Association between intrahepatic triglyceride content in subjects with metabolically healthy abdominal obesity and risks of pre-diabetes plus diabetes: an observational study.

OBJECTIVE: We aimed to evaluate the association of intrahepatic triglyceride (IHTG) content in subjects with metabolically healthy abdominal obesity (MHAO) on risks of pre-diabetes plus diabetes.
DESIGN: Cross-sectional survey.
SETTING: Lianqian community, the First Affiliated Hospital of Xiamen University, Xiamen, China. PARTICIPANTS: Among 1523 community-living healthy adults aged 40 years or older with abdominal obesity recruited at baseline, 428 subjects who underwent IHTG content measurement were selected. OUTCOME MEASURES: Risk of pre-diabetes plus diabetes.
RESULTS: Non-alcoholic fatty liver disease (NAFLD) was diagnosed as 203 (69.1%) in MHAO and 121 (90.3%) in metabolically unhealthy abdominal obesity (MUAO) (p<0.001). The prevalence rates of pre-diabetes plus diabetes were 81.1%, 88.8% and 90.9% across the tertiles of IHTG content (p=0.037). Both MUAO (vs MHAO) and NAFLD (vs non-NAFLD) were independently associated with increased risks of pre-diabetes plus diabetes, the adjusted ORs (95% CIs) were 10.90 (3.15 to 37.69, p<0.001) and 3.02 (1.47 to 6.20, p=0.003), respectively. Higher IHTG content was significantly associated with increased risk of pre-diabetes plus diabetes with the adjusted OR (95% CI) of per SD increase of IHTG content of 1.62 (1.07 to 2.46, p=0.024). And there was a significantly positive trend between increasing categories of IHTG content tertiles and excessive risks of pre-diabetes plus diabetes (trend test p value=0.011). Stratified analyses showed similar results on the associations of NAFLD and IHTG content with risks of pre-diabetes plus diabetes for subjects with MHAO but not for those with MUAO.
CONCLUSIONS: NAFLD and higher IHTG content were independently associated with increased risks of pre-diabetes plus diabetes in MHAO subjects. NAFLD or quantity of liver fat should be considered as additional criterion when defining and diagnosing MHO. Screening of NAFLD and intervention to reduce liver fat should be strengthened even for those seemly metabolically healthy obese. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This study was a cross-sectional analysis of baseline information on the ongoing cohort study to evaluate the independent association of intrahepatic triglyceride (IHTG) content with risk of pre-diabetes plus diabetes.

All subjects were abdominally obese and were not randomly sampled from their living communities.

IHTG content was determined using magnetic resonance spectroscopy which was relatively quantitative measurement of liver fat.

The sample size was relatively small, especially for the metabolically unhealthy abdominal obesity subgroup, and we might not have enough power to determine their true associations.

Introduction

The prevalence of diabetes has quadrupled during the past three decades with an estimated prevalence of 9.3% (463 million people) in 2019 and it is expected to rise to 10.2% (578 million) by 2030 and 10.9% (700 million) by 2045.1–3 Obesity has been well documented to be a risk factor for a broad array of chronic non-communicable diseases, including diabetes, hypertension, coronary heart disease, chronic kidney disease and certain sites of cancer.4–6 A subgroup of individuals with obesity who are devoid of obesity-related metabolic complications, such as diabetes and atherosclerosis, arise the concept of metabolically healthy obese (MHO).7–9 However, there is no unique definition and diagnose criteria for MHO by now. For example, some defined MHO when two or fewer of the four criteria of metabolism syndrome10 for those subjects with obesity while others defined as none of them,11 which made evidence on the association of MHO with diabetes was limited and controversial.12

Non-alcoholic fatty liver disease (NAFLD) is well documented to be associated with risk of diabetes13; however, NAFLD has not been considered as additional criterion for MHO although it usually occurs simultaneously when obesity happens. Therefore, little evidence is available on the risk of NAFLD or liver fat with diabetes for those with MHO. In the present study with 428 community-living Chinese adults with abdominal obesity, we mainly aimed to evaluate associations of intrahepatic triglyceride (IHTG) content and NAFLD in subjects with metabolically healthy abdominal obesity (MHAO) on risks of pre-diabetes plus diabetes.

