Prediction and validation of nonalcoholic fatty liver disease by fatty liver index in a Japanese population

Abstract

Fatty liver index (FLI) calculated by using body mass index (BMI), waist circumference and levels of γ-glutamyl transferase and triglycerides is a non-invasive predictor of nonalcoholic fatty liver disease (NAFLD). The original study in Italy showed that the cutoff level for prediction of NAFLD was FLI ≥60. However, the sex difference in FLI was not taken into consideration, and it is unclear whether the cutoff value can be applied to other races. We investigated the cutoff value of FLI for prediction of NAFLD determined by abdominal ultrasonography using receiver operating characteristic curve analyses in 14,471 Japanese subjects (men/women: 9,240/5,231; mean age: 48 ± 9 years). There was a significant interaction between sex and FLI for detection of NAFLD (p < 0.001). The cutoff values of FLI in men and women were 35.1 (area under the curve [AUC]: 0.82) and 15.6 (AUC: 0.91), respectively. When the subjects were divided by the absence and presence of obesity (BMI ≥25), there was a significant interaction between FLI and obesity for detection of NAFLD in women (p < 0.001) but not in men (p = 0.679). The cutoff values of FLI in non-obese/obese men and women were 22.6/52.6 and 11.2/33.2, respectively. In conclusion, the cutoff value of FLI for prediction of NAFLD in Japanese individuals was lower than that in the original study, and there is a significant sex difference. The simple and useful cutoff values in Japanese men and women are FLI ≥35 (non-obese/obese: 23/53) and FLI ≥16 (non-obese/obese: 11/33), respectively.

Nonalcoholic fatty liver disease (NAFLD) is a prevalent chronic liver disease and is closely related to obesity and lifestyle-related diseases [1, 2]. The frequency of NAFLD in adult subjects who received health examinations has been reported to be 9–30% in Japan [3] and has been increasing in recent years [4]. It has been shown that NAFLD is common even in lean subjects [5], and the prevalence of non-obese NAFLD patients with body mass index (BMI) <25 was approximately 15% in annual health checkup examinations in Japan [6]. NAFLD is a multisystem disease affecting extra-hepatic organs and regulatory pathways [7] in association with increased risks of insulin resistance [8], type 2 diabetes mellitus [9], hypertension [10], cardiovascular disease [11] and chronic kidney disease [12] as well as the development of liver cirrhosis [13] and liver cancer [14]. It is noteworthy that a new concept of metabolic dysfunction-associated fatty liver disease (MAFLD) has recently been proposed regardless of alcohol consumption [15].

In epidemiological studies, several non-invasive biochemical indicators are used for the diagnosis of NAFLD/MAFLD. Among the biomarkers, fatty liver index (FLI), which is calculated by using BMI, waist circumference (WC) and levels of γ-glutamyl transferase (γ-GTP) and triglycerides, was originally reported in Italy as an index for prediction of fatty liver detected by abdominal ultrasonography, and the cutoff value was reported to be FLI ≥60 [16]. The ability of FLI to predict fatty liver has been verified, and its usefulness has been reported in several countries [17-23] as well as in Japan [24] (Supplementary Table S1). However, the sex difference in FLI level was not taken into consideration in most of the studies [17-21, 24] despite the fact that there is a sex difference in the level of FLI [22, 23]. Furthermore, it is not clear whether categorization of FLI using FLI ≥60 is optimal because of the possibility of a racial difference in cutoff levels of FLI for diagnosis of NAFLD [16-24]. Since Asians have lower BMI and WC than those in other ethnic groups [25], their own FLI cutoff values for identification of fatty liver would be necessary.

It has been reported that the cutoff level of FLI for diagnosis of NAFLD seems to be lower in Asians than in Europeans: FLI ≥30 in China [18] and FLI ≥60 in Italy [16]. However, the subjects were not divided by sex in either of those studies. These findings indicate that there are racial and sex differences in cutoff levels of FLI for diagnosis of NAFLD. To the best of our knowledge, the sex difference in the FLI value for diagnosis of NAFLD was considered in only two studies performed in Taiwan. One study showed that the cutoff levels in men (n = 16,098) and women (n = 13,699) were FLI ≥35 and FLI ≥20, respectively [22]. The other study showed that the cutoff levels in men (n = 424) and women (n = 947) were FLI ≥20 and FLI ≥10, respectively [23]. Since the cutoff values have varied even in a same country, we investigated prediction and validation of NAFLD, which is defined by abdominal ultrasonography, exclusion of excess alcohol consumption and absence of viral hepatitis, by FLI in a large number of Japanese subjects divided by sex in the present study.

Methods

Study subjects

All of the subjects who received annual health examinations at Keijinkai Maruyama Clinic, Sapporo, Japan in 2006 were enrolled in this registry (n = 28,990) [26, 27]. A self-administered questionnaire survey was performed to obtain information on smoking habit, alcohol drinking habit and use of drugs for diabetes mellitus, hypertension and dyslipidemia. A flow chart of the study participants is shown in Fig. 1. Exclusion criteria were subjects with hepatitis B surface antigen positive, hepatitis C antibody positive and excess alcohol consumption (alcohol equivalent: ≥30 g/day for men, ≥20 g/day for women) and subjects with the absence of data for abdominal ultrasonography and components of FLI calculation including BMI, WC and levels of γ-GTP and triglycerides. After prespecified exclusion, a total of 14,471 Japanese subjects (men/women: 9,240/5,231; mean age: 48 ± 9 years) were finally recruited in the present study. The study conformed to the principles outlined in the Declaration of Helsinki and was performed with the approval of the institutional ethical committee of Sapporo Medical University (Number: 30-2-32). Written informed consent was obtained from all of the subjects.

