PNPLA3 rs 738409 and Other Nongenetic Factors Associated with Hepatic Steatosis Estimated by Magnetic Resonance Imaging Proton Density Fat Fraction in Adult Greek Subjects with Type 2 Diabetes Mellitus

Abstract

Objectives:

Nonalcoholic fatty liver disease is dramatically increasing in parallel with the pandemic of type 2 diabetes mellitus. We investigated factors associated with hepatic steatosis (HS) in adult Greek individuals with established type 2 diabetes mellitus.

Materials and Methods:

We investigated 120 consecutive people with type 2 diabetes attending the Diabetic Outpatient Clinic at an Academic Hospital in Athens, Greece. All of them had demographic, clinical, and biochemical data recorded. HS was estimated by magnetic resonance imaging determined by proton density fat fraction software and defined as the percentage of total liver fat divided by the liver volume. HS of >5% was considered abnormal. The PNPLA3 (I148M) variant was evaluated as a genetic factor by standard molecular techniques. FibroMax™ was also calculated.

Results:

Of the 120 participants, median age was 61.7, 46% were females, diabetes duration was 10 years, and HbA1c (glycated hemoglobin) was 6.7%. The median value of HS was 7.8. The PNPLA3 rs738409 CC/CG/GG genotype frequencies were 54.2%, 35%, and 10.8%, respectively. In multivariate analysis, PNPLA3 rs738409 (β = 0.425, P = 0.001), waist circumference (β = 2.448, P = 0.001), and female sex (β = 0.419, P = 0.002) had a direct association with HS, while duration of diabetes (β = −0.179, P = 0.011) had an inverse association with HS.

Conclusions:

HS in type 2 diabetes is the sum of interplay of various factors exerting a direct or an inverse association, the most prominent among them being abdominal obesity and PNPLA3 molecular variability.

Introduction

Nonalcoholic fatty liver disease (NAFLD)¹ has attracted the particular interest of medical research over the last two decades, as it affects almost a quarter of the world's population.² A significant number of epidemiological, translational, and basic research efforts have been made to understand the complex pathogenetic factors leading to NAFLD focusing on obesity,³ insulin resistance,^4

–7 type 2 diabetes, dyslipidemia, hypertension, and other causes associated with metabolic deregulation.⁸

Recent studies suggest that type 2 diabetes is an independent risk factor for the development of NAFLD.⁹ Almost 70% of people with type 2 diabetes are estimated to have NAFLD^10,11 and 20%–30% to have nonalcoholic steatohepatitis (NASH). This particular population is at high risk of disease progression to advanced fibrosis and cirrhosis.^10,12 Furthermore, studies have shown that this specific group has been quite pharmacologically challenging as beyond the histologic hallmarks of NASH, and alterations in the expression and function of xenobiotic metabolizing enzymes and transporters have been described.^12,13 It is clear, therefore, that NAFLD has become a major issue creating an immense clinical and economic burden.

As with most complex traits, the phenotypic patterns and severity of NAFLD appear to be the outcome of multiple interactions between genetic, epigenetic, and environmental factors.¹⁴ The substitution I to M at position 148 (rs738409 C>G) encoding for the enzyme patatin-like phospholipase domain-containing protein 3 (PNPLA3 I148M) was identified to be the strongest variant associated with the whole spectrum of NAFLD among different ethnicities.¹⁵ PNPLA3 is an enzyme that implicates in lipid regulation with its double activity as triacylglycerol lipase and acylglycerol 0-acyltransferase.¹⁶ Moreover, as a multifunctional enzyme, it has a retinyl ester activity in hepatic stellate cells based on in vitro and in vivo studies.¹⁷ Homozygosity for the I148M variant increases the risk of development of fibrosis by more than threefold¹⁸ and of HCC by more than eightfold.¹⁹

In this study, we aimed to investigate factors associated with hepatic steatosis (HS) as measured by proton density fat fraction in magnetic resonance imaging (MRI PDFF) in Greek people with type 2 diabetes unselected for NAFLD. First, we wanted to determine whether the presence of PNPLA3 variant I148M and its genotype frequencies CC(normal)/CG(heterozygote)/GG(homozygote) were associated with liver steatosis in our study population. Second, we wanted to assess the performance of several plasma biomarkers and noninvasive clinical models/scores on determining the presence of NAFLD in this population. To our knowledge, there are a few studies focused specifically on a European population with type 2 diabetes and no data in adult Greek population with established diagnosis of type 2 diabetes.