Methods

Study design and subjects

Details on study design and subject recruitment have been described previously.14–16 Briefly, 1523 community-living healthy adults aged 40 years or older with abdominal obesity (waist circumference greater than 90 cm for men and 80 cm for women) living in Lianqian community, Xiamen, China were recruited at baseline of the cohort study in 2011. Of them, 92 (6%) who had incomplete data on clinical and biochemistry measurements were excluded, and a random sample of 428 subjects who underwent IHTG content measurement was left for the present analysis (figure 1). Of the 428 study subjects, 319 (74.5%) were female with the mean age of 53.6±6.5 years old, 109 (25.5%) were male with the mean age of 53.2±7.1 years old, and there was no significant difference in age between male and female subjects (p=0.592).

Figure 1

Study subjects’ selection diagram. IHTG, intrahepatic triglyceride.

Measurements

Details on study measurements have been described previously.15 16 For each subject, face-to-face interview was conducted to collect sociodemographic status, lifestyle habits, present and previous history of health and medications. Subjects were excluded if they drank regularly with alcohol consumption ≥140 g/week for men or ≥70 g/week for women, had cancer, or received current treatment with systemic corticosteroids, biliary obstructive diseases, acute or chronic virus hepatitis, drug-induced liver diseases, total parenteral nutrition, autoimmune hepatitis, Wilson’s disease, known hyperthyroidism or hypothyroidism. Subjects underwent weight, height and waist circumference measurements by using a calibrated scale after removing shoes and heavy clothes. Waist circumference was measured at the midpoint between the inferior costal margin and the superior border of the iliac crest on the midaxillary line. Body mass index (BMI) was calculated as weight in kilograms divided by height in squared metres. Arterial blood pressure (BP) was measured with a mercury sphygmomanometer after sitting for at least 15 min.

Blood samples were obtained after 12-hour fasting and 75 g oral glucose tolerance test were conducted for each subject. All biochemical measurements were tested in the central laboratory of the First Affiliated Hospital, Xiamen University. Plasma glucose and serum lipid profiles, including triglyceride (TG), total cholesterol (TC) and high-density lipoprotein cholesterol (HDL-C) were determined on a HITACHI 7450 analyzer (HITACHI, Tokyo, Japan). Low-density lipoprotein cholesterol (LDL-C) was calculated by Friedewald’s formula. Fasting plasma glucose (FPG) and 2-hour plasma glucose (2-h PG) concentrations were measured by the hexokinase method and Hemoglobin A1c (HbA1c) by the Bio-Rad Variant Hemoglobin A1c assay.

Ultrasonography and definition of NAFLD

Details on ultrasonography and definition of NAFLD have been described previously.15 16 Hepatic ultrasonography scanning was performed by an experienced radiologist using GE LOGIQ P5 scanner (GE Healthcare, Milwaukee, USA) with a 4-MHz probe, who was blinded to the subjects’ health status. Hepatic steatosis was diagnosed on the basis of characteristic sonographic features, including hepatorenal echo contrast, liver parenchymal brightness, deep beam attenuation and vessel blurring.17 The definition of NAFLD was based on hepatic ultrasonography diagnosis of hepatic steatosis without excessive alcohol consumption, viral or autoimmune liver disease.15 16

IHTG content measurement

Details on IHTG content measurement has been described previously.18 IHTG content was determined by an experienced radiologist using magnetic resonance spectroscopy (1H MRS, 3.0-T Avanto, Siemens AG, Erlangen, German). Images of a sagittal, coronal and axial cube of a 2 cm3 volume in the right lobe of liver was acquired. Quantification of the spectra (water and methylene resonances) was performed as described previously.19 Areas of resonance from water protons and methylene groups in fatty acid chains were obtained with a time-domain non-linear fitting routine by using Syngo MR B15V software (Siemens AG). The percentage of IHTG content was calculated as the ratio of the area under the resonance of peak for methylene groups in fatty acid chains of IHTG and the combined area under the resonance peaks for both methylene groups and water.18 19