Fig. 1

Flow chart of the selected study participants.

Among 28,990 subjects enrolled in 2006, a total of 15,965 subjects (male/female: 9,466/6,499) were finally included for analyses in the present study.

Measurements

Medical examinations, blood pressure measurements and samplings of urine and blood were performed after an overnight fast. Body height and weight were measured in light clothing without shoes, and BMI was calculated as body weight in kilograms divided by height in meters squared. FLI was calculated using the algorithm reported by Bedogni et al. [16]: FLI = [e^{(0.953 × ln(triglycerides) + 0.139 × BMI + 0.718 × ln(γ-GTP) + 0.053 × WC – 15.745)}]/[1 + e ^{(0.953 × ln(triglycerides) + 0.139 × BMI + 0.718 × ln(γ-GTP) + 0.053 × WC – 15.745)}] × 100. The FIB-4 index, a marker of hepatic fibrosis, was calculated by the following formula: age (years) × aspartate aminotransferase (AST; U/L))/(platelet count [10⁹/L] × alanine aminotransferase [ALT; U/L]^1/2) [28].

NAFLD was diagnosed by the presence of fatty liver determined by abdominal ultrasonography in the recruited subjects who had no findings of viral hepatitis and no excess alcohol consumption. Obesity was defined as BMI ≥25. Diabetes mellitus was diagnosed in accordance with the guideline of the American Diabetes Association [29]: fasting plasma glucose ≥126 mg/dL, hemoglobin A1c ≥6.5% or self-reported use of anti-diabetic drugs. Hypertension was diagnosed as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg or self-reported use of anti-hypertensive drugs. Dyslipidemia was diagnosed as low-density lipoprotein (LDL) cholesterol ≥140 mg/dL, high-density lipoprotein (HDL) cholesterol <40 mg/dL, triglycerides ≥150 mg/dL or self-reported use of anti-dyslipidemic drugs.

Abdominal ultrasonography

Abdominal ultrasonography was performed using SSA-250A or SSA-340A (Toshiba Medical, Otawara, Japan) by 10 well-experienced echographers with at least 5 years of experience who were trained by gastroenterologists with more than 5 years of experience. Fatty liver was determined by any findings of high-intensity bright liver, hepato-renal contrast, vascular obscuration and deep attenuation in the liver [30, 31]. The images and the presence of hepatic steatosis were independently reviewed by certified gastroenterologists who were blinded to clinical data.

Statistical analysis

Numeric variables are expressed as means ± standard deviation (SD) for parameters with normal distributions and as medians (interquartile ranges) for parameters with skewed distributions. The distribution of each parameter was tested for its normality using the Shapiro-Wilk W test. Comparison between two groups was done with Mann-Whitney’s U test. Intergroup differences of demographic parameters were tested by the chi-square test. Logistic regression analysis after adjustment of sex was performed to identify subjects with NAFLD using FLI, and the interaction between sex and FLI for the presence of NAFLD was also investigated. The ability of FLI, BMI, WC, triglycerides, γ-GTP, AST, ALT or FIB-4 index to predict NAFLD was investigated by using receiver operating characteristic (ROC) curve analyses. The area under the curve (AUC) and 95% confidence interval (CI) were calculated, and the cutoff values were obtained by the Youden’s index [32]. Cutoff values were also investigated when subjects were divided by the presence and absence of obesity. A p value <0.05 was considered statistically significant. All data were analyzed by using EZR [33] and R3.6.2.

Results

Characteristics of the study subjects

Basal characteristics of the recruited subjects are shown in Table 1. NAFLD was found in 5,166 subjects (35.7%, men/women: 4,322 [46.8%]/844 [16.1%]). Components of FLI calculation, including BMI, WC, γ-GTP and triglycerides, were significantly higher in men than in women. FLI level was significantly higher in men (median [interquartile ranges]: 36 [18-60]) than in women (7 [4-19]). FIB-4 index was significantly higher in men than in women. The frequency of current smoking habit was higher in men than in women.

Table 1 Characteristics of recruited subjects (n = 14,471)

