Abstract
Background
While non-alcoholic fatty liver disease (NAFLD) has been well characterised in patients with diabetes mellitus (DM), less is known about NAFLD in non-DM patients. We investigated the clinical characteristics of NAFLD patients with and without DM and accuracy of the NAFLD fibrosis score (NFS) in these two NAFLD groups.
Methods
Clinical, biochemical and histological variables were evaluated in this prospective cross-sectional study of 503 patients with biopsy proven NAFLD. Comparisons between patients with and without DM were analysed. NFS was correlated with liver histology to assess its robustness in patients with and without DM.
Results
There were 503 biopsy proven NAFLD patients with 48% of the cohort being diabetic. Relative to patients without DM, patients with DM were older (52 vs. 46 years, p < 0.001), with higher proportion of females (70% vs. 54%, p < 0.001), higher BMI (37 vs. 35, p = 0.009), higher prevalence of hypertension (73% vs. 44%, p < 0.001), higher prevalence of NASH (80.2% vs. 64.4%; p < 0.001) and advanced fibrosis (40.3% vs. 17.0%; p < 0.001). A considerable amount of patients without DM still had NASH (64%) and advanced fibrosis (17%). The clinical utility of the NFS differed between NAFLD patients with and without DM, with sensitivity to exclude advanced fibrosis being 90% of NAFLD patients with DM but only 58% of patients without DM.
Conclusion
Patients with DM have more severe NAFLD based on histology. However, NASH and advanced fibrosis also occur in a considerable proportion of NAFLD patients without DM. The lower utility of the NFS in NAFLD patients without DM emphasises the heterogeneous nature of the NAFLD phenotype.
Abbreviations: ALP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ACE-I, angiotensin-converting enzyme-inhibitor; ARB, angiotensin receptor blocker; apoB-100, apolipoprotein B-100; BMI, body mass index; CIs, confidence intervals; Chol, total cholesterol; DM, type 2 diabetes mellitus; ER, endoplasmic reticulum; FFAs, free-fatty acids; HDL, high density lipoprotein cholesterol; HOMA-IR, Homeostatic model assessment—insulin resistance; INR, international normalised ratio; LDL, low density lipoprotein cholesterol; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; NFS, NAFLD fibrosis score; NASH CRN, Non-alcoholic Steatohepatitis Clinical Research Network; NAS, NAFLD activity score; ORs, odd ratios; SDs, standard deviations; TGs, triglycerides; VLDL, very-low-density lipoproteins
Keywords: NAFLD, Diabetic, Non-diabetic, NAFLD fibrosis score
Highlights
-
•
Patients with diabetes have more severe NAFLD based on histology.
-
•
Severe NAFLD can occur in a considerable proportion of non-diabetic NAFLD patients.
-
•
The NAFLD fibrosis score may be less accurate in non-diabetics.
1. Introduction
Non-alcoholic fatty liver disease (NAFLD), the hepatic manifestation of metabolic syndrome [1], represents a spectrum of histopathologic abnormalities ranging from simple steatosis to the more aggressive non-alcoholic steatohepatitis (NASH), characterised by steatosis, parenchymal inflammation, hepatocellular ballooning and other evidence of hepatic injury [2]. Patients with NASH are at risk of developing progressive fibrosis; reported in up to 50% of cases over 6 years [3]. There is increasing recognition that NAFLD is a heterogeneous disease with multiple pathways of pathogenesis and patients with different phenotypes of NAFLD can present with diverse disease manifestations [4]. Insulin resistance plays a dominant role in the pathogenesis of NAFLD [5]. Patients with type 2 diabetes mellitus (DM) have an increased risk of developing NAFLD, NASH and hepatic fibrosis/cirrhosis [6], [7], [8], [9]. Furthermore, NAFLD patients with DM have three times the mortality compared to non-diabetic NAFLD patients [10]. The importance of DM in NAFLD is reflected by its inclusion in the majority of the non-invasive composite predictive scores for NASH and advanced fibrosis [11], [12], [13], [14]. One such composite predictive score for predicting advanced fibrosis in NAFLD is the NAFLD fibrosis score (NFS), which has been validated and recommended for use in the American society guidelines [2], [14]. Reiterating disease heterogeneity and that NAFLD may not conform to a “one size fits all approach”, McPherson and colleagues had reported a difference in the reliability of NFS in the context of normal and abnormal ALT levels [15]. Other non-invasive fibrosis scores such as the BARD score and AST/ALT ratio have also been used to predict advanced fibrosis in NAFLD. We sought to characterise the clinical spectrum of NAFLD in patients with and without DM. In addition, we explored the utility of NFS and other established non-invasive fibrosis scores among these two groups.