Definition of MHAO

Abdominal obesity was defined as WC ≥90 cm for men and 80 cm for women.20 All subjects in the present study were abdominal obesity which was considered as one of the recruitment criteria. Subjects were defined as MHAO if two or fewer of the following criteria were met: (1) systolic BP ≥130 or diastolic BP ≥85 mmHg; (2) FPG ≥100 mg/dL (5.6 mmol/L); (3) TG ≥150 mg/dL (1.7 mmol/L); (4) HDL cholesterol <40 mg/dL (1.03 mmol/L) in men and <50 mg/dL (1.30 mmol/L) in women.16 21 22 Otherwise, subjects meeting 3 or more of the criteria were defined as metabolically unhealthy abdominal obesity (MUAO). Therefore, all subjects in the present study were dichotomised as either MHAO or MUAO.

Definitions of diabetes and pre-diabetes

According to American Diabetes Association 2020 criteria, diabetes was defined as (1) a self-reported history of diabetes previously diagnosed by healthcare professionals; (2) FGP ≥126 mg/dL (7.0 mmol/L); (3) 2-hour plasma glucose (2-h PG, oral glucose tolerance test)≥200 mg/dL (11.1 mmol/L); or (4) HbA1c≥6.5%. Pre-diabetes was defined as (1) FPG levels between 100 mg/dL (5.6 mmol/L) and 125 mg/ dL (6.9 mmol/L), (2) 2-h PG levels between 140 mg/dL (7.8 mmol/L) and 199 mg/dL (11.0 mmol/L), or (3) HbA1c between 5.7% and 6.4% in participants without a prior diabetes diagnosis.15 16 23

Statistical analyses

Methods on statistical analyses were similar to our previous publications.15 16 18 Data were presented as the mean±SD for continuous variables or number and percentage for categorical variables. Skewness and kurtosis tests for continuous variables were conducted and found them followed approximation of normal distributions. Differences between subjects categorised by MHAO and tertiles of IHTG content were analysed using one-way analysis of variance for continuous variables and Χ2 test for categorical variables. Bar graphs showing prevalence rates of diabetes, pre-diabetes and normal glucose test (NGT) were made by MHAO (vs MUAO) and tertiles of IHTG content.

Multivariable logistic regression models were used to calculate the adjusted ORs and 95% CIs of abdominal obesity (MUAO vs MHAO), NAFLD (yes vs no) and IHTG content (both the originally continuous values and the tertiles categories) for pre-diabetes plus diabetes with adjustment for potential confounders (including age, sex, educational level, smoking and drinking habits, regular physical exercise, BMI, systolic and diastolic BP, TG, TC, HDL-C and LDL-C and serum uric acid). And multivariable logistic regression analyses stratified by MHAO and MUAO groups were further conducted. All p values were two sided and p value<0.05 was considered statistically significant. All statistical analyses were performed using Stata V.14.0 (StatCorp).

Patient and public involvement

There were no funds or time allocated for patient and public involvement.

Results

Prevalence of diabetes and pre-diabetes stratified by MHAO and tertiles of IHTG content