	All (n = 14,471)	Men (n = 9,240)	Women (n = 5,231)	p
Age (years)	48 ± 9	49 ± 9	47 ± 9	<0.001
Body mass index	23 ± 3	24 ± 3	22 ± 3	<0.001
Waist circumference (cm)	84 ± 9	86 ± 9	79 ± 9	<0.001
Systolic blood pressure (mmHg)	117 ± 17	120 ± 16	111 ± 16	<0.001
Diastolic blood pressure (mmHg)	75 ± 11	77 ± 11	70 ± 11	<0.001
Pulse rate (beats/min)	63 ± 9	63 ± 9	64 ± 9	<0.001
Current smoking habit	4,729 (34.0)	3,842 (43.5)	887 (17.5)	<0.001
Comorbidity
NAFLD	5,166 (35.7)	4,322 (46.8)	844 (16.1)	<0.001
Obesity*	5,342 (36.9)	4,273 (46.2)	1,069 (20.4)	<0.001
Hypertension	2,631 (18.2)	2,040 (22.1)	591 (11.3)	<0.001
Diabetes mellitus	844 (5.8)	742 (8.0)	102 (1.9)	<0.001
Dyslipidemia	3,558 (24.6)	2,246 (24.3)	1,312 (25.1)	0.306
Biochemical data
Hemoglobin (g/dL)	14.3 ± 1.6	15.1 ± 1.1	12.9 ± 1.2	<0.001
Platelet (104/μL)	23.8 ± 5.2	23.3 ± 5.0	24.6 ± 5.5	<0.001
Albumin (g/dL)	4.4 ± 0.2	4.4 ± 0.2	4.3 ± 0.2	<0.001
Blood urea nitrogen (mg/dL)	14.2 ± 3.5	14.7 ± 3.5	13.4 ± 3.4	<0.001
Creatinine (mg/dL)	0.73 ± 0.24	0.81 ± 0.24	0.60 ± 0.18	<0.001
eGFR (mL/min/1.73 m2)	84.6 ± 14.9	83.3 ± 14.5	86.9 ± 15.4	<0.001
Uric acid (mg/dL)	5.5 ± 1.4	6.1 ± 1.2	4.4 ± 1.0	<0.001
AST (U/L)	21 (18–26)	22 (19–28)	19 (16–22)	<0.001
ALT (U/L)	21 (15–31)	25 (18–36)	15 (12–20)	<0.001
γ-GTP (U/L)	31 (19–57)	42 (27–73)	18 (14–27)	<0.001
FPG (mg/dL)	93 ± 20	97 ± 22	88 ± 14	<0.001
Hemoglobin A1c (%)	5.4 ± 0.7	5.4 ± 0.8	5.2 ± 0.5	<0.001
Total cholesterol (mg/dL)	206 ± 34	206 ± 34	205 ± 34	0.006
LDL cholesterol (mg/dL)	125 ± 33	127 ± 33	120 ± 32	<0.001
HDL cholesterol (mg/dL)	60 ± 16	55 ± 14	69 ± 15	<0.001
Triglycerides (mg/dL)	93 (63–138)	111 (79–161)	67 (50–94)	<0.001
FIB-4 Index	0.94 (0.73–1.22)	0.95 (0.73–1.24)	0.93 (0.72–1.19)	<0.001
FLI	23 (8–50)	36 (18–60)	7 (4–19)	<0.001

Variables are expressed as number (%), means ± SD or medians (interquartile ranges).

* Body mass index ≥25. ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomerular filtration rate; FLI, fatty liver index; FPG, fasting plasma glucose; γ-GTP, γ-glutamyl transpeptidase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NAFLD, nonalcoholic fatty liver disease.

Basal characteristics of the included subjects with and without NAFLD are shown in Table 2. Both men and women with NAFLD were older and had significantly larger BMI and WC, higher systolic and diastolic blood pressures and pulse rate, higher frequencies of obesity, hypertension, diabetes mellitus, and dyslipidemia and higher levels of hemoglobin, platelet count, albumin, uric acid, ALT, γ-GTP, fasting plasma glucose, hemoglobin A1c, total cholesterol, LDL cholesterol, triglycerides, and FLI and lower levels of eGFR, HDL cholesterol and FIB-4 index than did those without NAFLD.

Table 2 Characteristics of recruited subjects with and without NAFLD (n = 14,471)

	Men (n = 9,240)		p	Women (n = 5,231)		p
	NAFLD (–) (n = 4,918)	NAFLD (+) (n = 4,322)		NAFLD (–) (n = 4,387)	NAFLD (+) (n = 844)
Age (years)	48 ± 10	50 ± 9	<0.001	47 ± 9	52 ± 9	<0.001
Body mass index	23 ± 2	26 ± 3	<0.001	21 ± 3	26 ± 4	<0.001
Waist circumference (cm)	82 ± 7	91 ± 8	<0.001	77 ± 8	90 ± 9	<0.001
Systolic blood pressure (mmHg)	118 ± 15	123 ± 16	<0.001	109 ± 16	120 ± 17	<0.001
Diastolic blood pressure (mmHg)	55 ± 9	77 ± 11	<0.001	69 ± 10	76 ± 10	<0.001
Pulse rate (beats/min)	62 ± 9	64 ± 9	<0.001	63 ± 9	66 ± 9	<0.001
Current smoking habit	2,161 (46.1)	1,681 (40.6)	<0.001	751 (17.6)	136 (16.8)	0.579
Comorbidity
Obesity*	1,260 (25.6)	3,013 (69.7)	<0.001	496 (11.3)	573 (67.9)	<0.001
Hypertension	797 (16.2)	1,243 (28.8)	<0.001	354 (8.1)	237 (28.1)	<0.001
Diabetes mellitus	202 (4.1)	540 (12.5)	<0.001	37 (0.8)	65 (7.7)	<0.001
Dyslipidemia	969 (19.7)	1,277 (29.5)	<0.001	1,038 (23.7)	274 (32.5)	<0.001
Biochemical data
Hemoglobin (g/dL)	15.0 ± 1.0	15.3 ± 1.1	<0.001	12.8 ± 1.2	13.4 ± 1.1	<0.001
Platelet (104/μL)	23.2 ± 4.8	23.4 ± 5.2	0.017	24.3 ± 5.3	26.4 ± 6.0	<0.001
Albumin (g/dL)	4.4 ± 0.2	4.5 ± 0.2	<0.001	4.3 ± 0.2	4.3 ± 0.2	0.007
Blood urea nitrogen (mg/dL)	14.7 ± 3.6	14.8 ± 3.4	0.249	13.3 ± 3.4	13.9 ± 3.6	<0.001
Creatinine (mg/dL)	0.81 ± 0.31	0.81 ± 0.14	0.612	0.60 ± 0.19	0.59 ± 0.10	0.126
eGFR (mL/min/1.73 m2)	83.8 ± 14.0	82.8 ± 14.9	0.001	87.1 ± 15.4	85.8 ± 15.7	0.029
Uric acid (mg/dL)	5.9 ± 1.2	6.3 ± 1.3	<0.001	4.3 ± 0.9	5.0 ± 1.0	<0.001
AST (U/L)	21 (18–25)	25 (20–32)	<0.001	22 (18–27)	19 (16–22)	<0.001
ALT (U/L)	21 (16–27)	33 (24–49)	<0.001	14 (12–18)	23 (17–34)	<0.001
γ-GTP (U/L)	34 (23–57)	53 (34–88)	<0.001	17 (14–24)	28 (20–44)	<0.001
FPG (mg/dL)	92 ± 16	102 ± 27	<0.001	86 ± 12	96 ± 20	<0.001
Hemoglobin A1c (%)	5.2 ± 0.6	5.6 ± 1.0	<0.001	5.2 ± 0.5	5.6 ± 0.8	<0.001
Total cholesterol (mg/dL)	201 ± 33	213 ± 35	<0.001	202 ± 34	218 ± 35	<0.001
LDL cholesterol (mg/dL)	122 ± 31	134 ± 33	<0.001	117 ± 31	138 ± 32	<0.001
HDL cholesterol (mg/dL)	59 ± 15	52 ± 12	<0.001	71 ± 15	59 ± 13	<0.001
Triglycerides (mg/dL)	94 (68–131)	136 (99–192)	<0.001	62 (47–85)	107 (77–146)	<0.001
FIB-4 Index	0.95 (0.74–1.26)	0.94 (0.72–1.23)	0.046	0.93 (0.73–1.20)	0.92 (0.70–1.16)	0.019
FLI	22 (11–38)	56 (36–77)	<0.011	6 (3–12)	38 (21–60)	<0.001