2. Materials and methods
2.1. Study design and population
This is a prospective cross sectional study with patients enrolled from two hepatology outpatient clinics in Cleveland, Ohio (Cleveland Clinic and MetroHealth Medical Center). Study received approval from the institutional review board.
The study included patients 18 years of age and over, with histologically proven NAFLD, who had not received any prior therapies that may have been beneficial for NAFLD, such as Vitamin E, pentoxifylline, pioglitazone and prescribed diet & exercise weight loss programmes. Patients with excessive alcohol consumption (> 21 drinks per week and > 14 drinks per week for males and females respectively) were excluded. Similarly, patients with other contributory causes of liver disease including those with hepatotoxic drug history, viral hepatitis, hemochromatosis, autoimmune hepatitis, Wilson's disease or alpha 1 antitrypsin disease were excluded.
2.2. Ascertainment of clinical data
Demographic and clinical information was obtained by two of the authors (SD or AM) for all patients from an electronic medical record system that is common to both hospitals. The diagnosis of DM was diagnosed based on American Diabetes Association (ADA) criteria with or without the use of antidiabetic medications [16]. Hypertension was diagnosed by the Joint National Committee (JNC) 7 criteria [17]. All diagnoses were verified based on documentation in the electronic medical records by one of the investigators (SD or AM). Body mass index (BMI) were collated, as were liver function tests [serum albumin, bilirubin, alkaline phosphatase (ALP), aspartate aminotransferase (AST), alanine aminotransferase (ALT)], platelet count, international normalised ratio (INR), total cholesterol (Chol), triglycerides (TGs), high density lipoprotein (HDL) cholesterol, low density lipoprotein (LDL) cholesterol, and HbA1C. To minimise inter-observer variability, liver biopsy specimens were read using standardised well defined histological criteria [18] established by the Non-alcoholic Steatohepatitis Clinical Research Network (NASH CRN). As recommended, the diagnosis of NASH was based on the overall impression of histopathologists and not the NAFLD activity score [NAS] [19]. Fibrosis was classified into 4 stages with advanced fibrosis defined as stage 3–4 fibrosis (bridging fibrosis–cirrhosis). Only clinical variables obtained within 6 months of the liver biopsy were included in the analysis. The use of statins and angiotensin-converting enzyme-inhibitor (ACE-I)/angiotensin receptor blocker (ARB) within 6 months prior to the liver biopsy was also examined. Insulin resistance was assessed with Homeostatic model assessment—insulin resistance (HOMA-IR) based on the formula: [fasting glucose (mg/dL) × insulin (μU/mL)] divided by 405 [20]. NFS was calculated according to the published formula; NFS: − 1.675 + 0.037 × age (years) + 0.094 × body mass index (BMI; kg/m2) + 1.13 × impaired fasting glycaemia or DM (yes = 1, no = 0) + 0.99 × AST/ALT ratio − 0.013 × platelet (× 109/L) − 0.66 × albumin (g/dL) [14]. The cut-off points of the NFS used to categorise fibrosis were: <− 1.455, − 1.455 to 0.676, and > 0.676 for low indeterminate and high probability for advanced fibrosis, respectively [14]. Calculation of the BARD score was also performed; the BARD score was a 4 point score derived from the weighted sum of three variables (BMI > 28 = one point, AST/ALT ratio > 0.8 = two points, diabetes = one point) where a score of two or more suggestive of advanced fibrosis [13]. Similarly, AST/ALT ratio more than 0.8 itself has been suggested to be useful in predicting advanced fibrosis [21].
2.3. Statistical analysis
Descriptive statistics was computed for all variables and reported as means and standard deviations (SDs) for continuous variables or frequencies and percentages for categorical variables. Baseline characteristics and differences in demographic, clinical, histological and laboratory indices between patients with and without DM were ascertained, using Student's T tests and Pearson's chi-square testing for continuous and categorical variables, respectively. Mann–Whitney U test was also performed when applicable. The utility of NFS was correlated with histological staging of fibrosis in both DM and non-DM patients. With regard to advanced fibrosis, Spearman's correlation analysis for each individual component of the NFS was performed in both patients with and without DM. In addition, the utility of using fibrosis scores such as the BARD score and AST/ALT ratio was also assessed in both patients with and without DM. All analyses were performed using SPSS version 21 statistical software (Chicago, Illinois, USA). Two sided p values were used. p values < 0.05 were considered statistically significant.