Among the 428 subjects with abdominal obesity, MHAO and MUAO were identified on 294 (68.7%) and 134 (31.3%) subjects. Of them, 46 (10.8%), 326 (76.2%) and 56 (13.1%) were diagnosed as diabetes, pre-diabetes and NGT, respectively. There was a significantly positive trend between increasing tertiles of IHTG content and higher prevalence of pre-diabetes plus diabetes (81.1%, 88.8% and 90.9% across the tertiles of IHTG content (p=0.037)). Figure 2A showed the prevalence rates of diabetes and pre-diabetes across the tertiles of IHTG content in MHAO subjects were 7.1% and 67.3%, 10.2% and 74.5%, 10.2% and 77.6% for the tertile 1, tertile 2 and tertile 3, respectively (p value>0.05). But there was a significantly positive trend of higher prevalence of diabetes plus diabetes with increasing categories of tertiles of IHTG content (p=0.039). Figure 2B showed the prevalence rates of diabetes and pre-diabetes across the tertiles of IHTG content in MUAO subjects were 8.9% and 86.7%, 8.9% and 88.9%, 25.0% and 72.7% for tertile 1, tertile 2 and tertile 3, respectively. Table 1 also showed MUAO subjects had significantly higher prevalence of pre-diabetes and pre-diabetes plus diabetes than MHAO subjects (both p values<0.05).

Figure 2

(A) Prevalence rates (%) of pre-diabetes and diabetes stratified by tertiles of intrahepatic triglyceride (IHTG) content in metabolically healthy abdominal obesity subjects. (B) Prevalence rates (%) of pre-diabetes and diabetes stratified by tertiles of IHTG content in metabolically unhealthy abdominal obesity subjects.

Table 1

Demographic, lifestyle and clinical characteristics of 428 subjects stratified by MHAO and tertiles of IHTG content

Variables	MHAO (n=294, 68.7%)				MUAO (n=134, 31.3%)
	Tertile 1	Tertile 2	Tertile 3	P value	Tertile 1	Tertile 2	Tertile 3	P value	P value*
Demographics and life style
N (%)	98 (33.3%)	98 (33.3%)	98 (33.3%)		45 (33.6%)	45 (33.6%)	44 (32.8%)
Male gender (n, %)	12 (12.4%)	27 (27.6%)	30 (30.6%)	0.005†	11 (24.4%)	17 (38.8%)	12 (27.3%)	0.347	0.160
Age (years)	52.8±6.5	53.6±6.6	52.4±6.8	0.422	54.8±6.1	54.1±7.5	55.0±6.0	0.821	0.014†
Low educational attainment, (n, %)	61 (62.2%)	48 (49.0%)	51 (52.0%)	0.241	30 (66.7%)	25 (55.6%)	27 (61.4%)	0.110	0.090
Ever smoking (n, %)	16 (16.3%)	25 (25.5%)	23 (23.5%)	0.262	9 (20.0%)	11 (24.4%)	11 (25.0%)	0.828	0.753
Ever drinking (n, %)	8 (8.2%)	13 (13.3%)	15 (15.3%)	0.291	2 (4.4%)	6 (13.3%)	4 (9.1%)	0.216	0.317
Regular physical exercise (n, %)	34 (34.7%)	42 (42.9%)	32 (32.7%)	0.292	13 (28.9%)	17 (37.8%)	10 (22.7%)	0.296	0.165
Clinical characteristics
IHTG content (%)	4.10±1.63	9.64±1.97	23.25±8.47	<0.001‡	6.86±2.44	14.54±2.91	27.68±6.21	<0.001‡	<0.001‡
NAFLD (n, %)	34 (34.7%)	75 (76.5%)	94 (95.9%)	<0.001‡	35 (77.8%)	42 (93.3%)	44 (100.0%)	<0.001‡	<0.001‡
BMI (kg/m2)	26.2±2.5	27.2±2.5	28.3±2.7	<0.001‡	27.3±2.5	27.7±3.5	28.1±2.8	0.485	0.121
Waist circumference (cm)	90.7±5.1	93.5±6.1	96.0±7.1	<0.001‡	94.0±5.3	93.7±7.8	95.3±6.8	0.518	0.158
Systolic blood pressure (mm Hg)	125.6±17.3	130.6±17.5	129.5±15.7	0.095	139.8±12.4	140.9±15.3	143.4±12.6	0.451	<0.001‡
Diastolic blood pressure (mm Hg)	74.0±10.0	77.8±9.8	78.4±9.7	0.004†	82.7±10.7	83.7±9.4	85.2±8.3	0.482	<0.001‡
Triglyceride (mmol/L)	1.22±0.57	1.56±0.82	1.89±0.98	<0.001‡	2.53±1.26	3.32±1.90	2.84±1.52	0.062	<0.001‡
Total cholesterol (mmol/L)	5.80±1.03	5.81±0.95	5.95±1.05	0.487	6.18±1.43	6.13±1.27	6.02±0.94	0.823	0.024†
HDL-cholesterol (mmol/L)	1.50±0.26	1.39±0.29	1.37±0.25	0.003†	1.20±0.19	1.14±0.22	1.18±0.20	0.401	<0.001‡
LDL-cholesterol (mmol/L)	3.75±0.95	3.72±0.79	3.75±1.01	0.973	3.83±1.30	3.50±1.41	3.56±0.84	0.391	0.297
Blood uric acid (μmol/L)	322.9±71.5	359.8±83.6	378.2±98.1	<0.001‡	348.3±84.2	381.6±90.9	426.0±108.7	<0.001‡	0.001*
Fasting plasma glucose (mmol/L)	5.56±0.46	5.55±0.54	5.54±0.51	0.967	5.85±0.44	5.98±0.45	5.94±0.42	0.338	<0.001‡
2-h PG (OGTT, mmol/L)	7.42±2.52	7.51±1.86	7.87±1.88	0.285	8.04±1.91	8.09±1.72	8.87±2.13	0.082	<0.001‡
HbA1c (%)	5.86±0.29	5.93±0.36	5.97±0.29	0.041†	6.01±0.32	6.00±0.38	6.04±0.36	0.882	0.005†
Diabetes (n, %)	7 (7.1%)	10 (10.2%)	10 (10.2%)	0.693	4 (8.9%)	4 (8.9%)	11 (25.0%)	0.043†	0.122
Pre-diabetes (n, %)	66 (67.3%)	73 (74.5%)	76 (77.6%)	0.255	39 (86.7%)	40 (88.9%)	32 (72.7%)	0.091	0.039†
Pre-diabetes plus diabetes (n, %)	73 (74.5%)	83 (84.7%)	86 (87.8%)	0.039†	43 (95.6%)	44 (97.8%)	43 (97.7%)	0.779	<0.001‡

