boys than in girls [10,21,25,26]. According to Schwimmer et al. [25], the reason for higher rates of NAFLD in males is because in boys, excess body fat is distributed in the intra-abdominal compartment and it is also due to the influence of sex hormones. In contrast to this finding, our study shows no significant correlation between gender and NAFLD. A similar finding was also reported by Alavian et al. [9] and Rafeey et al. [22] The female gender association was only noticed in the 15–18-year-old children, which can be explained by the fact that in this age group, obesity and overweight were more prevalent in girls (23.8% and 21.4%) than in boys (14.3% in each category), respectively, but a similar distribution was noted in the other age groups. Several studies confirmed that children with NAFLD have a higher BMI and that the prevalence of NAFLD was higher among overweight and obese children [9,10,20,27–30]. Findings in this study were consistent with other researchers as most of the cases were overweight and obese. BMI, WC, Z-score were significantly higher among cases suggesting that obesity is a very important risk factor. The previous finding matched the result of Sartorio et al. [31] who found that an increase of 1 standard deviation score of Z–BMI was associated with a fivefold increase in the odds of NAFLD. Our results also confirmed a significant relation between different obesity indices (BMI, WC) and fatty liver grade. This coincided with other studies [10,29]. In the ROC curve analysis of NAFLD in relation to obesity indices, the area under the curve showed a slight difference, but BMI appeared to be the most sensitive obesity index for detecting NAFLD among our studied children at a cutoff of 22.9. On the other hand, Tominaga et al. [10] found that WC is the most sensitive index at a cutoff of 74 (sensitivity = 81.8%). IR and hyperinsulinaemia are recognised as essential components in the development of NAFLD. Increased free fatty acid in hepatocytes leads to steatosis. Insulin inhibits oxidation of free fatty acids and thus hyperinsulinaemia may enhance free fatty acid hepatotoxicity [32]. In this study, HOMA-IR and fasting insulin are significantly associated with NAFLD. This comes in agreement with other studies [9,10,20,28,29,33,34]. Interestingly, our study showed that OR of NAFLD increased with increasing percentile of HOMA-IR which consequently confirms the strong association and linear relationship between IR and NAFLD. This finding matched the result of Tominaga et al. [10].
80
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
Table 3 Differences between cases and controls regarding laboratory findings. Lab results
FPG ALT TG
HDL LDL
Cholesterol
HOMA-IR
Cases n = 75
Normal (70–140 mg/dl) Elevated Normal (7–55 U/L) Elevated Normal (<90 mg/dl) Borderline high High Low Normal (40–60 mg/dl) Normal (<110 mg/dl) Borderline high High Normal (<170 mg/dl) Borderline high High Minimum Maximum Median Mean SD
v2 (p)
Controls n = 75
No.
%
No.
%
74 1 73 2 36 31 8 35 40 53 10 12 54 10 11 1.07 10.33 2.79 3.35 1.99
98.70 1.30 97.30 2.70 48.00 41.30 10.70 46.70 53.30 70.70 13.30 16.00 72.00 13.30 14.70 0.01 8.10 2.43 2.44 1.21
74 1 75 0 58 14 3 15 60 61 13 1 56 17 2
98.70 1.30 100.00 0.00 77.30 18.70 4.00 20.00 80.00 81.30 17.30 1.30 74.70 22.70 2.70
OR (95% CI)
(1.000!)
–
(0.497)
–
13.8 (0.001)*
1 3.7 (1.7–7.6)* 4.3 (1.1–17.3)* 3.5 (1.7–7.2)*
12.0 (0.001)* 10.3 (0.006)*
8.1 (0.018)*
Z = 2.8 (0.005)*
1 0.98 (0.34–2.2) 13.8 (1.7–89.7)* 1 0.61 (0.26–1.5) 5.7 (1.2–26.9)* (1.1–1.9)*
FPG: Fasting Plasma Glucose; ALT: Alanine aminotransferase enzyme; TG: Triglycerides (TG) (Biosystems, Spain, Barcelona); HDL: High-density lipoprotein cholesterol; LDL: Low-density lipoprotein cholesterol; HOMA-IR: Homoeostasis model of insulin resistance. * p < 0.05 (significant). p value based on Fishers exact probability.
Fig. 3. Distribution of HOMA-IR in NAFLD cases and controls.
Genetic, environmental, habitual, and lifestyle factors seem to play a role in fatty liver development. In the present study, a strong association was revealed between family history of DM, HTN, obesity, and liver disease and NAFLD development in children. The results of Schwimmer et al. [35] matched ours, in that family history of obesity was associated with NAFLD in children. NAFLD was observed in 13.1% of patients with bronchial asthma. It reached 17.6% in nonobese Japanese children with atopic dermatitis or suffering from bronchial asthma and/or allergic rhinitis [36]. Furthermore, Minakata et al. [37] reported that the prevalence of chronic obstructive pulmonary disease
(COPD) was significantly higher in patients with liver diseases (18.8%) particularly NAFLD (about 21.4%). The positive correlation between NAFLD and bronchial asthma may be attributed to the enhanced production of interleukin-8 (IL-8) in liver diseases, which also plays a role in the pathogenesis of COPD [37]. Moreover, bronchial asthma might be indirectly related to NAFLD through receiving corticosteroid therapy, which may be a predisposing factor for NAFLD [38]. The liver is a unique organ because of its capability in cholesterol metabolism and bile production. Cholesterol is excreted from the body in the bile. On the contrary, very low density lipoprotein (VLDL) is produced in the liver and intestine and transformed to intermediate-density lipoproteins and LDL which are taken up by the LDL receptors in the liver. Therefore, elevated non-HDL cholesterol could lead to the accumulation of lipids within the hepatocyte, and thus it is reasonable that non-HDL cholesterol might be a useful marker for NAFLD [29]. Several studies demonstrated that significant dyslipidaemia is a major finding encountered in NAFLD patients [9,20,28,29,34]. Findings in this study were mostly consistent with the previous studies in that hypertriglyceridemia and low HDL levels were significantly associated with NAFLD patients. As discussed earlier, all the major elements of MS, namely obesity, IR, dyslipidaemia, and family history of HTN, were significantly associated with fatty liver. All children with NAFLD had at least one feature of MS. Schwimmer et al. [39] reported the same finding. They reported that children with MS had five times the odds of having NAFLD as children without MS.