Variables are expressed as number (%), means ± SD or medians (interquartile ranges).

* Body mass index ≥25. ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomerular filtration rate; FLI, fatty liver index; FPG, fasting plasma glucose; γ-GTP, γ-glutamyl transpeptidase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NAFLD, nonalcoholic fatty liver disease.

Prediction and validation of NAFLD by biomarkers

Logistic regression analysis after adjustment of sex showed that FLI was an independent determinant of NAFLD, and there was a significant interaction between FLI and sex for the presence of NAFLD (p < 0.001). Therefore, further analyses for prediction of NAFLD using FLI as well as its related parameters including BMI, WC, triglycerides, γ-GTP, ALT, AST and FIB-4 index were investigated in subjects divided by sex.

ROC curve analyses showed that the cutoff values of FLI for prediction of NAFLD in men (Fig. 2A) and women (Fig. 2B) were 35.1 (sensitivity/specificity: 76.7%/71.3%; AUC [95% CI]: 0.82 [0.81–0.83]; positive predictive value (PPV) [95% CI]: 70.1% [68.8–71.4]; negative predictive value (NPV) [95% CI]: 77.7 [76.4–78.9]; positive likelihood ratio (LR+) [95% CI]: 2.67 [2.55–2.80]; negative likelihood value (LR–) [95% CI]: 0.33 [0.31–0.35]) and 15.6 (sensitivity/specificity: 85.2%/81.4%; AUC [95% CI]: 0.91 [0.90–0.92]; PPV [95% CI]: 46.8 [44.3–49.4]; NPV [95% CI]: 96.6 [96.0–97.2]; LR+ [95% CI]: 4.58 [4.28–4.90]; LR– [95% CI]: 0.182 [0.16–0.21]), respectively. The diagnostic accuracy of FLI to predict NAFLD by 5-value intervals of FLI is shown in Supplementary Table S2. When values of FLI in men and women were 35 and 15, respectively, levels of Youden’s index were highest in both sexes.

Fig. 2

Prediction of fatty liver by FLI.

A, B. Receiver operating characteristic (ROC) curves of fatty liver index (FLI) to predict nonalcoholic fatty liver disease determined by abdominal ultrasonography in men (A) and women (B). AUC, area under the curve; CI, confidence interval.

Cutoff values of components of FLI calculation, including BMI, WC, triglycerides and γ-GTP, as well as the related markers, including AST, ALT and FIB-4 index, are shown in Table 3. The AUCs of FLI in both men and women were the highest among those of the biomarkers.

Table 3 Cutoff values of parameters for prediction of NAFLD

	Men (n = 9,240)					Women (n = 5,231)
	Cutoff	AUC	95% CI	Sensitivity (%)	Specificity (%)	Cutoff	AUC	95% CI	Sensitivity (%)	Specificity (%)
FLI	35.1	0.82	0.81–0.83	76.7	71.3	15.6	0.91	0.90–0.92	85.2	81.4
Body mass index	25	0.80	0.79–0.81	63.9	79.5	24	0.88	0.87–0.89	75.6	85.1
Waist circumference (cm)	86.8	0.80	0.79–0.81	70.0	75.0	83.2	0.87	0.86–0.88	79.1	79.3
Triglycerides (mg/dL)	111	0.71	0.70–0.72	67.1	63.6	76	0.79	0.77–0.80	77.0	66.3
γ-GTP (U/L)	41	0.67	0.66–0.68	66.1	60.3	20	0.75	0.73–0.77	78.1	62.3
AST (U/L)	24	0.68	0.67–0.69	57.4	68.8	22	0.69	0.67–0.71	52.8	75.7
ALT (U/L)	27	0.77	0.76–0.78	66.5	73.1	19	0.81	0.79–0.82	70.7	75.0
FIB-4 Index	0.69	0.51	0.50–0.52	22.3	80.3	0.64	0.53	0.50–0.55	86.2	20.0

ALT, alanine aminotransferase; AST, aspartate aminotransferase; AUC, area under the curve; CI, confidence interval; FLI, fatty liver index; γ-GTP, γ-glutamyl transpeptidase; NAFLD, nonalcoholic fatty liver disease.