3. Results
Of the 503 NAFLD patients in the data set, 62% were female, 58% had concomitant hypertension, 48% had concomitant DM and the mean BMI was 36.1 kg/m2. NASH and advanced fibrosis were present in 71.8% and 28.1% of the cohort respectively.
Table 1 illustrates the characteristics of the whole cohort and also patients with and without DM. Compared to NAFLD patients without DM, patients with DM were older (52.5 vs. 46.0 years; p < 0.001), were more likely to be female (70.7% vs. 54.0%; p < 0.001) and have a higher BMI (37.2 vs. 35.2 kg/m2; p = 0.009). Diabetic patients were also more likely to be hypertensive (73.2 vs. 43.9%; p < 0.001) with greater use of statins (42.0 vs. 15.2%; p < 0.001) and ACE-I/ARBs (51.9 vs. 28.8%; p < 0.001). However, there was no difference in the prevalence of advanced fibrosis among the DM patients taking or not taking statins (p = 0.182) and also among the DM patients taking or not taking ACE-I/ARB (p = 0.357). There were differences in ALT, albumin, platelet count, TG, LDL and HbA1C between the two groups. There was no difference in aggregated length of liver tissue examined between patients with and without DM (19.9 mm vs. 18.2 mm; p = 0.225). As shown in Table 2, patients with diabetes had more lobular inflammation (p = 0.017), ballooning (p < 0.001) and NASH (p < 0.001). The median NAS was higher in DM patients compared to non-DM patients (p = 0.022). More of the DM patients had grade 2 ballooning (41.7% vs. 24.0%; p < 0.001) and higher prevalence of NASH (80.2% vs. 64.4%; p < 0.001) compared to non-DM patients, while there was no differences in steatosis between patients with and without DM. DM patients also had a higher prevalence of advanced fibrosis (40.3% vs. 17.0%; p < 0.001) [Table 1]. When cirrhosis was considered specifically, significantly more of the DM patients had cirrhosis relative to the non-DM patients (20.6% vs. 5.7%) [Table 2]. Among the DM patients, the duration of DM did not differ between those with and without advanced fibrosis (5.79 vs. 5.42 years, p = 0.791). Similarly, HbA1C levels were comparable between diabetic patients with and without advanced fibrosis (7.5 vs. 7.3, p = 0.414). Among the patients without DM, there was no relationship between insulin resistance and presence of advanced fibrosis (p = 0.732) [Table 3]. Table 4 represents the correlation of NFS to the stage of fibrosis. In the whole cohort of patients, there were 334 patients without advanced fibrosis on histology, of which 26 patients (7.8%) were wrongly classified by the NFS. Hence, the NFS was able to correctly identify 92% of the patients as being without advanced fibrosis. Of the 129 patients with advanced fibrosis, 26 patients (20.2%) were wrongly classified by the NFS. Analysing the utility of the NFS in subgroups stratified by DM, 14.4% (19/132) of DM patients without advanced fibrosis as opposed to 3.5% (7/202) of non-DM patients without advanced fibrosis were wrongly predicted as the high risk NFS category. Of the DM patients with advanced fibrosis 10.1% (9/89) were inaccurately predicted as the low risk NFS category, while 42.5% (17/40) of the non-DM patients with advanced fibrosis were inaccurately predicted as the low risk NFS category. Spearman's correlation analysis of the individual components of NFS to advanced fibrosis in both the patients with and without DM was performed [Table 5]. This showed that in relation to advanced fibrosis, the correlation coefficients of the individual components used to calculate NFS were correspondingly lower in the non-DM patients compared to the DM patients. Other fibrosis scores developed for NAFLD, namely the BARD score and AST to ALT ratio were also tested in our cohort for their utility in patients with and without DM [Table 6]. Similar to our findings with the NFS, there were discrepancies in the utility of the BARD score and AST to ALT ratio between patients with and without DM.
Table 1.
Characteristics of patients with NAFLD by the presence of DM status.