Table 4 Odds ratio for NAFLD according to the number of metabolic syndrome components. Number of metabolic syndrome components
0 3+ *
p < 0.05 (significant). p value based on Monte Carlo exact probability.
Cases n = 75
Controls n = 75
No.
%
No.
%
0 37
0.0 49.3
9 4
12.0 5.3
OR (95% C.I)
v2 (p)
1 158.3 (87.4–202.9)*
55.4 (0.000)
*
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
Additionally, our study showed that the risk of NAFLD increased linearly with clustering of the components of MS. Tominaga et al. [10] and Park et al. [11] had the same conclusion. Paediatric NAFLD is a substantial problem in Egypt. It has a close relationship with obesity, dyslipidaemia (high TG and low HDL particularly), IR, and consequently MS. Simple obesity indices particularly BMI can be used as a screening tool of NAFLD in schoolchildren. Urgent public health measures are needed to prevent and control NAFLD in Egyptian schoolchildren. Conflicts of interest The authors declared that there was no conflict of interest. Acknowledgment We thank all children and their parents who agreed to participate in this study. Many thanks and appreciation go to the personnel in the Radiology department at the Sporting Health Insurance Paediatric Hospital in Alexandria who offered help in accomplishing this work. References [1] De Bruyne RM, Fitzpatrick E, Dhawan A. Fatty liver disease in children: eat now pay later. Hepatol Int 2010;4(1):375–85. [2] Barshop NJ, Sirlin CB, Schwimmer JB, Lavine JE. Review article: Epidemiology, pathogenesis and potential treatments of paediatric non-alcoholic fatty liver disease. Aliment Pharmacol Ther 2008;28(1):13–24. [3] Papandreou D, Rousso I, Mavromichalis I. Update on non-alcoholic fatty liver disease in children. Clin Nutr 2007;26(4):409–15. [4] Kosti RI, Panagiotakos DB. The epidemic of obesity in children and adolescents in the world. Cent Eur J Public Health 2006;14(4):151–9. [5] Kelly D. Diseases of the liver and biliary system in children. 3rd ed. UK: Blackwell Publishing; 2008. [6] Alisi A, Manco M, Panera N, Nobili V. Association between type two diabetes and non-alcoholic fatty liver disease in youth. Ann Hepatol 2009;8(Suppl. 1):S44–50. [7] Souza MR, Diniz Mde F, Medeiros-Filho JE, Araújo MS. Metabolic syndrome and risk factors for non-alcoholic fatty liver disease. Arq Gastroenterol 2012;49(1): 89–96. [8] 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(11):1538–44. [9] Alavian SM, Mohammad-Alizadeh AH, Esna-Ashari F, Ardalan G, Hajarizadeh B. Non-alcoholic fatty liver disease prevalence among school-aged children and adolescents in Iran and its association with biochemical and anthropometric measures. Liver Int 2009;29(2):159–63. [10] Tominaga K, Fujimoto E, Suzuki K, Hayashi M, Ichikawa M, Inaba Y. Prevalence of non-alcoholic fatty liver disease in children and relationship to metabolic syndrome, insulin resistance, and waist circumference. Environ Health Prev Med 2009;14(2):142–9. [11] Park HS, Han JH, Choi KM, Kim SM. Relation between elevated serum alanine aminotransferase and metabolic syndrome in Korean adolescents. Am J Clin Nutr 2005;82(5):1046–51. [12] Roldan-Valadez E, Favila R, Martinez-Lopez M, Uribe M, Mendez-Sanchez N. Imaging techniques for assessing hepatic fat content in nonalcoholic fatty liver disease. Ann Hepatol 2008;7(3):212–20. [13] Ma X, Holalkere NS, Kambadakone RA, Mino-Kenudson M, Hahn PF, Sahani DV. Imaging-based quantification of hepatic fat: methods and clinical applications. Radiographics 2009;29(5):1253–77. [14] The children’s hospital of Philadelphia, c2008–12. Available from: http:// stokes.Chop.Edu/web/zscore/index.Php. [internet, cited 2012 Nov 10]
81
[15] Centers for disease control and prevention. Available from: http:// www.Cdc.Gov/healthyweight/assessing/bmi/childrens_BMI/childrens_BMI_ formula.html. [internet, cited 2012 Nov 10] [16] Centers for disease control and prevention. Available from: http:// apps.Nccd.Cdc.Gov/dnpabmi/ calculator. Aspx?Calculatortype=metric. [internet, cited 2012 Nov 10] [17] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18(6):499–502. [18] HOMA-IR Calculator. Available from: http://Www.Hcvsociety.Org/Files/ Homacalc.Htm. [internet] [19] Denzer C, Thiere D, Muche R, Koenig W, Mayer H, Kratzer W, et al. Genderspecific prevalences of fatty liver in obese children and adolescents: roles of body fat distribution, sex steroids, and insulin resistance. J Clin Endocrinol Metab 2009;94(10):3872–81. [20] El-Karaksy HM, El-Koofy NM, Anwar GM, El-Mougy FM, El-Hennawy A, Fahmy ME. Predictors of non-alcoholic fatty liver disease in obese and overweight Egyptian children: single center study. Saudi J Gastroenterol 2011;17(1):40–6. [21] Schwimmer JB, Deutsch R, Kahen T, et al. Prevalence of fatty liver in children and adolescents. Pediatrics 2006;118(4):1388–93. [22] Rafeey M, Mortazavi F, Mogaddasi N, et al. Fatty liver in children. Ther Clin Risk Manag 2009;5(2):371–4. [23] Roberts EA. Pediatric nonalcoholic fatty liver disease (NAFLD): a ‘‘growing’’ problem? J Hepatol 2007;46(6):1133–42. [24] Loomba R, Sirlin CB, Schwimmer JB, et al. Advances in pediatric nonalcoholic fatty liver disease. Hepatology 2009;50(4):1282–93. [25] Schwimmer JB, Mcgreal N, Deutsch R, Finegold MJ, Lavine JE. Influence of