All percentages are column percentage; except for percentages, all values are mean±SD.

*P value for difference between MHAO and MUAO.

†P<0.05.

‡P<0.001.

BMI, body mass index; HDL, high-density lipoprotein; 2-h PG, 2-hour plasma glucose; IHTG, intrahepatic triglyceride; LDL, low-density lipoprotein cholesterol; MHAO, metabolically healthy abdominal obesity; MUAO, metabolically unhealthy abdominal obesity; NAFLD, non-alcoholic fatty liver disease; OGTT, oral glucose tolerance test.

Demographic and clinical characteristics stratified by MHAO and tertiles of IHTG content

For all the 428 subjects, the means (±SD) of age were 53.6 (±6.5) years for women (n=319, 74.5%) and 53.2 (±7.1) years for men (n=109, 25.5%) (p=0.592). Table 1 showed differences of demographics, life style habits and clinical characteristics stratified by MHAO and tertiles of IHTG content. For 294 MHAO subjects, with increasing categories of the tertiles of IHTG content (from tertile 1, tertile 2 to tertile 3), subjects were more likely to be male and had significantly higher levels of indices of obesity (BMI, waist circumference), diastolic BP, TG, HbA1c, serum uric acid as well as higher prevalence of NAFLD and significantly lower level of HDL-C. As for 134 MUAO subjects, increasing categories of the tertiles of IHTG content were significantly related to higher prevalence of NAFLD and serum uric acid levels. Furthermore, table 1 showed that, compared with subjects with MHAO, those with MUAO had significantly increased age, IHTG content, prevalence of NAFLD, systolic and diastolic BP, TG, TC, FPG, 2-h PG, HbA1c, serum uric acid and significantly lower level of HDL-C.