| Characteristics | Total (n = 503) | Presence of DM |
p value⁎ | |
|---|---|---|---|---|
| Yes (n = 239) | No (n = 264) | |||
| Age (years) | 49 ± 12 | 52 ± 10 | 46 ± 12 | < 0.001 |
| Proportion female (%) | 311 (62.0) | 169 (70.7) | 142 (54.0) | < 0.001 |
| BMI (kg/m2) | 36.13 ± 8.43 | 37.16 ± 7.84 | 35.20 ± 8.84 | 0.009 |
| Presence of HTN (%) | 291 (57.9) | 175 (73.2) | 116 (43.9) | < 0.001 |
| Bilirubin (mg/dL) | 0.67 ± 0.39 | 0.68 ± 0.42 | 0.66 ± 0.37 | 0.648 |
| AST (U/L) | 56.15 ± 63.65 | 53.52 ± 36.03 | 58.52 ± 80.85 | 0.388 |
| ALT (U/L) | 71.70 ± 55.60 | 64.18 ± 49.21 | 78.49 ± 60.09 | 0.004 |
| AST/ALT ratio | 0.85 ± 0.37 | 0.92 ± 0.35 | 0.79 ± 0.37 | < 0.001 |
| Albumin (g/dL) | 4.25 ± 0.49 | 4.19 ± 0.53 | 4.30 ± 0.45 | 0.017 |
| INR | 1.03 ± 0.17 | 1.04 ± 0.18 | 1.02 ± 0.16 | 0.262 |
| Platelet count (k/uL) | 238.38 ± 77.68 | 230.85 ± 78.72 | 245.11 ± 76.27 | 0.045 |
| Chol (mg/dL) | 197.74 ± 50.87 | 192.99 ± 53.57 | 201.88 ± 48.13 | 0.068 |
| TG (mg/dL) | 206.7 ± 181.20 | 231.8 ± 223.50 | 184.7 ± 130.20 | 0.007 |
| HDL (mg/dL) | 42.86 ± 10.85 | 41.94 ± 10.72 | 43.69 ± 10.93 | 0.096 |
| LDL (mg/dL) | 122.11 ± 42.86 | 114.6 ± 44.77 | 128.75 ± 40.04 | 0.001 |
| HbA1C (%) | 6.58 ± 1.47 | 7.40 ± 1.51 | 5.62 ± 0.56 | < 0.001 |
| Ferritin (ng/mL) | 231.3 ± 230.90 | 209.1 ± 233.90 | 250.9 ± 227.00 | 0.076 |
| Presence of NASH (%) | 361 (71.8) | 190 (80.2) | 168 (64.4) | < 0.001 |
| Presence of advanced fibrosis (%) | 141 (28.1) | 96 (40.3) | 45 (17.0) | < 0.001 |
| Length of specimen (mm) | 19.0 ± 10.8 | 19.9 ± 13.3 | 18.2 ± 8.0 | 0.225 |
| Use of statin (%) | 140 (27.9) | 100 (42.0) | 40 (15.2) | < 0.001 |
| Use of ACE-I/ARBs (%) | 200 (39.8) | 124 (51.9) | 76 (28.8) | < 0.001 |
Data are expressed as mean ± standard deviation or as number and percentage. BMI: body mass index, HTN: hypertension, AST: aspartate aminotransferase, ALT: alanine aminotransferase, INR: international normalised ratio, Chol: total cholesterol, TG: triglyceride, HDL: high density lipoprotein cholesterol, LDL: low density lipoprotein cholesterol, ACE-I: angiotensin-converting enzyme-inhibitor, ARB: angiotensin receptor blocker, DM: diabetes mellitus.
p values were derived using Student's T tests and Pearson's chi-square tests on continuous and categorical variables respectively.
Table 2.
Comparison of liver histology between DM and non-DM patients.
| Histology feature | Presence of DM |
p value⁎ | |
|---|---|---|---|
| Yes (n = 238) | No (n = 263) | ||
| Grade of steatosis | 0.500 | ||
| 0 | 5 (2.1%) | 6 (2.3%) | |
| 1 | 85 (36.0%) | 81 (30.9%) | |
| 2 | 79 (33.5%) | 104 (39.7%) | |
| 3 | 67 (28.4%) | 71 (27.1%) | |
| Lobular inflammation | 0.017 | ||
| 0 | 8 (3.4%) | 28 (10.7%) | |
| 1 | 104 (44.1%) | 113 (43.1%) | |
| 2 | 111 (47.0%) | 109 (41.6%) | |
| 3 | 13 (5.5%) | 12 (4.6%) | |
| Ballooning | < 0.001 | ||
| 0 | 39 (16.6%) | 70 (26.7%) | |
| 1 | 98 (41.7%) | 129 (49.2%) | |
| 2 | 98 (41.7%) | 63 (24.0%) | |
| NAFLD activity score | 0.200 | ||
| NAS < 5 | 108 (46.0%) | 136 (51.7%) | |
| NAS ≥ 5 | 127 (54.0%) | 127 (48.3%) | |
| Median NAS (interquartile range) | 5 (3) | 4 (3) | 0.022 |
| Stage of fibrosis | < 0.001 | ||
| 0 | 40 (16.8%) | 98 (37.3%) | |
| 1 | 55 (23.1%) | 79 (30.0%) | |
| 2 | 47 (19.7%) | 41 (15.6%) | |
| 3 | 47 (19.7%) | 30 (11.4%) | |
| 4 | 49 (20.6%) | 15 (5.7%) | |
DM: diabetes mellitus, NAS: NAFLD activity score, NASH: non-alcoholic steatohepatitis.
p values were derived using Pearson's chi-square tests and Mann–Whitney U test.