gender, race, and ethnicity on suspected fatty liver in obese adolescents. Pediatrics 2005;115(5):e561–5. [26] Tominaga K, Kurata JH, Chen YK, Fujimoto E, Miyagawa S, Abe I, et al. Prevalence of fatty liver in Japanese children and relationship to obesity. An epidemiological ultrasonographic survey. Dig Dis Sci 1995;40(9):2002–9. [27] Franzese A, Vajro P, Argenziano A, Puzziello A, Iannucci MP, Saviano MC, et al. Liver involvement in obese children. Ultrasonography and liver enzyme levels at diagnosis and during follow-up in an Italian population. Dig Dis Sci 1997;42(7):1428–32. [28] Arslan N, Buyukgebiz B, Ozturk Y, Cakmakci H. Fatty liver in obese children: Prevalence and correlation with anthropometric measurements and hyperlipidemia. Turk J Pediatr 2005;47(1):23–7. [29] Fu CC, Chen MC, Li YM, Liu TT, Wang LY. The risk factors for ultrasounddiagnosed non-alcoholic fatty liver disease among adolescents. Ann Acad Med Singapore 2009;38(1):15–7. [30] Damaso AR, Do Prado WL, De Piano A, Tock L, Caranti DA, Lofrano MC. Relationship between nonalcoholic fatty liver disease prevalence and visceral fat in obese adolescents. Dig Liver Dis 2008;40(2):132–9. [31] Sartorio A, Del Col A, Agosti F, Mazzilli G, Bellentani S, Tiribelli C, et al. Predictors of non-alcoholic fatty liver disease in obese children. Eur J Clin Nutr 2007;61(7):877–83. [32] Roberts EA. Non-alcoholic fatty liver disease (NAFLD) in children. Front Biosci 2005;10:2306–18. [33] Lin YC, Chang PF, Yeh SJ, Liu K, Chen HC. Risk factors for liver steatosis in obese children and adolescents. Pediatr Neonatol. 2010;51(3):149–54. [34] Yoo J, Lee S, Kim K, Yoo S, Sung E, Yim J. Relationship between insulin resistance and serum alanine aminotransferase as a surrogate of NAFLD (nonalcoholic fatty liver disease) in obese Korean children. Diabetes Res Clin Pract 2008;81(3):321–6. [35] Schwimmer JB, Celedon MA, Lavine JE, Salem R, Campbell N, Schork NJ, et al. Heritability of nonalcoholic fatty liver disease. Gastroenterology 2009;136(5): 1585–92. [36] Kimata H. Prevalence of fatty liver in non-obese Japanese children with atopic dermatitis. Indian Pediatr 2005;42(6):587–93. [37] Minakata Y, Sugiura H, Yamagata T, Yanagisawa S, Ichikawa T, Koarai A, et al. Prevalence of COPD in primary care clinics: correlation with non-respiratory diseases. Intern Med 2008;47(2):77–82. [38] Salt WB. Nonalcoholic fatty liver disease (NAFLD): a comprehensive review. J Insur Med 2004;36(1):27–41. [39] Schwimmer JB, Pardee PE, Lavine JE, Blumkin AK, Cook S. Cardiovascular risk factors and the metabolic syndrome in pediatric nonalcoholic fatty liver disease. Circulation 2008;118(3):277–83.
Contents lists available at ScienceDirect
Arab Journal of Gastroenterology journal homepage: www.elsevier.com/locate/ajg
Original Article
Prevalence, risk factors, and predictors of nonalcoholic fatty liver disease among schoolchildren: A hospital-based study in Alexandria, Egypt Yasmine M. Alkassabany, Azza G. Farghaly, Engy M. El-Ghitany ⇑ Tropical Health Department, High Institute of Public Health, Alexandria University, Alexandria, Egypt
a r t i c l e
i n f o
Article history: Received 10 July 2013 Accepted 22 May 2014
Keywords: Fatty liver Children Risk factors Obesity Predictor Metabolic syndrome
a b s t r a c t Background and study aims: Nonalcoholic fatty liver disease (NAFLD) is an emerging problem in children and adolescents worldwide. This study was done to investigate the prevalence of NAFLD in children and adolescents as well as to determine the associated risk factors of fatty liver and to explore the ability of some obesity indices to predict and consequently be used as a screening method of fatty liver disease at certain cutoff points in schoolchildren. Patients and methods: A cross-sectional, nested case–control study was carried out. Cases and controls were randomly selected from outpatient schoolchildren aged 6–18 years attending the radiology clinic at Sporting Health Insurance Paediatric Hospital in Alexandria. They were subjected to ultrasonic examination as well as complete anthropometric and laboratory measurements including fasting plasma glucose (FPG) level, fasting insulin, alanine aminotransferase (ALT) level, and lipid profile. Results: Fatty liver was prevalent in schoolchildren (15.8%) and increased significantly with age (p = 0.004). Positive family history of diabetes mellitus (DM), hypertension (HTN), obesity, and liver disease were all statistically significant risk factors for fatty liver. Waist circumference (WC), body mass index (BMI) and its Z-score were significantly sensitive predictors. BMI was considered the best predictor of paediatric NAFLD at a cutoff = 22.9. NAFLD was significantly associated with high triglycerides (TGs), low high-density lipoprotein cholesterol (HDL), homoeostatic model assessment (HOMA) percentile, and the number of metabolic syndrome (MS) components. Conclusion: Paediatric NAFLD is a substantial problem in schoolchildren and has a close relationship with obesity, dyslipidaemia, insulin resistance (IR), and consequently MS. BMI and WC can be used as useful predictors and screening tools for NAFLD in schoolchildren. Ó 2014 Arab Journal of Gastroenterology. Published by Elsevier B.V. All rights reserved.