Materials and Methods

Subjects

A total of 140 consecutive people followed up at the Diabetes Outpatient Clinic of the 2nd Department of Internal Medicine, Hippokration General Hospital, Athens, Greece consented to participate in this study. All of them were diagnosed with type 2 diabetes according to the American Diabetes Association criteria.²⁰

Participants were eligible if they met the following criteria: age 20 to 70 years; being ambulatory without a recent acute illness; alcohol consumption of less than 30 grams/day for men and less than 20 grams/day for women; glycated hemoglobin (HbA1c) <8.00%; absence of nonhepatic malignancy; no use of immunosuppressive medications in the last 6 months; negative test for hepatitis B surface antigen (HBsAg), antibodies against hepatitis C (anti-HCV), anti-smooth muscle (SMA), and antinuclear antibodies (ANA); thyroid-stimulating hormone (TSH) and immunoglobulin G and M (IgG, IgM) levels in normal reference range; and absence of unrelated to type 2 diabetes chronic, heart, or renal disease.

Each participant underwent an interview by a trained investigator, full clinical examination, anthropometric measurements, and laboratory measurements. Information regarding participants' demographic characteristics, medical history, duration of diabetes, diabetic complications, and comorbidities such as hypertension and dyslipidemia and current medication was obtained through interviews and their medical records.

Blood pressure, height, bodyweight, and waist and hip circumference were measured without shoes or outer clothing. Body mass index (BMI) was calculated as the weight (kg) divided by the height squared (m²). Blood pressure was recorded as the mean of three consecutive measurements taken 5 min apart and in the sitting position.

Metabolic syndrome was defined using the most updated criteria of the International Diabetes Federation: (1) central obesity (waist circumference 94 cm in men and 80 cm in women) and (2) at least two of the following factors; (i) serum triglycerides >150 mg/dL or specific treatment for this lipid abnormality; (ii) serum high-density lipoprotein (HDL) <40 mg/dL in men and <50 mg/dL in women or specific treatment for this lipid abnormality; (iii) systolic blood pressure >130 mmHg or diastolic blood pressure >85 mmHg or specific antihypertensive treatment; and (iv) fasting blood glucose >100 mg/dL or previously diagnosed type 2 diabetes.

The study was approved by our regional Ethics Committee. Written informal consent was obtained from all patients before recruitment in the study.

Biological parameters

After a fasting period of 12 hr, blood was collected for a complete blood count and HbA1c, HBsAg, anti-HCV, ANA, SMA, IgG, IgM, creatinine, fasting blood glucose, uric acid, serum alkaline phosphatase, total, HDL, low-density lipoprotein (LDL) and very low density lipoprotein (VLDL) cholesterol, total serum protein, ferritin, TSH, albumin, gamma globulin, and International Normalized Ratio determination in the central autoanalyzer of our hospital.

Proprietary scores

For the prediction of HS we calculated SteatoTest,¹⁷ a proprietary score based on serum total bilirubin, gamma-glutamyl transpeptidase (GGT), alpha 2 macroglobulin, haptoglobin, apolipoprotein A1 (ApoA1), fasting glucose, triglycerides, total cholesterol, alanine transaminase (ALT) based on participants sex, age, height, and weight (Table 1). Hepatic necroinflammatory activity and NASH were determined by ActiTest²¹ and NASH-2 test,²¹ respectively, as parts of the FibroMax™ ²¹ (BioPredictive S.A.S., Paris, France). Measurement of the ActiTest was based on serum total bilirubin, GGT, alpha 2 macroglobulin, haptoglobin, ApoA1, and ALT, and the NASH-2 test on serum alpha 2 macroglobulin, ApoA1, haptoglobin, total bilirubin, GGT, aspartate aminotransferase (AST), total cholesterol, and triglycerides.

Table 1.

Interpretation of Proprietary Scores (BioPredictive)

Score (s)	Grade	Interpretation
SteatoTest
0.00 < s < 0.30	S0	No steatosis (<1%)
0.31 < s < 0.38	S0–S1	No steatosis (<1%)
0.39 < s < 0.48	S1	Minimal steatosis (1%–5%)
0.49 < s < 0.57	S1–S2	Minimal steatosis (1%–5%)
0.58 < s < 0.67	S2	Significant steatosis (6%–32%)
0.68 < s < 0.69	S2–S3	Significant steatosis (6%–32%)
0.70 < s < 1.00	S3	Severe steatosis (>32%)

Score (a)	Grade	Interpretation
ActiTest
0.00 < a <0.17	A0	No necroinflammatory activity
0.18 < a <0.52	A1	Minimal necroinflammatory activity
0.53 < a <0.62	A2	Significant necroinflammatory activity
0.63 < a <1.00	A3	Severe necroinflammatory activity

Score (n)	Grade	Interpretation
NashTest 2
0.00 < n < 0.25	N0	No NASH
0.26 < n < 0.50	N1	Mild NASH
0.51 < n < 0.75	N2	Moderate NASH
0.76 < n < 1.00	N3	Severe NASH

Score (f)	Stage	Interpretation
FibroTest
0.00 < f < 0.21	F0	No fibrosis
0.22 < f < 0.27	F0–F1	No fibrosis
0.28 < f < 0.31	F1	Minimal fibrosis
0.32 < f < 0.48	F1–F2	Minimal fibrosis
0.49 < f < 0.58	F2	Moderate fibrosis
0.59 < f < 0.72	F3	Advanced fibrosis
0.73 < f < 0.74	F3–F4	Advanced fibrosis
0.75 < f < 1.00	F4	Severe fibrosis (Cirrhosis)

NASH, nonalcoholic steatohepatitis.