Associations of MHAO, NAFLD and IHTG content with pre-diabetes plus diabetes for all subjects

Table 2 showed that, for all subjects, both MUAO (vs MHAO) and NAFLD (yes vs no) were independently associated with increased risk of pre-diabetes plus diabetes, and the adjusted ORs (95% CIs) were 10.90 (3.15 to 37.69, p<0.001) and 3.02 (1.47 to 6.20, p=0.003), respectively. Higher IHTG content was significantly associated with increased risk of pre-diabetes plus diabetes with the adjusted OR (95% CI) of per SD increase of IHTG content of 1.62 (1.07 to 2.46, p=0.024). With the tertile 1 of IHTG content as the reference, the tertile 3 showed significantly higher risk of pre-diabetes plus diabetes (adjusted OR (95% CI): 3.13 (1.28 to 7.61), p=0.012). And there was a significantly positive trend of increasing categories of IHTG content tertiles with excessive risk of pre-diabetes plus diabetes (trend test: p=0.011). There was no significant interaction between MHAO with either NAFLD or tertiles of IHTG content for risk of pre-diabetes plus diabetes (both p values>0.05).

Table 2

Adjusted ORs with associated 95% CI of MUAO, NAFLD and IHTG content for pre-diabetes plus diabetes

Variables	Pre-diabetes plus diabetes
	OR	95% CI	P value
All subjects
MUAO versus MHAO	10.90	3.15 to 37.69	<0.001*
NAFLD versus non-NAFLD	3.02	1.47 to 6.20	0.003*
IHTG content (%)†	1.62	1.07 to 2.46	0.024*
IHTG content tertiles‡
Tertile 1	1.00
Tertile 2	1.81	0.86 to 3.81	0.117
Tertile 3	3.13	1.28 to 7.61	0.012*
Trend test			0.011*
Interaction test
MUAO*NAFLD			0.956
MUAO*Tertiles of IHTG			0.869
MHAO subjects
NAFLD versus non-NAFLD	2.65	1.25 to 5.60	0.011*
IHTG content (%)†	1.55	1.00 to 2.40	0.051
IHTG content tertiles‡
Tertile 1	1.00
Tertile 2	2.31	1.03 to 5.17	0.042*
Tertile 3	2.81	1.14 to 6.90	0.024*
Trend test			0.021*
MUAO subjects
NAFLD versus Non-NAFLD	4.77	0.07 to 327.48	0.469
IHTG content (%)†	0.81	0.13 to 5.26	0.830
IHTG content tertiles‡
Tertile 1	1.00
Tertile 2	3.22	0.24 to 43.54	0.378
Tertile 3	1.90	0.15 to 23.69	0.620
Trend test			0.558

OR was adjusted for age, sex, educational level, ever smoking, ever drinking, physical activity, BMI, systolic and diastolic BP, triglyceride, total cholesterol, HDL-cholesterol and LDL-cholesterol and serum uric acid.

*P<0.05.

†OR and 95% CI were expressed by per SD increase of IHTG content.

‡OR and 95% CI was expressed by the first quartile of IHTG content as the reference.

IHTG, intrahepatic triglyceride; MHAO, metabolically healthy abdominal obesity; MUAO, metabolically unhealthy abdominal obesity; NAFLD, non-alcoholic fatty liver disease.;

Stratified analyses on associations of NAFLD and IHTG content with pre-diabetes plus diabetes by MHAO and MUAO

Multivariable logistic regression analyses stratified by MHAO and MUAO separately were conducted (table 2). For MHAO subjects, NAFLD was independently associated with increased risk of pre-diabetes plus diabetes (adjusted OR (95% CI): 2.65 (1.25 to 5.60), p=0.011). Per SD increase of IHTG content was marginally associated with excessive risk of pre-diabetes plus diabetes with the adjusted OR (95% CI) of 1.55 (1.00 to 2.40, p=0.051). Compared with the tertile 1 of IHTG content, both the tertile 2 and tertile 3 groups showed significantly increased risks of pre-diabetes plus diabetes with the adjusted ORs (95% CI) of 2.31 (1.03 to 5.17, p=0.042) and 2.81 (1.14 to 6.90, p=0.024), respectively. And there was also a significantly positive trend between increasing categories of IHTG content tertiles and excessive risk of pre-diabetes plus diabetes (trend test: p=0.021). For MUAO subjects, neither NAFLD nor IHTG content was found to be significantly associated with risk of pre-diabetes plus diabetes.