Introduction Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in developed countries and presents an emerging problem in children and adolescents worldwide [1,2]. It refers to a wide spectrum of liver diseases ranging from steatosis, when fat accumulates in the liver, nonalcoholic steatohepatitis (NASH), when fat in the liver causes liver inflammation, to cirrhosis, when chronic inflammation progresses to advanced scarring of the liver [3]. It is highly associated with obesity, which is becoming a serious public health problem globally, and in turn shifted the attention from steatosis in adults to children and adolescents in recent years [4]. It is also frequently associated with insulin ⇑ Corresponding author. Address: Tropical Health Department, High Institute of Public Health, Alexandria University 165 El-Horreya Avenue, Alexandria, Egypt. Mobile: +20 1001781333; fax: +20 3 5457037. E-mail address: [email protected] (E.M. El-Ghitany).
resistance (IR), type 2 diabetes mellitus (DM), and dyslipidaemia, which are the main components of metabolic syndrome (MS) [5–7]. Most of the children diagnosed to have NAFLD (83.3%) presented with at least one feature of MS, whereas overt MS (i.e., more than three features) was present in 28.8% of them [8]. Reports of paediatric fatty liver disease have increased over the past decades. However, the actual prevalence of NAFLD in children remains unknown because of the lack of population-based studies and the lack of reliable concrete screening methods [2]. Fatty liver prevalence differs according to the country, the method of screening, the age range, and the clinical characteristics of the target population. Some of the available population-based studies of paediatric NAFLD suggest a prevalence of 7.1%, 4.4%, and 3.2% in Iran, Japan, and Korea, respectively [9–11]. The problem of fatty liver in Egyptian children has not been sufficiently studied in terms of prevalence and risk factors. Studying the risk factors behind the development of fatty liver in children is crucial for the development of a primary prevention programme
http://dx.doi.org/10.1016/j.ajg.2014.05.002 1687-1979/Ó 2014 Arab Journal of Gastroenterology. Published by Elsevier B.V. All rights reserved.
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
aiming at modification of lifestyle in children at risk. Moreover, early detection of steatosis in children is mandatory for early intervention and prevention of further complications. The aim of the present study was to estimate the prevalence and determine the associated risk factors of fatty liver in the studied schoolchildren and to explore the ability of some obesity indices to predict fatty liver disease and consequently be used as a screening method for fatty liver disease at certain cutoff points in children.
Patients and methods This study was a hospital-based cross-sectional, nested case– control study. The study participants were selected by simple random sampling from outpatient schoolchildren aged 6–18 years attending the radiology clinic for screening of fatty liver at the Sporting Health Insurance Paediatric Hospital in Alexandria. Children who were diabetics, had malignancies, on immunosuppressive therapy, and/or had a known liver disease or hepatitis C virus (HCV) infection by history were excluded from the study. The study comprised 800 students who were screened by ultrasonography (US) for the presence of NAFLD after getting their informed assent and their parents’ written informed consent. Although liver biopsy is the gold standard for fatty liver diagnosis, its relatively high expense and invasiveness make it unsuitable for screening. Therefore, we used US for screening because of its low cost, noninvasiveness, safety, and absence of radiation exposure. We used a convex-type transducer of an ultrasound device with 3.5–5-MHz frequency. NAFLD was diagnosed according to the following features: [12,13] (a) the echo level of the liver in contrast to that of the kidney, (b) the clarity of the hepatic vessels, and (c) posterior attenuation and visibility of the diaphragm. Fatty liver was graded as follows. Grade 1 (mild): a slight diffuse increase in fine echoes in the hepatic parenchyma with normal visualisation of the diaphragm and intrahepatic vessel borders. Grade 2 (moderate): a moderate diffuse increase in fine echoes with slightly impaired visualisation of the intrahepatic vessels and diaphragm. Grade 3 (severe): a marked increase in fine echoes with poor or no visualisation of the intrahepatic vessel borders, diaphragm, and posterior portion of the right lobe of the liver. Age and sex were the only recorded data from the children during the initial screening. Among the US confirmed cases, only 75 children who agreed to continue in the study were further studied together with a similar number of age- and sex-matched controls who were free from fatty liver. Among the 75 fatty liver cases, 47 were mild and 28 were moderate but none were severe. A questionnaire was used to obtain information on sociodemographics, medical history, physical activities, and nutritional habits. Cases and controls were also subjected to complete anthropometric measurements and laboratory investigations. Weight (kg) and height (cm) were measured and body mass index (BMI) and Z-score were calculated using the Child and Teen BMI Calculator [14]. BMI was calculated using the following formula (BMI = weight (kg)/(height (m))2) [15]. The BMI-for-age percentile was used to interpret the BMI value which is age- and sex-specific for children and teens. The Centers for Disease Control and Prevention (CDC) BMI-for-age growth charts for girls and boys allow translation of a BMI number into a percentile for a child’s or teen’s sex and age [16]. The weight status category for the calculated BMI-for-age percentile was defined as follows: underweight (<5th percentile), normal weight (5th percentile to <85th percentile), overweight (85th to <95th percentile), and obese (P95th percentile). Waist circumference (WC; cm) was also measured at the umbilical level. Blood samples were collected from the 150 children in the morning after 12 hours of fasting using the universal sterile
77
precautions. Using the enzymatic colorimetric method, and measurement by spectrophotometry, plasma levels of fasting plasma glucose (FPG; Linear chemicals, Barcelona, Spain), alanine aminotransferase enzyme (ALT; Human, Wiesbaden, Germany), triglycerides (TG; Biosystems, Barcelona, Spain), high-density lipoprotein cholesterol (HDL; Biosystems, Barcelona, Spain), and cholesterol (Biosystems, Barcelona, Spain) were measured. Low-density lipoprotein cholesterol (LDL) level was calculated using the Friedewald (1972) formula [17]. Serum fasting insulin was measured using insulin ELISA Kit (LDN, Nordhorn, Germany). IR was calculated using the homeostasis model of insulin resistance (HOMA-IR) according to the following formula: [18] fasting Glucose (mg/dl) fasting insulin (lU/ml)/405. Children were diagnosed as having MS if they met at least three of the following criteria: (1) WC P90th percentile, (2) elevated blood pressures (BPs; systolic BP (SBP) and/or diastolic blood pressure (DBP) P90th percentile), (3) low HDL (<40 mg/dl) and/or high fasting serum TG (P120 mg/dl), and (4) high FPG (P100 mg/dl). Ethical consideration The study procedure strictly followed the international ethical consideration of the Helsinki Declaration. The study protocol was approved by the ethics committee of HIPH and students medical insurance sector in Alexandria. Statistical analysis Data entry and analyses were done using SPSS version 16. All statistical analyses were done using two-tailed tests and alpha error of 0.05 p value. Descriptive statistics were presented using frequencies as well as means and standard deviation to describe the categorical and numeric data, respectively. For comparing numeric parametric data, t-test was used. Mann–Whitney test was used to compare ranks (medians) for two independent groups of cases. Pearson’s chi-square test was used to test for the association between the categories of two independent samples. Monte Carlo exact and Fisher’s exact tests were used if there were many small expected values. Odds ratio (OR) and their confidence intervals (CIs) were used to assess the risk of NAFLD if a certain factor is present. Receiver Operating Characteristic (ROC) curve was used to test for the effect of some obesity indices (BMI, Z-score, and WC) on fatty liver and to have the optimal cutoff point (discriminating between groups) based on sensitivity and specificity for each selected point. This was created by plotting the fraction of true positives out of the positives (true positive rate, TPR) versus the fraction of false positives out of the negatives (FPR, false positive rate), at various threshold settings. Results Among the initially screened 800 students, 349 (43.6%) were boys and 451 (56.4%) were girls. NAFLD was diagnosed in 126 children giving a prevalence of 15.8%, 95% CI 13.2–18.3. Mild, moderate, and severe fatty liver represented 65.9%, 31.7%, and 2.4%, respectively of fatty liver cases. Prevalence of fatty liver increased significantly with age (p = 0.004) ranging from 8.7% for ages 6–10 years up to 20.4% for ages 15–18 years. Although there was no significant relation between gender and fatty liver (p = 0.561) in the overall population, a significant difference was noted in the oldest age group (15–18 years) where NAFLD was significantly associated with female gender (Fig. 1). Table 1 shows the basic demographic characteristics, medical as well as family history of the 150 participants in the case–control study. It is well shown that both groups were matched with regard
78
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
50.0
Male
45.0
Female
40.0 35.0
25.2
%
30.0 19.3
25.0 20.0
14.3 11.4
7.9
9.6
15.0 10.0 5.0 0.0 6-
10-
15- 18
Age
Fig. 1. The adjusted prevalence for age and sex among the studied sample. v2 test for difference within males = 5.5; p = 0.061. v2 test for difference within females = 14.2; p = 0.001.