Biochemical analyses were also performed in Biomedicine SA (Athens, Greece) and the respective scores with their equivalence to liver biopsies, were provided by the BioPredictive SAS. The ActiTest was reported as A0–A3²¹ (absent, minimal, significant, and severe necroinflammatory activity), and the NASH-2 test as N0–N3¹⁷ (absent, mild, moderate, and severe NASH).

For the prediction of advanced fibrosis, we calculated FibroTest²¹ another proprietary score based on serum alpha 2 macroglobulin, ApoA1, haptoglobin, total bilirubin, and GGT (Table 1).

For determination of SteatoTest, ActiTest, NashTest 2, and FibroTest (BioPredictive algorithms) samples were blindly provided to Biomedicine SA Athens to measure haptoglobin, alpha 2 macroglobulin, Apo1, total bilirubin, GGT, AST, ALT, triglycerides, total cholesterol, and calculate the above noninvasive biomarkers.

Liver fat content

The liver fat content was estimated with the MRI PDFF technique^22,23 using the IQ IDEAL (IDEAL-IQ General Electric Healthcare, Waukesha, WI), a Gradient Multi-Echo Acquisition Sequence with six echos. The examination was performed at the MRI Section of Biomedicine. A single experienced observer measured the fat fraction value for each hepatic segment and the total hepatic fat fraction was calculated as the average value of all segments. The scan protocol also included a 3D LAVA (GE) sequence. Images from that sequence were reformed on axial and coronal planes of 10 mm slices, with no gap. Using the GE ReportCard software, the liver volume was measured for each plane, excluding big vessels and other parenchymal lesions. From the average of each plane volume, the total liver volume was calculated in cm³.

HS was defined as the percentage of total liver fat divided by the respective liver volume. A fat fraction of <5% was considered normal (no steatosis). Higher percentages were reported as grades 1–3 (G1 mild steatosis, 5%–33%; G2 moderate steatosis, 34%–67%; G3 severe steatosis, ≥67%).

DNA preparation and SNP genotyping

Testing for PNPLA3 single nucleotide polymorphism included DNA extraction from 200 μL of whole blood. Genomic DNA was extracted using the MagNa Pure LC DNA Isolation Kit (Roche Diagnostics, GmbH Mannheim, Germany) applying magnetic bead isolation technology on the MagNa Pure LC automated extraction instrument (Roche Diagnostics).

High-purity genomic DNA was measured by NanoDrop 1000 quantitation system and 40 ng of the extracted DNA was amplified by conventional polymerase chain reaction assay (Primer set: rs738409_F CCC-TGC-TCA-CTT-GGA-GAA-AG and rs738409_R CTG-CAG-GCA-GGA-GAT-GTG-T) on BIORAD c1000 Touch thermocycler. The 227 bp polymerase chain reaction product was tested using restriction fragment length ploymorphism (RFLP) digestion using the NIAIII enzyme cutting the Inline image sequence according the manufacturer's reaction protocol (New England Biolabs, Hitchin, United Kingdom).

The RFLP digestion pattern was the following: CC allele: 227 bp, GG allele: 112,115 bp. CG allele: 227,115, 112 bp. Some of the results were confirmed by sequencing, using the BigDye Terminator v3.1 Cycle Sequencing Kit and analyzed on the ABI3500 automated Genetic Analyzer. Both techniques are complementary to identify and verify the existence or not of the rs738409 PNPLA3 polymorphism.

Statistical analysis

Data were analyzed using the statistical package SPSS version 25 (IBM Corp. Armonk, NY). The Kolmogorov–Smirnov test was used to assess normality of distribution for all continuous variables. The variables are not normally distributed and nonparametric test was chosen.

To describe the sample, descriptive statistics such as median, interquartile range, absolute number, and percentages were performed. Variables were compared by the Mann–Whitney or the Kruskal–Wallis tests, as appropriate. Categorical variables were reported as frequencies and percentages and compared by χ² test or the Fisher's test. Continuous variables were correlated using Spearman's testing. Multiple linear regression analysis using the Backward method was conducted to investigate the associated factors with HS. In regression, all variables were log-transformed. A level of significance of <0.05 was used in all instances.

Results

Of the 140 people recruited, 120 were actually included in our study; 20 participants who had not had MRI PDFF were excluded. Table 2 shows the demographic, clinical, and biochemical characteristics of the cohort. Table 3 shows participants' same characteristics as per MRI liver steatosis grading, as well.

Table 2.