Prediction and validation of NAFLD by FLI in non-obese and obese subjects

As subgroup analyses, the subjects were divided by the absence and presence of obesity (BMI ≥25). Basal characteristics of men and women in the subgroup analysis are shown in Supplementary Table S3 and Supplementary Table S4, respectively. There was a significant interaction between FLI and obesity for detection of NAFLD in women (p < 0.001) but not in men (p = 0.679). The cutoff values of FLI for prediction of NAFLD in non-obese and obese men were 22.6 (sensitivity/specificity: 71.4%/65.9%, AUC [95% CI]: 0.75 [0.74–0.77]) and 52.6 (sensitivity/specificity: 69.6%/67.0%, AUC [95% CI]: 0.74 [0.73–0.76]), respectively (Table 4). On the other hand, the cutoff values of FLI for prediction of NAFLD in non-obese and obese women were 11.2 (sensitivity/specificity: 78.2%/80.7%, AUC [95% CI]: 0.86 [0.84–0.88]) and 33.2 (sensitivity/specificity: 75.9%/68.2%, AUC [95% CI]: 0.78 [0.76–0.81]), respectively (Table 4).

Table 4 Cutoff values of FLI for prediction of NAFLD in non-obese and obese subjects

	n	Cutoff	AUC	95% CI	Sensitivity (%)	Specificity (%)
Men
Non-obese	4,967	22.6	0.75	0.74–0.77	71.4	65.9
Obese	4,273	52.6	0.74	0.73–0.76	69.6	67.0
Women
Non-obese	4,162	11.2	0.86	0.84–0.88	78.2	80.7
Obese	1,069	33.2	0.78	0.76–0.81	75.9	68.3

AUC, area under the curve; CI, confidence interval; FLI, fatty liver index; NAFLD, nonalcoholic fatty liver disease.

Discussion

To the best of our knowledge, we showed for the first time that the cutoff values of FLI for prediction of NAFLD determined by abdominal ultrasonography in Japanese men and women were 35.1 and 15.6, respectively. There was a significant interaction between sex and FLI for detection of NAFLD (p < 0.001), suggesting that there is a significant sex difference in the cutoff value of FLI for prediction of NAFLD. As shown in Supplementary Table S1, possible racial and sex differences in FLI for the cutoff level have been reported. A previous study using Asians in China showed that the cutoff value was FLI ≥30, but the sex difference was not taken into consideration [18]. A study in Taiwan showed that the cutoff values in men (n = 16,098) and women (n = 13,699) were FLI ≥35 and FLI ≥20, respectively [22]. Another study in Taiwan also showed that the cutoff values in men (n = 424) and women (n = 947) were FLI ≥20 and FLI ≥10, respectively [23]. The cutoff values of FLI for prediction of NAFLD obtained in Asian individuals in those studies [22, 23] as well as the present study were obviously lower than that in the original study (FLI ≥60) [16]. Setting cutoff values of FLI in different races and sexes is essential for predicting NAFLD.

NAFLD is common in not only obese subjects but also lean subjects [5], and the incidence of NAFLD has been increasing even in lean individuals in recent years [34]. In Asia, 15–20% of subjects with NAFLD were reported to be non-obese subjects [6, 35, 36]. Lean patients with NAFLD have been reported to require medical attention due to the same unhealthy metabolic environment as that for obese patients with NAFLD [37, 38]. The usefulness of FLI for detection of NAFLD in lean subjects has also been reported [39, 40] (Supplementary Table S1). A study in Taiwan showed that the cutoff value for prediction of NAFLD in lean subjects with BMI <24 was FLI ≥15 [39]. Another study in China also showed that the cutoff value for prediction of NAFLD in lean subjects with BMI <23 and normal WC (<90 cm for men and <80 cm for women) was FLI ≥25.2 [40]. In the present study, when divided by the absence and presence of obesity (BMI ≥25), the cutoff values of FLI in non-obese/obese subjects in men and women were 22.6/52.6 and 11.2/33.2, respectively. The ability of FLI to predict NAFLD may be useful in both non-obese and obese subjects, though there was a significant interaction between FLI and obesity for detection of NAFLD in women (p < 0.001) but not in men (p = 0.679) in the present study.

It has been reported that detection of fatty liver by abdominal ultrasonography is difficult when fat content in the liver is less than 30% [41, 42], though performance of new equipment is improving nowadays. Therefore, FLI may not be suitable for detection of fatty liver at an early stage. However, FLI is a non-invasive, inexpensive and convenient indicator that is useful for identification of patients with NAFLD. Calculating FLI may trigger early efforts to prevent metabolic complications in patients with NAFLD. On the other hand, a new concept of fatty liver associated with metabolic disorders has recently been proposed as MAFLD, and FLI has been recommend as a biomarker for detection of fatty liver [15]. There have been several studies showing that FLI is associated with the development of hypertension [43, 44], diabetes mellitus [45, 46], chronic kidney disease [47, 48], heart failure [49] and cardiovascular disease [50, 51]. Determination of the cutoff value of FLI will make it easier to find several NAFLD-associated diseases.