to age, gender, and residence. No relation was found between birth weight or being full term and the development of NAFLD. Only two cases had DM versus none in the controls. A history of chronic disease including bronchial asthma and autoimmune disease was associated with a 27-fold more risk to develop fatty liver. Five cases versus none of the controls reported history of current cortisone therapy which increased the risk of NAFLD development (OR = 11.8, CI = 0.89–20.2). Positive family history of DM, hypertension (HTN), obesity, and liver disease were all statistically significant risk factors. Family history of obesity carried the highest risk (OR = 5.2). Obesity and overweight were distributed nearly similarly in both genders in children <15 years old. In older children, normalweight boys constituted 71.4% compared with 54.8% in girls, while obesity and overweight were higher in girls (23.8% and 21.4%, respectively) compared with 14.3% in each category in boys. Mean BMI and mean WC were not only significantly associated with fatty liver but also with its grade (Table 2). ROC curve analysis
of NAFLD according to the three obesity indices (BMI, Z-score, and WC) revealed that all of them were significant predictors at certain cutoff points with considerable areas under the curve, sensitivities, and specificities (Fig. 2). BMI was the most sensitive indicator at a cutoff of 22.9. The best discriminating (cutoff) points for other predictors were 80 for WC and 0.89 for Z-score. The relationships between fatty liver and laboratory findings are shown in Table 3. Fatty liver cases had significantly higher levels of lipid profile components. HDL levels were significantly of lower values in the cases. HOMA-IR value ranged in cases from 1.0 to 10.3 with a median of 2.8 compared with a significantly lower range of 0.01–8.1 and a median of 2.4 in the controls (p = 0.005). However, there was no significant difference between cases and controls regarding mean levels of FPG. Fig. 3 explores the relationship between IR and paediatric NAFLD. About half of the cases were above the 75th percentile compared to about one-third of the controls. Children with a HOMA level above the 75th percentile were significantly more at risk to develop NAFLD (OR = 2.9, CI = 1.0–9.0). The relationship between the clustering of MS components and the risk of NAFLD is illustrated in Table 4. All the cases have at least one component of the MS. The ORs for NAFLD increased with the clustering of the number of components (the linear trend was significant). The presence of three or more components indicated highest risk, and almost half of the cases had typical MS compared to 5.3% of the controls. Discussion The estimated prevalence of NAFLD in this study (15.8%) is higher than that reported by other studies, which also used US for screening because the patients are a selected group of hospital referrals. A similar research has been conducted in Iran resulting in 7.1% prevalence categorised as mild (84.1%), moderate (14.3%), and severe (1.6%) [9]. This difference in prevalence may be attributed to the fact that only 15.9% were obese in this study compared with
Table 1 Distribution of the fatty liver cases and controls according to sociodemographic characteristics, medical, and family history. Variable
Fatty liver cases n = 75
v2 (p)
Controls n = 75
OR (95% CI)
No.
%
No.
%
Age in years 6 to <10 10 to <15 15–18 Mean ± SD
14 37 24 12.9 ± 3.1
18.7 49.3 32.0
14 36 25 12.6 ± 3.1
18.7 48.0 33.3
Gender Male Female
21 54
28.0 72.0
20 55
26.7 73.3
0.05 (0.855)
1.1 (0.52–2.2)
Residence Urban Rural
71 4
94.7 5.3
68 7
90.7 9.3
0.88 (0.374)
1.8 (0.51–6.5)
Any other chronic disease Yes No
20 55
26.7 73.3
1 74
1.3 98.7
7.2 (0.005)
26.9 (3.5–120.6)2)
45
60.0
18
24.0
19.9 (0.000)
4.8 (2.4–9.6)
36
48.0
18
24.0
9.4 (0.002)
2.9 (1.5–5.9)
39
52.0
13
17.3
19.6 (0.000)
5.2 (2.4–10.9)
31
41.3
14
18.7
9.2 (0.007)
1.0 1.1 (0.53–2.2) 1.0 (0.42–2.6) t = 0.45 (0.654)
Family history of DM Family history of hypertension
Family history of obesity Family history of liver disease
*
p value based on Monte Carlo exact probability. p < 0.05 (significant).
*
3.1 (1.6–5.7)
79
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81 Table 2 Relation between anthropometric measurements and NAFLD. Anthropometric measurements Cases n = 75
Weight category Underweight Overweight Obese Normal BMI Mean ± SD Z-score Mean ± SD Waist circumference Mean ± SD
Controls n = 75
Statistical test (p) OR (95% CI)
No.