Discussion

In the present study of 428 subjects with abdominal obesity, 294 (68.7%) and 134 (31.3%) were identified as MHAO and MUAO, respectively. Both MUAO (vs MHAO) and NAFLD (vs non-NAFLD) were independently associated with increased risks of pre-diabetes plus diabetes. Furthermore, higher IHTG content was significantly associated with increased risk of pre-diabetes plus diabetes, and there was a significantly positive trend between increasing categories of IHTG content tertiles and excessive risks of pre-diabetes plus diabetes. Stratified analyses showed similar results for subjects with MHAO but not for those with MUAO.

The concept of MHO has been established for a subgroup of subjects with obesity who do not exhibit metabolic and cardiovascular complications at a given time point, such as diabetes and atherosclerosis, for a few decades.24 25 Compared with subjects with MUO, those with MHO are characterised by lower liver and visceral fat, higher subcutaneous leg fat, greater cardiorespiratory fitness, physical activity and insulin sensitivity, lower levels of inflammation, and normal adipose tissue function.26 However, it could be debated whether MHO predicts the risk of diabetes compared with metabolically healthy normal weight or MUO. Hinnouho based on the Whitehall II cohort study found a significantly decreased risk of diabetes for MHO compared with metabolically unhealthy obesity (MUO) (HR=1.98 (MUO vs MHO), 95% CI 1.39 to 2.83).27 The present study found similar results that MUAO was significantly associated with increased risk of pre-diabetes plus diabetes compared with MHAO but with a much higher adjusted OR(95% CI) (10.90 (3.15 to 37.69)). Hinnouho and others further found that MHO showed a significant increased risk of type 2 diabetes mellitus (T2DM) incidence compared with metabolically healthy normal weight.27–29 We cannot evaluate the risk of MHAO on diabetes compared with metabolically healthy normal weight since all the subjects in the present study were central obese and none of them could be classified as metabolically healthy or unhealthy normal weight. And because we had a relatively small sample size, we might find the adjusted OR was much higher than those from other.27–29

Little evidence is available on differences of prevalence of NAFLD or liver fat content between MHO and MUO. In the present study, we found subjects with MUO, compared with those with MHO, showed significantly higher prevalence of NAFLD (90.3% vs 69.1%) and IHTG content (16.3±9.5 vs 12.3%±9.5%) (both p values<0.001). Our findings indicated the prevalence of NAFLD and IHTG content are still common and high even for those with MHO. Meanwhile, MHO is commonly identified based on the presence of obesity and absence of metabolic syndrome, neither NAFLD nor liver fat content has been considered as an additional criterion when defining and diagnosing MHO. Therefore, our findings implied that screening of NAFLD and intervention to reduce IHTG content for those seemly healthy obese should be strengthened.

Our previous findings showed that NAFLD was significantly associated with increased risk of T2DM prevalence.15 The present study expanded the positive association of NAFLD to risk of pre-diabetes plus diabetes for all subjects as well as for those with MHAO, and the adjusted ORs (95% CI) were 3.02 (1.47 to 6.20) and 2.65 (1.25 to 5.60) (both p values<0.05), respectively. NAFLD has been generally diagnosed by hepatic ultrasonography scanning. In the present study, we conducted IHTG content measurement by using magnetic resonance spectroscopy to quantify the extent of liver fat in these abdominal obese subjects. And we found that IHTG content was significantly associated with increased risk of pre-diabetes plus diabetes with the adjusted OR (95% CI) of per SD increase of IHTG content of 1.62 (1.07 to 2.46, p=0.024). Moreover, we found a significantly positive trend between increasing categories of IHTG content tertiles and excessive risks of pre-diabetes plus diabetes.