Demographical, Clinical, and Biological Features of Subjects with Type 2 Diabetes Mellitus Included in the Study

Variables	Median (IQR) or N (%)
N	120
Age	61.7 (56.91–65.83)
Sex
Male	74 (61.70)
Female	46 (38.30)
Duration of diabetes, years	10 (5.00–16.75)
BMI	30.5 (27.30–33.47)
Waist circumference, cm
Male	107.50 (95.75–116.00)
Female	102 (96.00–104.25)
Systolic blood pressure, mmHg	125 (120–130)
Diastolic blood pressure, mmHg	60 (60–70)
HbA1c, %	6.70 (6.20–7.23)
Glucose, mg/dL	121 (106.00–143.00)
AST, U/L	19 (15.00–24.00)
ALT, U/L	21 (16.00–34.00)
TSH, mg/dL	1.40 (1.07–2.18)
Total cholesterol, mg/dL	155 (132.00–176.00)
HDL cholesterol, mg/dL
Male	42.50 (36.00–50.00)
Female	49 (42.50–56.50)
LDL cholesterol, mg/dL	87.50 (74.00–103.00)
VLDL cholesterol, mg/dL	24 (18.00–33.75)
Liver fat content >5% (MRI PDFF)	97 (80.80)
Triglycerides, mg/dL	116.50 (93.00–165.50)
SteatoTest grading
S0	25 (21.55)
S0–S1	13 (11.20)
S1	16 (13.79)
S1–S2	24 (20.69)
S2	14 (12.07)
S2–S3	5 (4.31)
S3	19 (16.38)
NashTest2 grading
N0	19 (17.92)
N1	48 (45.28)
N2	33 (31.13)
N3	6 (5.66)
ActiTest grading
A0	89 (76.72)
A1	24 (20.69)
A2	1 (0.86)
A3	2 (1.72)
FibroTest grading
F0	48 (41.37)
F0–F1	15 (12.93)
F1	9 (7.76)
F1–F2	25 (21.55)
F2	4 (3.44)
F3	11 (9.48)
F3–F4	2 (1.72)
F4	2 (1.72)

ALT, alanine transaminase; AST, aspartate aminotransferase; BMI, body mass index; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MRI PDFF, magnetic resonance imaging proton density fat fraction; TSH, thyroid-stimulating hormone; VLDL, very low density lipoprotein.

Table 3.

Participants' Characteristics as per Magnetic Resonance Imaging Liver Steatosis Grading

Variables	Median (IQR) or N (%)
Variables	≤5%	5%–33%	34%–66%	>66%
N	23	64	19	14
Age	63 (61–67.24)	61.16 (56.80–65.64)	64.36 (55.16–68.27)	59.07 (50.61–61.85)
Sex
Male	18 (78.3)	36 (56.3)	13 (68.4)	7 (50)
Female	5 (21.70)	28 (43.8)	6 (31.6)	7 (50)
Duration of diabetes, years	12 (6–19)	10 (5–17)	8 (3–10)	8 (1.75–10.5)
BMI	26.6 (25–30.80)	30.9 (27.87–34.77)	31 (28.75–32)	33 (29.24–38.65)
Waist circumference, cm
Male	95 (90–109.25)	109.5 (98–115.5)	111 (103–121)	120 (108–127)
Female	95 (82.5–103)	102 (96–107.25)	97 (96–100.25)	103 (100–104)
Systolic blood pressure, mmHg	120 (110–130)	130 (120–130)	125 (120–130)	120 (120–130)
Diastolic blood pressure, mmHg	65 (60–70)	60 (60–70)	60 (60–80)	60 (58.75–70)
HbA1c, %	6.6 (6.20–7.23)	6.6 (6.2–7.3)	6.8 (6.3–7.3)	6.79 (6.24–6.93)
Glucose, mg/dL	122 (108.75–147.75)	121 (102.5–142.75)	127 (106–144)	128 (107.75–143.−75)
AST, U/L	17 (14–20)	18 (15–23)	20 (16–23)	31 (21.75–40)
ALT, U/L	14.5 (12–19.25)	20 (17–29)	24 (19–38)	48.5 (32.5–70.75)
TSH, mg/dL	1.57 (1.09–18.6)	1.41 (10.7–2.19)	1.2 (0.8–1.99)	2 (1.12–2.77)
Total cholesterol, mg/dL	147.5 (129.25–167.75)	157 (136.25–174.5)	153 (132–169)	177.5 (126–189.5)
HDL cholesterol, mg/dL
Male	48 (37–53.75)	41.5 (35.2–49)	40 (35–49.5)	41 (35–46)
Female	65 (50–75.5)	51 (43.2–57.5)	43 (38.75–55)	44 (39–50)
LDL cholesterol, mg/dL	85 (71–97)	90.5 (76.5–108)	82 (62–94)	91.5 (63.5–114.25)
VLDL cholesterol, mg/dL	21 (15–33)	24 (18–32.75)	24 (17–48)	32.5 (19.7–41.5)
Triglycerides, mg/dL	106 (74–127)	116 (92.5–161.75)	109 (97–193)	141 (117.75–218)
SteatoTest grading
S0	8 (40)	14 (22.2)	2 (10.5)	1 (7.1)
S0–S1	6 (30)	3 (4.8)	4 (21.1)	0
S1	2 (10)	12 (19)	2 (10.5)	0
S1–S2	1 (5)	14 (22.2)	4 (21.1)	5 (35.7)
S2	2 (10)	9 (14.3)	2 (10.5)	1 (7.1)
S2–S3	0	3 (4.8)	1 (5.3)	1 (7.1)
S3	0	8 (12.7)	4 (21.1)	6 (42.9)
NashTest2 grading
N0	8 (44.4)	9 (15.5)	2 (11.8)	0
N1	5 (27.8)	29 (50)	9 (52.9)	5 (38.5)
N2	5 (27.8)	18 (31)	5 (29.4)	5 (38.5)
N3	0	2 (3.4)	1 (5.9)	3 (23.1)
ActiTest grading
A0	18 (78.3)	53 (84.1)	14 (73.7)	4 (28.6)
A1	2 (8.7)	8 (12.7)	5 (26.3)	9 (64.3)
A2	0	0	0	1 (7.1)
A3	0	2 (3.2)	0	0
FibroTest grading
F0	5 (25)	29 (46)	7 (36.8)	7 (50)
F0–F1	5 (25)	7 (11.1)	2 (10.5)	1 (7.1)
F1	2 (10)	2 (3.2)	3 (15.8)	2 (14.3)
F1–F2	8 (40)	13 (20.6)	3 (15.8)	1 (7.1)
F2	0	3 (4.8)	0	1 (7.1)
F3	0	6 (9.5)	3 (15.8)	2 (14.3)
F3–F4	0	2 (3.2)	0	0
F4	0	1 (1.6)	1 (5.3)	0