%
No.
%
0 18 46 11
0.0 24.0 61.3 14.7
2 2 2 69
2.7 v2 = 97.2 (0.000)* 2.7 2.7 92.0
Fatty liver grade
Statistical test (p)
Mild (n = 47)
Moderate (n = 28)
No.
%
No.
%
2.1 (0.14–4.2) 56.5 (11.5–110.3)* 144.3 (30.6–425.1)* 1.0
0 12 26 9
0.0 66.7 56.5 81.8
0 6 20 2
0.0 33.3 43.5 18.2
v2 = 2.6 (0.274)
27.6 ± 5.9
19.2 ± 3.4
t = 10.7 (0.000)*
1.6 (1.4–1.9)*
25.9 ± 4.6
30.4 ± 6.8
t = 3.1 (0.003*)
1.7 ± 0.7
1.8 ± 9.4
Z = 8.5 (0.000)*
0.99 (0.95–1.0)*
1.6 ± 0.8
1.9 ± 0.6
t = 2.2 (0.034*)
90.7 ± 15.6
68.6 ± 4.9
t = 8.9 (0.000)*
1.2 (1.1–1.3)*
87.3 ± 13.7
96.4 ± 17.0
t = 2.5 (0.014)*
Z: Mann–Whitney test. p < 0.05 (significant).
*
ROC Curve 1.0
Sensitivity
0.8
0.6
0.4
Source of the Curve BMI 0.2
Z-score Waist circumference Reference Line
0.0 0.0
0.2
0.6
0.4
0.8
1.0
1 - Specificity Diagonal segments are produced by ties.
Obesity index BMI Z-score WC
Area under the curve 0.911 0.901 0.879
Cutoff
Sensitivity
Specificity
22.9 0.89 80.0
82.7 89.3 80.0
92.0 88.0 86.7
Fig. 2. Receiver Operating Characteristic curves of BMI, Z-score, and WC.
33.5% in our study. Similarly, mild fatty liver constituted the majority of cases (65.9%), and severe fatty liver were of low frequency (2.4%), but we reported moderate fatty liver more frequently (31.7%), which may also be attributed to the higher obesity prevalence which we also found to be closely related to fatty liver grade. In Japan, Tominaga et al. [10] have reported that the NAFLD prevalence in 846 children aged 6–15 years old was 4.4%. Most of the studies that showed higher prevalence of NAFLD by US than that detected in this study were in overweight and/or obese children. In Germany, a higher percentage (28%) was reported among 532 obese children aged 8–19 years old [19]. A higher percentage (38%) was also reported by El-Karaksy et al. [20] among 76 obese Egyptian children aged 2–15 years. Several studies demonstrated that the prevalence of fatty liver is higher in adolescents than in younger children and that fatty
liver prevalence significantly increases with age [9,21,22]. This was also confirmed in the present study. Reported factors explaining the higher rate of NAFLD in adolescents include hormonal changes surrounding puberty, which may potentiate accumulation of fat in the liver [23], their increased tendency for unhealthy food consumption, and lower physical activity with age [24]. Most of the published reports indicated that NAFLD is more common in boys than in girls [10,21,25,26]. According to Schwimmer et al. [25], the reason for higher rates of NAFLD in males is because in boys, excess body fat is distributed in the intra-abdominal compartment and it is also due to the influence of sex hormones. In contrast to this finding, our study shows no significant correlation between gender and NAFLD. A similar finding was also reported by Alavian et al. [9] and Rafeey et al. [22] The female gender association was only noticed in the 15–18-year-old children, which can be explained by the fact that in this age group, obesity and overweight were more prevalent in girls (23.8% and 21.4%) than in boys (14.3% in each category), respectively, but a similar distribution was noted in the other age groups. Several studies confirmed that children with NAFLD have a higher BMI and that the prevalence of NAFLD was higher among overweight and obese children [9,10,20,27–30]. Findings in this study were consistent with other researchers as most of the cases were overweight and obese. BMI, WC, Z-score were significantly higher among cases suggesting that obesity is a very important risk factor. The previous finding matched the result of Sartorio et al. [31] who found that an increase of 1 standard deviation score of Z–BMI was associated with a fivefold increase in the odds of NAFLD. Our results also confirmed a significant relation between different obesity indices (BMI, WC) and fatty liver grade. This coincided with other studies [10,29]. In the ROC curve analysis of NAFLD in relation to obesity indices, the area under the curve showed a slight difference, but BMI appeared to be the most sensitive obesity index for detecting NAFLD among our studied children at a cutoff of 22.9. On the other hand, Tominaga et al. [10] found that WC is the most sensitive index at a cutoff of 74 (sensitivity = 81.8%). IR and hyperinsulinaemia are recognised as essential components in the development of NAFLD. Increased free fatty acid in hepatocytes leads to steatosis. Insulin inhibits oxidation of free fatty acids and thus hyperinsulinaemia may enhance free fatty acid hepatotoxicity [32]. In this study, HOMA-IR and fasting insulin are significantly associated with NAFLD. This comes in agreement with other studies [9,10,20,28,29,33,34]. Interestingly, our study showed that OR of NAFLD increased with increasing percentile of HOMA-IR which consequently confirms the strong association and linear relationship between IR and NAFLD. This finding matched the result of Tominaga et al. [10].
80
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
Table 3 Differences between cases and controls regarding laboratory findings. Lab results
FPG ALT TG
HDL LDL
Cholesterol
HOMA-IR
Cases n = 75
Normal (70–140 mg/dl) Elevated Normal (7–55 U/L) Elevated Normal (<90 mg/dl) Borderline high High Low Normal (40–60 mg/dl) Normal (<110 mg/dl) Borderline high High Normal (<170 mg/dl) Borderline high High Minimum Maximum Median Mean SD
v2 (p)
Controls n = 75
No.
%
No.
%
74 1 73 2 36 31 8 35 40 53 10 12 54 10 11 1.07 10.33 2.79 3.35 1.99
98.70 1.30 97.30 2.70 48.00 41.30 10.70 46.70 53.30 70.70 13.30 16.00 72.00 13.30 14.70 0.01 8.10 2.43 2.44 1.21
74 1 75 0 58 14 3 15 60 61 13 1 56 17 2
98.70 1.30 100.00 0.00 77.30 18.70 4.00 20.00 80.00 81.30 17.30 1.30 74.70 22.70 2.70
OR (95% CI)
(1.000!)