Quantitative MRI proton-density fat fraction method has been proved to serve as accurate noninvasive biomarkers for quantifying liver steatosis30 and liver fat content was found to be correlated with insulin resistance,31 but evidence was scarce on association between the quantity of liver fat and risk of diabetes. Our results on the association between IHTG content and risks of pre-diabetes plus diabetes might account for possibly a novel finding for the present study.

We further conducted stratified analyses on the associations of IHTG content with risk of pre-diabetes plus diabetes for subjects with MHAO and MUAO separately. For those with MHAO, the association of IHTG content with risk of pre-diabetes plus diabetes was marginally significant, and the adjusted OR (95% CI) of per SD increase of IHTG content was 1.55 (1.00 to 2.40, p=0.051). With the first tertile of IHTG content as the reference group, the adjusted ORs (95% CI) of risks of pre-diabetes plus diabetes for the second and third tertiles were 2.31 (1.03 to 5.17) and 2.81 (1.14 to 6.90) (both p values<0.05), respectively. The positive trend between increasing categories of IHTG content tertiles and excessive risks of pre-diabetes plus diabetes was also statistically significant for the subgroup with MHAO (trend test p value<0.05). Our findings implied that increased IHTG content was associated with excessive risk of pre-diabetes and diabetes even for MHO subjects. To the best of our knowledge, we were probably the first to find the positive associations of IHTG content with risks of diabetes and pre-diabetes for MHAO subjects. The reason for non-significant results for MUAO subgroups may be due to the relatively small sample size of subjects with MUAO (n=134).

NAFLD and liver fat quantity has not been currently considered in the definitions and diagnose criteria of MHO,16 26 although liver is one of the main parts of fat accumulation when obesity occurs. The present study found that around 69% of subjects with MHAO were diagnosed as NAFLD. Most importantly, even for these apparently MHO individuals, NAFLD and higher IHTG content were both significantly associated with increased risks of pre-diabetes plus diabetes. Therefore, our findings implied that the current criteria of MHO may not be appropriate. NAFLD, quantity of liver fat or abdominal fat content (obtained from ultrasonography or CT-scanning techniques) should be considered as additional criterion when defining and diagnosing MHO if more evidence could be proved in future, especially from the prospective cohort studies with larger sample sizes.

A few limitations of the present study should be recognised when generalising our findings to other populations. First, all subjects were abdominally obese and were not randomly sampled from their living communities; therefore, we could not assess the effect of MHAO as compared with metabolically healthy non-obesity and we might also under-estimate the true associations of MHAO as compared with MUAO on risks of pre-diabetes plus diabetes. Second, the present analyses were based on the baseline information of our ongoing cohort study, therefore, we cannot determine the temporal sequence among MHAO and pre-diabetes plus diabetes. Third, our sample size was small, especially for the MUAO subgroup and we may not have enough power to determine their true associations. On the other hand, we still have some strengths in the present study. For example, we used IHTG content by magnetic resonance spectroscopy, which was relatively measurement of liver fat. And we were probably the first to find the positive associations of IHTG content with risks of diabetes and pre-diabetes, especially for subjects with MHAO.

Conclusions

NAFLD were diagnosed in 69% of MHAO and 90% of MUAO subjects, and the prevalence rates of pre-diabetes plus diabetes were linearly increased across the tertiles of IHTG content. NAFLD and higher IHTG content were independently associated with increased risks of pre-diabetes plus diabetes for all subjects as well as for the MHAO subgroups. Therefore, our findings imply that NAFLD or quantity of liver fat should be considered as additional criterion when defining and diagnosing MHO. Furthermore, screening of NAFLD and intervention to reduce liver fat should be strengthened even for the seemly healthy obese subjects.