Table 3.
Insulin resistance in relation to advanced fibrosis in patients without DM.
| HOMA-IR | Advanced fibrosis |
||
|---|---|---|---|
| Yes | No | ||
| Tertile 1 | 7 (36.8%) | 28 (43.8%) | p = 0.732 |
| Tertile 2 | 8 (42.1%) | 27 (42.2%) | |
| Tertile 3 | 4 (21.1%) | 9 (14.1%) | |
HOMA-IR: Homeostatic model assessment—insulin resistance.
Table 4.
Correlation of NFS and histology in patients with DM and without DM.
| NFS | DM |
Non-DM |
||||||
|---|---|---|---|---|---|---|---|---|
| Low risk | Indeterminate | High risk | Total | Low risk | Indeterminate | High risk | Total | |
| Stages 0–2 | 25 | 88 | 19 | 132 | 141 | 54 | 7 | 202 |
| Stages 3–4 | 9 | 36 | 44 | 89 | 17 | 17 | 6 | 40 |
| Total | 34 | 124 | 63 | 158 | 71 | 13 | ||
| Sensitivity | 89.9% | 49.4% | 57.5% | 15% | ||||
| Specificity | 18.9% | 85.6% | 69.8% | 96.5% | ||||
| PPV | 42.8% | 69.8% | 27.4% | 46.2% | ||||
| NPV | 73.5% | 71.5% | 89.2% | 85.1% | ||||
NFS: NAFLD fibrosis score, DM: diabetes mellitus, PPV: positive predictive value, NPV: negative predictive value.
Table 5.
Correlation of NFS components with advanced fibrosis in patients with and without DM.
| Advanced fibrosis | Age | BMI | Platelet | Albumin | AST/ALT | |
|---|---|---|---|---|---|---|
| DM | Rho | 0.21 | − 0.72 | − 0.33 | − 0.18 | 0.36 |
| p value | 0.001 | 0.269 | < 0.001 | 0.007 | < 0.001 | |
| Non-DM | Rho | 0.20 | 0.05 | − 0.19 | − 0.14 | 0.31 |
| p value | 0.001 | 0.459 | 0.003 | 0.027 | < 0.001 |
NFS: NAFLD fibrosis score, DM: diabetes mellitus, rho: correlation coefficient, AST: aspartate aminotransferase, ALT: alanine aminotransferase.
Table 6.
Correlation of BARD score and AST/ALT ratio to histology in patients with and without DM.
| DM | Non-DM | ||||||
|---|---|---|---|---|---|---|---|
| BARD | < 2 | ≥ 2 | Total | BARD | < 2 | ≥ 2 | Total |
| Stages 0–2 | 5 | 132 | 137 | Stages 0–2 | 145 | 65 | 210 |
| Stages 3–4 | 0 | 92 | 92 | Stages 3–4 | 16 | 26 | 42 |
| Total | 5 | 224 | 229 | Total | 161 | 91 | 252 |
| Sensitivity: 100%, specificity: 3.6% PPV: 41.1%, NPV: 100% |
Sensitivity: 61.9%, specificity: 69.0% PPV: 28.6%, NPV: 90.0% |
||||||
| AST:ALT | < 0.8 | ≥ 0.8 | Total | AST:ALT | < 0.8 | ≥ 0.8 | Total |
| Stages 0–2 | 78 | 59 | 137 | Stages 0–2 | 146 | 65 | 211 |
| Stages 3–4 | 21 | 71 | 92 | Stages 3–4 | 16 | 26 | 42 |
| Total | 99 | 130 | 229 | Total | 162 | 91 | 253 |
| Sensitivity: 77.2%, specificity: 56.9% PPV: 54.6%, NPV: 78.8% |
Sensitivity: 61.9%, specificity: 69.2% PPV: 28.6%, NPV: 90.1% |
||||||
4. Discussion
The present study shows that the clinical spectrum of NAFLD is different among patients with and without DM. More aggressive NAFLD as suggested by the higher prevalence of NASH and advanced fibrosis was seen in DM patients compared to non-DM patients. Other factors contributing to more severe NAFLD in the DM patients include older age, higher BMI and increased prevalence of hypertension. This is compatible with previous reports observing that patients with DM have higher risk of NAFLD, higher rates of NASH and advanced fibrosis relative to patients without DM [10], [22], [23]. In addition, DM is associated with accelerated progression of hepatic fibrosis [24], which is consistent with the higher prevalence of cirrhosis among the DM patients relative to the non-DM patients in our study. NAFLD patients with DM have twice the mortality risk compared to the general population with DM and no NAFLD [25]. Similarly, NAFLD patients with DM have three times the overall mortality risk and twenty two times the liver related mortality risk compared to NAFLD patients without DM [10]. In part, this mortality risk relationship is contributed by DM patients having a higher prevalence of NASH and advanced fibrosis. Studies have observed a higher overall mortality in patients with NASH relative to the general population, in particular attributed to cardiovascular and liver related mortality [26], [27]. Comparing patients with and without NASH, NASH was associated with higher rates of cirrhosis and liver-related mortality [28]. This reiterates the importance of DM in the context of NAFLD.