–
(0.497)
–
13.8 (0.001)*
1 3.7 (1.7–7.6)* 4.3 (1.1–17.3)* 3.5 (1.7–7.2)*
12.0 (0.001)* 10.3 (0.006)*
8.1 (0.018)*
Z = 2.8 (0.005)*
1 0.98 (0.34–2.2) 13.8 (1.7–89.7)* 1 0.61 (0.26–1.5) 5.7 (1.2–26.9)* (1.1–1.9)*
FPG: Fasting Plasma Glucose; ALT: Alanine aminotransferase enzyme; TG: Triglycerides (TG) (Biosystems, Spain, Barcelona); HDL: High-density lipoprotein cholesterol; LDL: Low-density lipoprotein cholesterol; HOMA-IR: Homoeostasis model of insulin resistance. * p < 0.05 (significant). p value based on Fishers exact probability.
Fig. 3. Distribution of HOMA-IR in NAFLD cases and controls.
Genetic, environmental, habitual, and lifestyle factors seem to play a role in fatty liver development. In the present study, a strong association was revealed between family history of DM, HTN, obesity, and liver disease and NAFLD development in children. The results of Schwimmer et al. [35] matched ours, in that family history of obesity was associated with NAFLD in children. NAFLD was observed in 13.1% of patients with bronchial asthma. It reached 17.6% in nonobese Japanese children with atopic dermatitis or suffering from bronchial asthma and/or allergic rhinitis [36]. Furthermore, Minakata et al. [37] reported that the prevalence of chronic obstructive pulmonary disease
(COPD) was significantly higher in patients with liver diseases (18.8%) particularly NAFLD (about 21.4%). The positive correlation between NAFLD and bronchial asthma may be attributed to the enhanced production of interleukin-8 (IL-8) in liver diseases, which also plays a role in the pathogenesis of COPD [37]. Moreover, bronchial asthma might be indirectly related to NAFLD through receiving corticosteroid therapy, which may be a predisposing factor for NAFLD [38]. The liver is a unique organ because of its capability in cholesterol metabolism and bile production. Cholesterol is excreted from the body in the bile. On the contrary, very low density lipoprotein (VLDL) is produced in the liver and intestine and transformed to intermediate-density lipoproteins and LDL which are taken up by the LDL receptors in the liver. Therefore, elevated non-HDL cholesterol could lead to the accumulation of lipids within the hepatocyte, and thus it is reasonable that non-HDL cholesterol might be a useful marker for NAFLD [29]. Several studies demonstrated that significant dyslipidaemia is a major finding encountered in NAFLD patients [9,20,28,29,34]. Findings in this study were mostly consistent with the previous studies in that hypertriglyceridemia and low HDL levels were significantly associated with NAFLD patients. As discussed earlier, all the major elements of MS, namely obesity, IR, dyslipidaemia, and family history of HTN, were significantly associated with fatty liver. All children with NAFLD had at least one feature of MS. Schwimmer et al. [39] reported the same finding. They reported that children with MS had five times the odds of having NAFLD as children without MS.
Table 4 Odds ratio for NAFLD according to the number of metabolic syndrome components. Number of metabolic syndrome components
0 3+ *
p < 0.05 (significant). p value based on Monte Carlo exact probability.
Cases n = 75
Controls n = 75
No.
%
No.
%
0 37
0.0 49.3
9 4
12.0 5.3
OR (95% C.I)
v2 (p)
1 158.3 (87.4–202.9)*
55.4 (0.000)
*
Y.M. Alkassabany et al. / Arab Journal of Gastroenterology 15 (2014) 76–81
Additionally, our study showed that the risk of NAFLD increased linearly with clustering of the components of MS. Tominaga et al. [10] and Park et al. [11] had the same conclusion. Paediatric NAFLD is a substantial problem in Egypt. It has a close relationship with obesity, dyslipidaemia (high TG and low HDL particularly), IR, and consequently MS. Simple obesity indices particularly BMI can be used as a screening tool of NAFLD in schoolchildren. Urgent public health measures are needed to prevent and control NAFLD in Egyptian schoolchildren. Conflicts of interest The authors declared that there was no conflict of interest. Acknowledgment We thank all children and their parents who agreed to participate in this study. Many thanks and appreciation go to the personnel in the Radiology department at the Sporting Health Insurance Paediatric Hospital in Alexandria who offered help in accomplishing this work. References [1] De Bruyne RM, Fitzpatrick E, Dhawan A. Fatty liver disease in children: eat now pay later. Hepatol Int 2010;4(1):375–85. [2] Barshop NJ, Sirlin CB, Schwimmer JB, Lavine JE. Review article: Epidemiology, pathogenesis and potential treatments of paediatric non-alcoholic fatty liver disease. Aliment Pharmacol Ther 2008;28(1):13–24. [3] Papandreou D, Rousso I, Mavromichalis I. Update on non-alcoholic fatty liver disease in children. Clin Nutr 2007;26(4):409–15. [4] Kosti RI, Panagiotakos DB. The epidemic of obesity in children and adolescents in the world. Cent Eur J Public Health 2006;14(4):151–9. [5] Kelly D. Diseases of the liver and biliary system in children. 3rd ed. UK: Blackwell Publishing; 2008. [6] Alisi A, Manco M, Panera N, Nobili V. Association between type two diabetes and non-alcoholic fatty liver disease in youth. Ann Hepatol 2009;8(Suppl. 