The present study has some limitations. First, since the study subjects had a yearly health check-up at a single urban clinic, the possibility of sample selection bias cannot be ruled out. Second, since diagnosis of hepatic steatosis was performed by abdominal ultrasonography but not by liver biopsy, the pathological severity of hepatic steatosis was not taken into consideration. Third, since the amount of alcohol consumption was obtained by a questionnaire, it may not reflect the accurate amount. Forth, since recruited subjects were recipients of annual health examinations, most of the subjects would be apparently healthy. However, among a total of 14,471 subjects (men/women: 9,240/5,231), the numbers of men and women with a high level of FIB-4 index (≥1.3) as a possible risk for hepatic fibrosis [52, 53] in the non-NAFLD group were 1,090 (11.8%) and 817 (15.6%), respectively. Since it has been reported that NAFLD patients with severe fibrosis have steatosis in less than 30% of hepatocytes [54, 55], diagnosis of NAFLD by FLI may cause false negative results in some patients with NAFLD. Finally, among 28,990 subjects who received health examinations in 2006, hepatitis B surface antigen and hepatitis C antibody were examined in only 15,623 subjects (53.9%) and 15,051 subjects (51.9%), respectively. Therefore, some patients with viral hepatitis might have been included in the present study, though the prevalence of hepatitis B (0.63%) and that of hepatitis C (0.49%) were reported to be relatively low in the Japanese population [56, 57].

In conclusion, the cutoff value of FLI for prediction of NAFLD in Japanese individuals was lower than that in the original study (FLI ≥60), and there is a significant sex difference. The simple and useful cutoff values for prediction of NAFLD in Japanese men and women are FLI ≥35 (non-obese/obese: 23/53) and FLI ≥16 (non-obese/obese: 11/33), respectively.

Acknowledgements

None.

Conflict of Interest Statement

The authors declare no conflicts of interest.