In this population, 8 participants (6.66%) were diet controlled and 31 (25.83%) were treated with insulin on top of oral medication. One hundred twelve participants (93.33%) were treated with metformin, and the most popular treatment combination was metformin with dipeptidyl peptidase-4 inhibitors, 42 participants (35.00%). The percentage of participants treated with statins was 62.50% (n = 75), whereas 10.00% (n = 12) were treated with statins and fibrates.

The PNPLA3 rs738409 CC/CG/GG genotype frequencies were 54.16% (n = 65), 35.00% (n = 42), and 10.83% (n = 13), respectively. The median liver burden of HS by MRI PDFF was 7.80 (range 4.1–14.5). Based on MRI PDDF only 23 participants (19.16%) had <5% of predicted hepatocytes loaded with fat characterized as grade 0 steatosis, whereas 97 participants (80.83%) had abnormal levels of steatosis. In 64 (53.33%) of the later, the predicted histological steatosis was grade 1 (5%–33%); 19 (15.83%) participants had 34%–66% of predicted hepatocytes loaded with fat (grade 2) and 14 (11.67%) had grade 3 steatosis (i.e., >66% of their hepatocytes predicted to be steatotic).

To identify whether there was statistically significant difference between MRI PDFF and SteatoTest grading, participants with SteatoTest grades S0–S1 (Table 1) were grouped together as in MRI PDFF grade 0 is defined as <5% of hepatocytes with steatosis (LBx). No statistically significant difference was found between MRI PDFF Steatosis Grading and SteatoTest Steatosis Grading (P = 0.077). The kappa coefficient was 0.062, which represents no agreement.

Moderate-to-severe NASH (stage N2 and above) was observed in 39 participants (36.79%) (Table 2).

Moderate-to-severe liver fibrosis (stage F2 and above) was observed in 16.36% of participants (Table 2). Subjects with moderate-to-severe fibrosis in comparison to subjects with no evidence or minimal fibrosis had higher HbA1c (6.99% vs. 6.65%) and ALT >38 U/L (47.1% vs. 14.1%), and they were mostly heterozygous for the G allele of PNPLA3 rs738409 (58.8% vs. 29.3%).

As far as the prevalence of diabetic macrovascular and microvascular complications is concerned, only two patients (1.7%) suffered from microvascular and one patient (0.8%) from macrovascular complications.

In a univariate analysis, HS was directly correlated with ActiTest (r = 0.507, P < 0.001), NashTest2 (r = 0.532, P < 0.001), VLDL cholesterol (r = 0.219, P = 0.016), and triglycerides (r = 0.262, P = 0.004). HS was also directly correlated with BMI (r = 0.379, P < 0.001), waist circumference (r = 0.290, P < 0.001), metabolic syndrome (r = 0.199, P = 0.029), AST (r = 0.377, P < 0.001), ALT (r = 0.594, P < 0.001), GGT (r = 0.449, P < 0.001), and the G-allele of PNPLA3 rs738409 (CC vs. CG/GG, P = 0.001) and inversely but significantly correlated with age (r = −0.286, P = 0.002) and duration of diabetes (r = −0.277, P = 0.002). Furthermore, a statistically significant difference was found between sex and HS; the median of HS for male sex was 9.1 and for female sex 7.2 (P = 0.008).