Liver biopsy remains the gold standard in diagnosing and predicting the severity of NAFLD, with several histological features that can predict clinically relevant outcomes in NAFLD [29]. With regard to liver histology, our study found that inflammation and ballooning, integral features required for the diagnosis of NASH, were more frequently seen among NAFLD patients with DM, relative to patients without DM. Hepatocyte injury in NAFLD is often noted in the form of hepatocellular ballooning, and is often considered the most important histological feature for distinguishing NASH [30], [31]. Ballooning is associated with more aggressive NAFLD and higher rates of cirrhosis [28]. Similarly, necroinflammation at index biopsy was reported to be the only predictor of fibrosis progression in an analysis of ten histo-pathological series with follow-up liver biopsies [32].
The lower serum albumin and platelet count in combination with a higher AST to ALT ratio seen with the DM patients reflect the higher prevalence of advanced fibrosis as they are recognised markers of hepatic synthetic function, portal hypertension and surrogates of advanced fibrosis respectively [21], [33].
Although DM has been proven to be an established risk factor for NASH and advanced fibrosis, the potential for severe disease in patients without DM should not be overlooked. Several epidemiological studies have reported prevalence rates of DM in NAFLD cohorts between 7 and 41% [1], [8], [14], [34]. Consequently, although prevalence rates of NASH and advanced fibrosis may be lower among non-diabetics, taking into account the larger number of patients in the non-diabetic group, the absolute number of patients with significantly severe NAFLD may not be significantly less than in diabetics. Indeed, in our study, a significant proportion of NAFLD patients without DM has NASH (64%) and advanced fibrosis (17%). This reiterates the need for cautious management of NAFLD in the non-DM patient without trivialising the potential for disease severity in this subgroup of patients. Our study did not find a significant association between severity of insulin resistance and advanced fibrosis among the patients without DM, which may suggest that factors other than insulin resistance play a more prominent role in fibrogenesis among non-diabetics. This is consistent with the evolving recognition that multiple factors (not just insulin resistance) are involved in the pathogenesis of NAFLD [35], [36].
With regard to non-invasive tools used in NAFLD, the NFS has been widely used as a reliable tool to predict advanced fibrosis in NAFLD [14], [37], [38], [39]. In our study, the NFS performed well in identifying patients with and without advanced fibrosis in the entire group. However subgroup analysis suggested that the utility of the NFS had divergent clinical reliability for NAFLD patients with and without DM, particularly in excluding advanced fibrosis in patients without DM although this must be tampered in the context of a relatively lower prevalence of advanced fibrosis among the non-DM patients. Larger studies are warranted to clarify the utility of the NFS in non-DM patients with NAFLD. Nevertheless, this may illustrate the heterogeneous nature of the NAFLD phenotype and that the NFS may have to be adjusted in different subsets of NAFLD patients. Similar discrepant findings using the BARD score and AST/ALT ratios for predicting advanced fibrosis between patients with and without DM support the individualised approach to this heterogeneous disease. A recent study observed differences in sensitivity and specificity of NFS between NAFLD patients with and without normal ALT levels; sensitivity of NFS (at lower cut-off) was good in patients with normal ALT at 82% but specificity was low at 51% [15]. Yoneda et al. had similar findings and suggested resetting the cut-off values for NFS to achieve higher sensitivity and specificities in NAFLD patients with normal ALT levels [40].