References

• 1   Marchesini  G,  Bugianesi  E,  Forlani  G,  Cerrelli  F,  Lenzi  M, et al. (2003) Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 37: 917–923.
• 2   Hamaguchi  M,  Kojima  T,  Takeda  N,  Nakagawa  T,  Taniguchi  H, et al. (2005) The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 143: 722–728.
• 3   Amarapurkar  DN,  Hashimoto  E,  Lesmana  LA,  Sollano  JD,  Chen  PJ, et al. (2007) How common is non-alcoholic fatty liver disease in the Asia-Pacific region and are there local differences? J Gastroenterol Hepatol 22: 788–793.
• 4   Younossi  ZM,  Koenig  AB,  Abdelatif  D,  Fazel  Y,  Henry  L, et al. (2016) Global epidemiology of nonalcoholic fatty liver disease-Meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology 64: 73–84.
• 5   Xu  C,  Yu  C,  Ma  H,  Xu  L,  Miao  M, et al. (2013) Prevalence and risk factors for the development of nonalcoholic fatty liver disease in a nonobese Chinese population: the Zhejiang Zhenhai Study. Am J Gastroenterol 108: 1299–1304.
• 6   Nishioji  K,  Sumida  Y,  Kamaguchi  M,  Mochizuki  N,  Kobayashi  M, et al. (2015) Prevalence of and risk factors for non-alcoholic fatty liver disease in a non-obese Japanese population, 2011–2012. J Gastroenterol 50: 95–108.
• 7   Byrne  CD,  Targher  G (2015) NAFLD: a multisystem disease. J Hepatol 62: S47–S64.
• 8   Souza  MR,  Diniz Mde  F,  Medeiros-Filho  JE,  Araujo  MS (2012) Metabolic syndrome and risk factors for non-alcoholic fatty liver disease. Arq Gastroenterol 49: 89–96.
• 9   Amiri Dash Atan  N,  Koushki  M,  Motedayen  M,  Dousti  M,  Sayehmiri  F, et al. (2017) Type 2 diabetes mellitus and non-alcoholic fatty liver disease: a systematic review and meta-analysis. Gastroenterol Hepatol Bed Bench 10: S1–S7.
• 10   Zhao  YC,  Zhao  GJ,  Chen  Z,  She  ZG,  Cai  J, et al. (2020) Nonalcoholic fatty liver disease: an emerging driver of hypertension. Hypertension 75: 275–284.
• 11   Hamaguchi  M,  Kojima  T,  Takeda  N,  Nagata  C,  Takeda  J, et al. (2007) Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease. World J Gastroenterol 13: 1579–1584.
• 12   Chang  Y,  Ryu  S,  Sung  E,  Woo  HY,  Oh  E, et al. (2008) Nonalcoholic fatty liver disease predicts chronic kidney disease in nonhypertensive and nondiabetic Korean men. Metabolism 57: 569–576.
• 13   Bhala  N,  Angulo  P,  van der Poorten  D,  Lee  E,  Hui  JM, et al. (2011) The natural history of nonalcoholic fatty liver disease with advanced fibrosis or cirrhosis: an international collaborative study. Hepatology 54: 1208–1216.
• 14   Baffy  G,  Brunt  EM,  Caldwell  SH (2012) Hepatocellular carcinoma in non-alcoholic fatty liver disease: an emerging menace. J Hepatol 56: 1384–1391.
• 15   Eslam  M,  Newsome  PN,  Sarin  SK,  Anstee  QM,  Targher  G, et al. (2020) A new definition for metabolic dysfunction-associated fatty liver disease: an international expert consensus statement. J Hepatol 73: 202–209.
• 16   Bedogni  G,  Bellentani  S,  Miglioli  L,  Masutti  F,  Passalacqua  M, et al. (2006) The fatty liver index: a simple and accurate predictor of hepatic steatosis in the general population. BMC Gastroenterol 6: 33.
• 17   Koehler  EM,  Schouten  JN,  Hansen  BE,  Hofman  A,  Stricker  BH, et al. (2013) External validation of the fatty liver index for identifying nonalcoholic fatty liver disease in a population-based study. Clin Gastroenterol Hepatol 11: 1201–1204.
• 18   Huang  X,  Xu  M,  Chen  Y,  Peng  K,  Huang  Y, et al. (2015) Validation of the fatty liver index for nonalcoholic fatty liver disease in middle-aged and elderly Chinese. Medicine (Baltimore) 94: e1682.
• 19   Chen  LD,  Huang  JF,  Chen  QS,  Lin  GF,  Zeng  HX, et al. (2019) Validation of fatty liver index and hepatic steatosis index for screening of non-alcoholic fatty liver disease in adults with obstructive sleep apnea hypopnea syndrome. Chin Med J (Engl) 132: 2670–2676.
• 20   Zhu  J,  He  M,  Zhang  Y,  Li  T,  Liu  Y, et al. (2018) Validation of simple indexes for nonalcoholic fatty liver disease in western China: a retrospective cross-sectional study. Endocr J 65: 373–381.
• 21   Kim  JH,  Kwon  SY,  Lee  SW,  Lee  CH (2011) Validation of fatty liver index and lipid accumulation product for predicting fatty liver in Korean population. Liver Int 31: 1600–1601.
• 22   Yang  BL,  Wu  WC,  Fang  KC,  Wang  YC,  Huo  TI, et al. (2015) External validation of fatty liver index for identifying ultrasonographic fatty liver in a large-scale cross-sectional study in Taiwan. PLoS One 10: e0120443.
• 23   Chen  LW,  Huang  PR,  Chien  CH,  Lin  CL,  Chien  RN (2020) A community-based study on the application of fatty liver index in screening subjects with nonalcoholic fatty liver disease. J Formos Med Assoc 119: 173–181.
• 24   Murayama  K,  Okada  M,  Tanaka  K,  Inadomi  C,  Yoshioka  W, et al. (2021) Prediction of nonalcoholic fatty liver disease using noninvasive and non-imaging procedures in Japanese health checkup examinees. Diagnostics (Basel) 11: 132.
• 25   Grundy  SM,  Cleeman  JI,  Daniels  SR,  Donato  KA,  Eckel  RH, et al. (2005) Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute scientific statement. Circulation 112: 2735–2752.
• 26   Higashiura  Y,  Tanaka  M,  Furuhashi  M,  Koyama  M,  Ohnishi  H, et al. (2020) Low urine pH predicts new onset of diabetes mellitus during a 10-year period in men: BOREAS-DM1 study. J Diabetes Investig 11: 1490–1497.
• 27   Mori  K,  Furuhashi  M,  Tanaka  M,  Numata  K,  Hisasue  T, et al. (2021) U-shaped relationship between serum uric acid level and decline in renal function during a 10-year period in female subjects: BOREAS-CKD2. Hypertens Res 44: 107–116.
• 28   Sterling  RK,  Lissen  E,  Clumeck  N,  Sola  R,  Correa  MC, et al. (2006) Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology 43: 1317–1325.