No association was found with FibroTest (P = 0.878), HDL (P = 0.099), LDL (P = 0.614), total cholesterol (P = 0.390), ApoA1 (P = 0.861), HbA1c (P = 0.781), and uric acid (P = 0.575).

In a multivariate analysis, the predictive variables for liver steatosis were the PNPLA3 rs738409 variant, waist circumference, female sex, and duration of diabetes (Table 4).

Table 4.

Multivariate Linear Regression for Hepatic Steatosis by Magnetic Resonance Imaging Proton Density Fat Fraction

Variables	Beta coefficient	95% CI	P	Correlations
Variables	Beta coefficient	95% CI	P	Zero-order	Partial	Part
PNPLA3 rs738409 (CG/GG)	0.425	0.182 to 0.667	0.001	0.328	0.308	0.267
Waist circumference, cm	2.448	1.334 to 3.563	0.001	0.267	0.235	0.218
Female sex	0.419	0.164 to 0.674	0.002	0.266	0.310	0.269
Duration of diabetes, years	−0.179	−0.316 to −0.042	0.011	−0.311	−0.281	−0.242

CI, confidence interval.

Discussion

There is a complex association between NAFLD and type 2 diabetes with each condition affecting negatively the other one.²⁴ Specifically, not only the presence of NAFLD is related to an increased risk of developing type 2 diabetes,^25
–27 but also subjects with type 2 diabetes and NAFLD are at risk of progressing faster to severe forms of NAFLD such as advanced fibrosis, cirrhosis, and hepatocellular carcinoma.^28

–33 Recently, it was observed by liver biopsy that steatohepatitis represents the sole feature of liver damage in type 2 diabetes, suggesting that NASH may be one of the early complications of type 2 diabetes due to its pathophysiological correlation with insulin resistance.³⁴

Findings from this study confirm the association, among lipid profile parameters examined, of specifically serum triglycerides and VLDL cholesterol with HS in keeping with other studies.²⁷

However, in our study the predictive variables for HS were mostly related to our population's demographic characteristics. A predominant association between women, but not with men, and HS was identified. This may be explained based on the fact that estrogen levels are different in men and women, with estrogen levels playing a critical role in lipid metabolism.²⁸

Furthermore, numerous studies have discussed the relationship between waist circumference and the increased risk of developing fatty liver. Ghaemi et al.²⁹ showed that waist circumference is a mediator of dietary pattern in NAFLD. In a meta-analysis study, Pang et al. examined 20 studies and concluded that waist circumference is the strongest anthropometric variable in predicting fatty liver (OR = 3.14, 95% CI: 2.07 to 4.77), an even stronger variable than BMI, as it was shown in our study as well. The same study showed superiority of waist circumference over BMI in predicting fatty liver, which was also confirmed in our study. This may be due to existing evidence that insulin resistance in NAFLD is more related to waist girth than BMI, and fatty liver is a liver manifestation of metabolic syndrome.³⁰

Furthermore, BMI is a composite variable consisting of two parameters (weight and height²), which could easily vary with weight changes, making it seem less accurate in predicting NAFLD.

In univariate analysis, participants with definite steatosis were younger and had lesser duration of diabetes. In multivariate analysis, duration of diabetes was also identified as an independent predictor of HS. Although most studies have shown that the prevalence of NAFLD increases with age, we did not find such evidence in our study population. Our findings are in accordance with Williamson et al.'s previous study.¹¹

First of all, it is possible that the prevalence of NAFLD in our study population to be greater than the general diabetic population. The results may be less applicable to other countries in which the prevalence of other causes of liver disease such as alcohol liver disease and viral hepatitis is markedly different. Second, in our study, we showed that age was negatively but significantly correlated with BMI (P = −0.029), which means that our participants as they are getting older, they might reduce their calorie intake and lose weight, the basic risk factor for both diabetes and NAFLD.

Furthermore, 93.33% of our participants were on metformin, which is known from previous studies³¹ to have a positive or neutral effect in individuals. As far as the association of shorter duration of diabetes and liver disease is concerned, it has been described in detail in the past.³² One possible explanation is that the significantly increased level of hyperinsulinemia in early type 2 diabetes drives uptake of free fatty acids by hepatocytes.

Findings from our study also confirm an association of the PNPLA3 rs738409 with liver fat content. In keeping with prior studies examining the relationship between the above genetic variant with HS, we found the minor allele to confer risk for increased fat deposition in the liver.¹⁴ The specific polymorphism could be useful to identify people needed specific strategies, first to prevent, and second to detect liver fibrosis and its complications, especially based on the fact that transaminase levels underestimate NAFLD in people with type 2 diabetes^33,35 (in our study the transaminase levels of our population not even doubled the upper reference range as shown in Table 2).

Several limitations of the current study need to be considered. First, it is a cross-sectional study. Second, the number of participants was fairly small so our results cannot be considered as definite and should be confirmed in large cohorts.