4.1. Strength and limitations
To our knowledge, this is the first study to explore the clinical utility of the commonly used NFS in NAFLD patients with and without DM. The strength of this study include the large cohort of patients with histologically proven NAFLD in addition to the wide range of available demographic, anthropometric, clinical, histological and laboratory data. However, there are a number of potential limitations. The cross-sectional nature of our study allows the evaluation of associations only. Furthermore, the use of electronic medical records to define the presence or absence of DM and hypertension may have led to some patients being misclassified. With regard to the NFS, the relatively small number of patients with advanced fibrosis in the non-diabetic group may impact on the test performance of the NFS, in particular the positive predictive value. Nevertheless, this study highlights important clinical concepts which are relevant to the everyday practical management of NAFLD.
5. Conclusion
While NAFLD patients with DM have higher prevalence of NASH and advanced fibrosis, the potential risk of severe NAFLD cannot be overlooked in patients without DM since both NASH and advanced fibrosis occur in a considerable proportion of these patients. The utility of the NFS and other fibrosis scores may have disparate clinical reliability for NAFLD patients with and without DM, which emphasises the heterogeneous nature of the NAFLD phenotype and that the NFS may have to be adjusted in different subsets of NAFLD patients. This warrants further studies to explore this relationship in greater detail.
Conflict of interest
All authors declare no conflicts of interest.
Acknowledgements
GBBG was supported by Singapore SingHealth HMDP. MRP was supported by NIH T32 DK061917. SD, JD, AJM, RS, and CH were supported by NIH DK U.
Footnotes
No grant support was provided for this study.
References
- 1.Marchesini G., Bugianesi E., Forlani G. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology. 2003;37(4):917–923. doi: 10.1053/jhep.2003.50161. [DOI] [PubMed] [Google Scholar]
- 2.Chalasani N., Younossi Z., Lavine J.E. The diagnosis and management of non-alcoholic fatty liver disease: practice Guideline by the American Association for the Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association. Hepatology. 2012;55(6):2005–2023. doi: 10.1002/hep.25762. [DOI] [PubMed] [Google Scholar]
- 3.Musso G., Gambino R., Cassader M., Pagano G. Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity. Ann. Med. 2011;43(8):617–649. doi: 10.3109/07853890.2010.518623. [DOI] [PubMed] [Google Scholar]
- 4.Younossi Z.M., Reyes M.J., Mishra A., Mehta R., Henry L. Systematic review with meta-analysis: non-alcoholic steatohepatitis — a case for personalised treatment based on pathogenic targets. Aliment. Pharmacol. Ther. 2014;39(1):3–14. doi: 10.1111/apt.12543. [DOI] [PubMed] [Google Scholar]
- 5.Bugianesi E., McCullough A.J., Marchesini G. Insulin resistance: a metabolic pathway to chronic liver disease. Hepatology. 2005;42(5):987–1000. doi: 10.1002/hep.20920. [DOI] [PubMed] [Google Scholar]
- 6.El-Serag H.B., Tran T., Everhart J.E. Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology. 2004;126(2):460–468. doi: 10.1053/j.gastro.2003.10.065. [DOI] [PubMed] [Google Scholar]
- 7.Hossain N., Afendy A., Stepanova M. Independent predictors of fibrosis in patients with nonalcoholic fatty liver disease. Clin. Gastroenterol. Hepatol. 2009;7(11):1224–1229. doi: 10.1016/j.cgh.2009.06.007. (9 e1-2) [DOI] [PubMed] [Google Scholar]
- 8.Neuschwander-Tetri B.A., Clark J.M., Bass N.M. Clinical, laboratory and histological associations in adults with nonalcoholic fatty liver disease. Hepatology. 2010;52(3):913–924. doi: 10.1002/hep.23784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Loomba R., Abraham M., Unalp A. Association between diabetes, family history of diabetes, and risk of nonalcoholic steatohepatitis and fibrosis. Hepatology. 2012;56(3):943–951. doi: 10.1002/hep.25772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Younossi Z.M., Gramlich T., Matteoni C.A., Boparai N., McCullough A.J. Nonalcoholic fatty liver disease in patients with type 2 diabetes. Clin. Gastroenterol. Hepatol. 2004;2(3):262–265. doi: 10.1016/s1542-3565(04)00014-x. [DOI] [PubMed] [Google Scholar]