1):S44–50. [7] Souza MR, Diniz Mde F, Medeiros-Filho JE, Araújo MS. Metabolic syndrome and risk factors for non-alcoholic fatty liver disease. Arq Gastroenterol 2012;49(1): 89–96. [8] 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(11):1538–44. [9] Alavian SM, Mohammad-Alizadeh AH, Esna-Ashari F, Ardalan G, Hajarizadeh B. Non-alcoholic fatty liver disease prevalence among school-aged children and adolescents in Iran and its association with biochemical and anthropometric measures. Liver Int 2009;29(2):159–63. [10] Tominaga K, Fujimoto E, Suzuki K, Hayashi M, Ichikawa M, Inaba Y. Prevalence of non-alcoholic fatty liver disease in children and relationship to metabolic syndrome, insulin resistance, and waist circumference. Environ Health Prev Med 2009;14(2):142–9. [11] Park HS, Han JH, Choi KM, Kim SM. Relation between elevated serum alanine aminotransferase and metabolic syndrome in Korean adolescents. Am J Clin Nutr 2005;82(5):1046–51. [12] Roldan-Valadez E, Favila R, Martinez-Lopez M, Uribe M, Mendez-Sanchez N. Imaging techniques for assessing hepatic fat content in nonalcoholic fatty liver disease. Ann Hepatol 2008;7(3):212–20. [13] Ma X, Holalkere NS, Kambadakone RA, Mino-Kenudson M, Hahn PF, Sahani DV. Imaging-based quantification of hepatic fat: methods and clinical applications. Radiographics 2009;29(5):1253–77. [14] The children’s hospital of Philadelphia, c2008–12. Available from: http:// stokes.Chop.Edu/web/zscore/index.Php. [internet, cited 2012 Nov 10]
81
[15] Centers for disease control and prevention. Available from: http:// www.Cdc.Gov/healthyweight/assessing/bmi/childrens_BMI/childrens_BMI_ formula.html. [internet, cited 2012 Nov 10] [16] Centers for disease control and prevention. Available from: http:// apps.Nccd.Cdc.Gov/dnpabmi/ calculator. Aspx?Calculatortype=metric. [internet, cited 2012 Nov 10] [17] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18(6):499–502. [18] HOMA-IR Calculator. Available from: http://Www.Hcvsociety.Org/Files/ Homacalc.Htm. [internet] [19] Denzer C, Thiere D, Muche R, Koenig W, Mayer H, Kratzer W, et al. Genderspecific prevalences of fatty liver in obese children and adolescents: roles of body fat distribution, sex steroids, and insulin resistance. J Clin Endocrinol Metab 2009;94(10):3872–81. [20] El-Karaksy HM, El-Koofy NM, Anwar GM, El-Mougy FM, El-Hennawy A, Fahmy ME. Predictors of non-alcoholic fatty liver disease in obese and overweight Egyptian children: single center study. Saudi J Gastroenterol 2011;17(1):40–6. [21] Schwimmer JB, Deutsch R, Kahen T, et al. Prevalence of fatty liver in children and adolescents. Pediatrics 2006;118(4):1388–93. [22] Rafeey M, Mortazavi F, Mogaddasi N, et al. Fatty liver in children. Ther Clin Risk Manag 2009;5(2):371–4. [23] Roberts EA. Pediatric nonalcoholic fatty liver disease (NAFLD): a ‘‘growing’’ problem? J Hepatol 2007;46(6):1133–42. [24] Loomba R, Sirlin CB, Schwimmer JB, et al. Advances in pediatric nonalcoholic fatty liver disease. Hepatology 2009;50(4):1282–93. [25] Schwimmer JB, Mcgreal N, Deutsch R, Finegold MJ, Lavine JE. Influence of gender, race, and ethnicity on suspected fatty liver in obese adolescents. Pediatrics 2005;115(5):e561–5. [26] Tominaga K, Kurata JH, Chen YK, Fujimoto E, Miyagawa S, Abe I, et al. Prevalence of fatty liver in Japanese children and relationship to obesity. An epidemiological ultrasonographic survey. Dig Dis Sci 1995;40(9):2002–9. [27] Franzese A, Vajro P, Argenziano A, Puzziello A, Iannucci MP, Saviano MC, et al. Liver involvement in obese children. Ultrasonography and liver enzyme levels at diagnosis and during follow-up in an Italian population. Dig Dis Sci 1997;42(7):1428–32. [28] Arslan N, Buyukgebiz B, Ozturk Y, Cakmakci H. Fatty liver in obese children: Prevalence and correlation with anthropometric measurements and hyperlipidemia. Turk J Pediatr 2005;47(1):23–7. [29] Fu CC, Chen MC, Li YM, Liu TT, Wang LY. The risk factors for ultrasounddiagnosed non-alcoholic fatty liver disease among adolescents. Ann Acad Med Singapore 2009;38(1):15–7. [30] Damaso AR, Do Prado WL, De Piano A, Tock L, Caranti DA, Lofrano MC. Relationship between nonalcoholic fatty liver disease prevalence and visceral fat in obese adolescents. Dig Liver Dis 2008;40(2):132–9. [31] Sartorio A, Del Col A, Agosti F, Mazzilli G, Bellentani S, Tiribelli C, et al. Predictors of non-alcoholic fatty liver disease in obese children. Eur J Clin Nutr 2007;61(7):877–83. [32] Roberts EA. Non-alcoholic fatty liver disease (NAFLD) in children. Front Biosci 2005;10:2306–18. [33] Lin YC, Chang PF, Yeh SJ, Liu K, Chen HC. Risk factors for liver steatosis in obese children and adolescents. Pediatr Neonatol. 2010;51(3):149–54. [34] Yoo J, Lee S, Kim K, Yoo S, Sung E, Yim J. Relationship between insulin resistance and serum alanine aminotransferase as a surrogate of NAFLD (nonalcoholic fatty liver disease) in obese Korean children. Diabetes Res Clin Pract 2008;81(3):321–6. [35] Schwimmer JB, Celedon MA, Lavine JE, Salem R, Campbell N, Schork NJ, et al. Heritability of nonalcoholic fatty liver disease. Gastroenterology 2009;136(5): 1585–92. [36] Kimata H. Prevalence of fatty liver in non-obese Japanese children with atopic dermatitis. Indian Pediatr 2005;42(6):587–93. [37] Minakata Y, Sugiura H, Yamagata T, Yanagisawa S, Ichikawa T, Koarai A, et al. Prevalence of COPD in primary care clinics: correlation with non-respiratory diseases. Intern Med 2008;47(2):77–82. [38] Salt WB. Nonalcoholic fatty liver disease (NAFLD): a comprehensive review. J Insur Med 2004;36(1):27–41. [39] Schwimmer JB, Pardee PE, Lavine JE, Blumkin AK, Cook S. Cardiovascular risk factors and the metabolic syndrome in pediatric nonalcoholic fatty liver disease. Circulation 2008;118(3):277–83.