Finally, we used FibroMax as a diagnostic tool composed of the combination of four noninvasive biomarker panels for the diagnosis of steatosis (SteatoTest), necrosis, and inflammation (ActiTest and NashTest 2) and fibrosis (FibroTest). All these biomarkers were originally used for populations with various liver conditions (alcohol liver disease, hepatitis B and C),^35
–37 but they have also been validated for patients with NAFLD.²⁶ However, these studies only included 30% of participants with type 2 diabetes and the performances of the biomarker panels were not specifically assessed in this subgroup of participants. Bril and Cusi³⁸ in a recent study showed that these specific noninvasive panels underperformed when applied to a large cohort of people with type 2 diabetes.

Because of lack of liver biopsies, we are also aware that our results should be interpreted with caution and be confirmed with studies, including histological evaluation in the future.

Conclusions

In conclusion, this study confirms the clinical and lipid profile of people with type 2 diabetes at risk of developing NAFLD. Furthermore, it suggests that PNPLA3 rs738409 is a major determinant of liver steatosis in individuals with type 2 diabetes; further studies should confirm the risk of this particular group of individuals to develop liver fibrosis and establish PNPLA3 rs738409 as a useful screening tool. Finally, although well-validated biomarker panels for the diagnosis of NAFLD are quite promising, people with type 2 diabetes may require predictive models that have been specifically developed for them as extrapolation of results from population with no diabetes may result in significant misclassification.

Footnotes

Authors' Contributions

A.M. conceived of the article, wrote the article, and researched data; E.M. conceived of the article, reviewed, and edited the article; P.K. analyzed MRI PDFF data; S.R. performed genetic analysis; D.L. performed statistical analysis; E.T. performed genetic analysis; H.M. reviewed the article; S.M. reviewed and edited the article; D.M. reviewed and edited the article; A.T. conceived of the article, wrote, reviewed, and edited the article; A.M. and A.T. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Author Disclosure Statement

No conflicting financial interests exist.

Funding Information

No funding was received for this article.

References

Chalasani

, Younossi

, Lavine

, et al. The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases: Hepatology, Vol. XX, No. X, 2017. Hepatology, 2018; 67:328–357.

Younossi

, Koenig

, Abdelatif

, et al. Global epidemiology of nonalcoholic fatty liver disease—Meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology, 2016; 64:73–84.

Cozzolino

, Sessa

, Salvatore

, et al. The involvement of the opioid system in human obesity: A study in normal weight relatives of obese people. J Clin Endocrinol Metab, 1996; 81:713–718.

Adinolfi

, Petta

, Fracanzani

, et al. Reduced incidence of type 2 diabetes in patients with chronic hepatitis C virus infection cleared by direct-acting antiviral therapy: A prospective study. Diabetes Obes Metab, 2020; 22:2408–2416.

Sasso

, Pafundi

, Caturano

, et al. Impact of direct acting antivirals (DAAs) on cardiovascular events in HCV cohort with pre-diabetes. Nutr Metab Cardiovasc Dis, 2021; 31:2345–2353.

Rinaldi

, Pafundi

, Galiero

, et al. Mechanisms of non-alcoholic fatty liver disease in the metabolic syndrome. A narrative review. Antioxidants (Basel), 2021; 10:270.

Caturano

, Aciemo

, Nevola

, et al. Non-alcoholic fatty liver disease: From pathogenesis to clinical impact. Processes, 2021; 9:135.

Marchesini

, Brizi

, Bianchi

, et al. Nonalcoholic fatty liver disease: A feature of the metabolic syndrome. Diabetes, 2001; 50:1844–1850.

Hossain

, Afendy

, Stepanova

, et al. Independent predictors of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol, 2009; 7:1224–1229, 1229.e1–e2.

10.

Bril

, Cusi

. Management of nonalcoholic fatty liver disease in patients with type 2 diabetes: A call to action. Diabetes Care, 2017; 40:419–430.

11.

Williamson

, Price

, Glancy

, et al. Prevalence of and risk factors for hepatic steatosis and nonalcoholic Fatty liver disease in people with type 2 diabetes: The Edinburgh Type 2 Diabetes Study. Diabetes Care, 2011; 34:1139–1144.

12.

El-Serag

, Tran

, Everhart

. Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology, 2004; 126:460–468.

13.

Dzierlenga

, Clarke

, Hargraves

, et al. Mechanistic basis of altered morphine disposition in nonalcoholic steatohepatitis. J Pharmacol Exp Ther, 2015; 352:462–470.

14.

Romeo

, Kozlitina

, Xing

, et al. Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease. Nat Genet, 2008; 40:1461–1465.

15.

Liu

Y-L

, Patman

, Leathart

JBS

, et al. Carriage of the PNPLA3 rs738409 C>G polymorphism confers an increased risk of non-alcoholic fatty liver disease associated hepatocellular carcinoma. J Hepatol, 2014; 61:75–81.

16.