• 29   American Diabetes  Association (2017) 2. Classification and diagnosis of diabetes. Diabetes Care 40: S11–S24.
• 30   Hamaguchi  M,  Kojima  T,  Itoh  Y,  Harano  Y,  Fujii  K, et al. (2007) The severity of ultrasonographic findings in nonalcoholic fatty liver disease reflects the metabolic syndrome and visceral fat accumulation. Am J Gastroenterol 102: 2708–2715.
• 31   Lee  SS,  Park  SH (2014) Radiologic evaluation of nonalcoholic fatty liver disease. World J Gastroenterol 20: 7392–7402.
• 32   Schisterman  EF,  Perkins  NJ,  Liu  A,  Bondell  H (2005) Optimal cut-point and its corresponding Youden Index to discriminate individuals using pooled blood samples. Epidemiology 16: 73–81.
• 33   Kanda  Y (2013) Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant 48: 452–458.
• 34   Kumar  R,  Mohan  S (2017) Non-alcoholic fatty liver disease in lean subjects: characteristics and implications. J Clin Transl Hepatol 5: 216–223.
• 35   Liu  CJ (2012) Prevalence and risk factors for non-alcoholic fatty liver disease in Asian people who are not obese. J Gastroenterol Hepatol 27: 1555–1560.
• 36   Fukuda  T,  Hamaguchi  M,  Kojima  T,  Hashimoto  Y,  Ohbora  A, et al. (2016) The impact of non-alcoholic fatty liver disease on incident type 2 diabetes mellitus in non-overweight individuals. Liver Int 36: 275–283.
• 37   Sookoian  S,  Pirola  CJ (2017) Systematic review with meta-analysis: risk factors for non-alcoholic fatty liver disease suggest a shared altered metabolic and cardiovascular profile between lean and obese patients. Aliment Pharmacol Ther 46: 85–95.
• 38   Younossi  Z,  Anstee  QM,  Marietti  M,  Hardy  T,  Henry  L, et al. (2018) Global burden of NAFLD and NASH: trends, predictions, risk factors and prevention. Nat Rev Gastroenterol Hepatol 15: 11–20.
• 39   Hsu  CL,  Wu  FZ,  Lin  KH,  Chen  YH,  Wu  PC, et al. (2019) Role of fatty liver index and metabolic factors in the prediction of nonalcoholic fatty liver disease in a lean population receiving health checkup. Clin Transl Gastroenterol 10: 1–8.
• 40   Zeng  J,  Yang  RX,  Sun  C,  Pan  Q,  Zhang  RN, et al. (2020) Prevalence, clinical characteristics, risk factors, and indicators for lean Chinese adults with nonalcoholic fatty liver disease. World J Gastroenterol 26: 1792–1804.
• 41   Hernaez  R,  Lazo  M,  Bonekamp  S,  Kamel  I,  Brancati  FL, et al. (2011) Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 54: 1082–1090.
• 42   Bohte  AE,  van Werven  JR,  Bipat  S,  Stoker  J (2011) The diagnostic accuracy of US, CT, MRI and 1H-MRS for the evaluation of hepatic steatosis compared with liver biopsy: a meta-analysis. Eur Radiol 21: 87–97.
• 43   Roh  JH,  Park  JH,  Lee  H,  Yoon  YH,  Kim  M, et al. (2020) A close relationship between non-alcoholic fatty liver disease marker and new-onset hypertension in healthy Korean adults. Korean Circ J 50: 695–705.
• 44   Higashiura  Y,  Furuhashi  M,  Tanaka  M,  Takahashi  S,  Mori  K, et al. (2021) Elevated fatty liver index is independently associated with new onset of hypertension during a 10-year period in both male and female subjects. J Am Heart Assoc 10: e021430.
• 45   Yadav  D,  Choi  E,  Ahn  SV,  Koh  SB,  Sung  KC, et al. (2016) Fatty liver index as a simple predictor of incident diabetes from the KoGES-ARIRANG study. Medicine (Baltimore) 95: e4447.
• 46   Higashiura  Y,  Furuhashi  M,  Tanaka  M,  Takahashi  S,  Koyama  M, et al. (2021) High level of fatty liver index predicts new onset of diabetes mellitus during a 10-year period in healthy subjects. Sci Rep 11: 12830.
• 47   Huh  JH,  Kim  JY,  Choi  E,  Kim  JS,  Chang  Y, et al. (2017) The fatty liver index as a predictor of incident chronic kidney disease in a 10-year prospective cohort study. PLoS One 12: e0180951.
• 48   Takahashi  S,  Tanaka  M,  Furuhashi  M,  Moniwa  N,  Koyama  M, et al. (2021) Fatty liver index is independently associated with deterioration of renal function during a 10-year period in healthy subjects. Sci Rep 11: 8606.
• 49   Furuhashi  M,  Muranaka  A,  Yuda  S,  Tanaka  M,  Koyama  M, et al. (2021) Independent association of fatty liver index with left ventricular diastolic dysfunction in subjects without medication. Am J Cardiol 158: 139–146.
• 50   Zou  B,  Yeo  YH,  Cheung  R,  Ingelsson  E,  Nguyen  MH (2021) Fatty liver index and development of cardiovascular disease: findings from the UK biobank. Dig Dis Sci 66: 2092–2100.
• 51   Olubamwo  OO,  Virtanen  JK,  Voutilainen  A,  Kauhanen  J,  Pihlajamaki  J, et al. (2018) Association of fatty liver index with the risk of incident cardiovascular disease and acute myocardial infarction. Eur J Gastroenterol Hepatol 30: 1047–1054.
• 52   Shah  AG,  Lydecker  A,  Murray  K,  Tetri  BN,  Contos  MJ, et al. (2009) Comparison of noninvasive markers of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol 7: 1104–1112.
• 53   Sumida  Y,  Yoneda  M,  Hyogo  H,  Itoh  Y,  Ono  M, et al. (2012) Validation of the FIB4 index in a Japanese nonalcoholic fatty liver disease population. BMC Gastroenterol 12: 2.
• 54   Caldwell  SH,  Oelsner  DH,  Iezzoni  JC,  Hespenheide  EE,  Battle  EH, et al. (1999) Cryptogenic cirrhosis: clinical characterization and risk factors for underlying disease. Hepatology 29: 664–669.
• 55   Mulhall  BP,  Ong  JP,  Younossi  ZM (2002) Non-alcoholic fatty liver disease: an overview. J Gastroenterol Hepatol 17: 1136–1143.
• 56   Tanaka  J,  Kumagai  J,  Katayama  K,  Komiya  Y,  Mizui  M, et al. (2004) Sex- and age-specific carriers of hepatitis B and C viruses in Japan estimated by the prevalence in the 3,485,648 first-time blood donors during 1995–2000. Intervirology 47: 32–40.
• 57   Tanaka  J,  Koyama  T,  Mizui  M,  Uchida  S,  Katayama  K, et al. (2011) Total numbers of undiagnosed carriers of hepatitis C and B viruses in Japan estimated by age- and area-specific prevalence on the national scale. Intervirology 54: 185–195.