Bruschi

, Claudel

, Tardelli

, et al. The PNPLA3 I148M variant modulates the fibrogenic phenotype of human hepatic stellate cells. Hepatology, 2017; 65:1875–1890.

17.

Pirazzi

, Valenti

, Motta

, et al. PNPLA3 has retinyl-palmitate lipase activity in human hepatic stellate cells. Hum Mol Genet, 2014; 23:4077–4085.

18.

Sookoian

, Castano

, Burgueno

, et al. A nonsynonymous gene variant in the adiponutrin gene is associated with non-alcoholic fatty liver disease severity. J Lipid Res, 2009; 50:2111–2116.

19.

Liu

, Patman

, Leathart

, et al. Carriage of PNPLA3 I148M is associated with increased risk of non-alcoholic fatty liver disease-associated hepatocellular carcinoma. J Hepatol, 2013; 58:S516.

20.

American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care, 2014; 37:S81–S90.

21.

Poynard

, Lassailly

, Diaz

, et al. Performance of biomarkers FibroTest, ActiTest, SteatoTest and NasTest in patients with severe obesity: Meta analysis of individual patient data. PLoS One, 2012; 7:e30325.

22.

Heba

, Desai

, Zand

, et al. Accuracy and the effect of possible subject-based confounders of magnitude-based MRI for estimating hepatic proton density fat fraction in adults, using MR spectroscopy as reference. J Magn Reson Imaging, 2016; 43:398–406.

23.

Doycheva

, Cui

, Nguyen

, et al. Non-invasive screening of diabetics in primary care for NAFLD and advanced fibrosis by MRI and MRE. Aliment Pharmacol Ther, 2016; 43:83–95.

24.

Adams

, Harmsen

, Sauver

, et al. Nonalcoholic fatty liver disease increases risk of death among patients with diabetes: A community-based cohort study. Am J Gastroenterol, 2010; 105:1567–1573.

25.

De Marco

, Locatelli

, Zoppini

, et al. Cause-specific mortality in type 2 diabetes. The Verona Diabetes Study. Diabetes Care, 1999; 22:756–761

26.

Angulo

, Keach

, Batts

, et al. Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. Hepatology, 1999; 30:1356–1362.

27.

Toledo

FGS

, Sniderman

, Kelley

. Influence of hepatic steatosis (fatty liver) on severity and composition of dyslipidemia in type 2 diabetes. Diabetes Care, 2006; 29:1845–1850.

28.

Chen

L-Z

, Xin

Y-N

, Geng

, et al. PNPLA3 I148M variant in nonalcoholic fatty liver disease: Demographic and ethnic characteristics and the role of the variant in nonalcoholic fatty liver fibrosis. World J Gastroenterol, 2015; 21:794.

29.

Ghaemi

, Hosseini

, Osati

, et al. Waist circumference is a mediator of dietary pattern in Non-alcoholic fatty liver disease. Sci Rep, 2018; 8:4788.

30.

Pang

, Zhang

J-Y

, Song

S-D

, et al. Central obesity and nonalcoholic fatty liver disease risk after adjusting for body mass index. World J Gastroenterol, 2015; 21:1650–1662.

31.

Bugianesi

, Gentilcore

, Manini

, et al. A randomized controlled trial of metformin versus vitamin E or prescriptive diet in nonalcoholic fatty liver disease. Am J Gastroenterol, 2005; 100:1082–1090.

32.

Lewis

, Carpentier

, Adeli

, et al. Disordered fat storage and mobilization in the pathogenesis of insulin resistance and type 2 diabetes, Endocr Rev, 2002; 23:201–299.

33.

Kotronen

, Juurinen

, Hakkarainen

, et al. Liver fat is increased in type 2 diabetic patients and underestimated by serum alanine aminotransferase compared with equally obese nondiabetic subjects. Diabetes Care, 2008; 31:165–169.

34.

Massarone

, Rosato

, Bucci

, et al. Liver biopsy in type 2 diabetes mellitus: Steatohepatitis represents the sole feature of liver damage. PLoS One, 2017; 12:e0178473.

35.

Myers

, Tainturier

, Ratziu

, et al. Prediction of liver histological lesions with biochemical markers in patients with chronic hepatitis B. J Hepatol, 2003; 39:222–230.

36.

Naveau

, Gaudé G, Asnacios A, et al. Diagnostic and prognostic values of noninvasive biomarkers of fibrosis in patients with alcoholic liver disease. Hepatology, 2009; 49:97–105.

37.

Ratziu

, Massard

, Charlotte

, et al. Diagnostic value of biochemical markers (FibroTest-FibroSURE) for the prediction of liver fibrosis in patients with non-alcoholic fatty liver disease. BMC Gastroenterol, 2006; 6:6.

38.

Bril

, McPhaul

, Caulfield

, et al. Performance of the SteatoTest, ActiTest, NashTest and FibroTest in a multiethnic cohort of patients with type 2 diabetes mellitus. J Investig Med, 2019; 67:303–311.