- 11.Dixon J.B., Bhathal P.S., O'Brien P.E. Nonalcoholic fatty liver disease: predictors of nonalcoholic steatohepatitis and liver fibrosis in the severely obese. Gastroenterology. 2001;121(1):91–100. doi: 10.1053/gast.2001.25540. [DOI] [PubMed] [Google Scholar]
- 12.Gholam P.M., Flancbaum L., Machan J.T., Charney D.A., Kotler D.P. Nonalcoholic fatty liver disease in severely obese subjects. Am. J. Gastroenterol. 2007;102(2):399–408. doi: 10.1111/j.1572-0241.2006.01041.x. [DOI] [PubMed] [Google Scholar]
- 13.Harrison S.A., Oliver D., Arnold H.L., Gogia S., Neuschwander-Tetri B.A. Development and validation of a simple NAFLD clinical scoring system for identifying patients without advanced disease. Gut. 2008;57(10):1441–1447. doi: 10.1136/gut.2007.146019. [DOI] [PubMed] [Google Scholar]
- 14.Angulo P., Hui J.M., Marchesini G. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology. 2007;45(4):846–854. doi: 10.1002/hep.21496. [DOI] [PubMed] [Google Scholar]
- 15.McPherson S., Anstee Q.M., Henderson E., Day C.P., Burt A.D. Are simple noninvasive scoring systems for fibrosis reliable in patients with NAFLD and normal ALT levels? Eur. J. Gastroenterol. Hepatol. 2013;25(6):652–658. doi: 10.1097/MEG.0b013e32835d72cf. [DOI] [PubMed] [Google Scholar]
- 16.American Diabetes A. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2013;36(Suppl. 1):S67–S74. doi: 10.2337/dc13-S067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chobanian A.V., Bakris G.L., Black H.R. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension. 2003;42(6):1206–1252. doi: 10.1161/01.HYP.0000107251.49515.c2. [DOI] [PubMed] [Google Scholar]
- 18.Kleiner D.E., Brunt E.M., Van Natta M. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. 2005;41(6):1313–1321. doi: 10.1002/hep.20701. [DOI] [PubMed] [Google Scholar]
- 19.Brunt E.M., Kleiner D.E., Wilson L.A. Nonalcoholic fatty liver disease (NAFLD) activity score and the histopathologic diagnosis in NAFLD: distinct clinicopathologic meanings. Hepatology. 2011;53(3):810–820. doi: 10.1002/hep.24127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Matthews D.R., Hosker J.P., Rudenski A.S. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–419. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
- 21.McPherson S., Stewart S.F., Henderson E., Burt A.D., Day C.P. Simple non-invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non-alcoholic fatty liver disease. Gut. 2010;59(9):1265–1269. doi: 10.1136/gut.2010.216077. [DOI] [PubMed] [Google Scholar]
- 22.Nakahara T., Hyogo H., Yoneda M. Type 2 diabetes mellitus is associated with the fibrosis severity in patients with nonalcoholic fatty liver disease in a large retrospective cohort of Japanese patients. J. Gastroenterol. 2013 doi: 10.1007/s00535-013-0911-1. [DOI] [PubMed] [Google Scholar]
- 23.Argo C.K., Caldwell S.H. Epidemiology and natural history of non-alcoholic steatohepatitis. Clin. Liver Dis. 2009;13(4):511–531. doi: 10.1016/j.cld.2009.07.005. [DOI] [PubMed] [Google Scholar]
- 24.Adams L.A., Sanderson S., Lindor K.D., Angulo P. The histological course of nonalcoholic fatty liver disease: a longitudinal study of 103 patients with sequential liver biopsies. J. Hepatol. 2005;42(1):132–138. doi: 10.1016/j.jhep.2004.09.012. [DOI] [PubMed] [Google Scholar]
- 25.Adams L.A., Harmsen S., St Sauver J.L. Nonalcoholic fatty liver disease increases risk of death among patients with diabetes: a community-based cohort study. Am. J. Gastroenterol. 2010;105(7):1567–1573. doi: 10.1038/ajg.2010.18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ekstedt M., Franzen L.E., Mathiesen U.L. Long-term follow-up of patients with NAFLD and elevated liver enzymes. Hepatology. 2006;44(4):865–873. doi: 10.1002/hep.21327. [DOI] [PubMed] [Google Scholar]
- 27.Soderberg C., Stal P., Askling J. Decreased survival of subjects with elevated liver function tests during a 28-year follow-up. Hepatology. 2010;51(2):595–602. doi: 10.1002/hep.23314. [DOI] [PubMed] [Google Scholar]
- 28.Matteoni C.A., Younossi Z.M., Gramlich T. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology. 1999;116(6):1413–1419. doi: 10.1016/s0016-5085(99)70506-8. [DOI] [PubMed] [Google Scholar]
- 29.Pagadala M.R., McCullough A.J. The relevance of liver histology to predicting clinically meaningful outcomes in nonalcoholic steatohepatitis. Clin. Liver Dis. 2012;16(3):487–504. doi: 10.1016/j.cld.2012.05.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yeh M.M., Brunt E.M. Pathology of nonalcoholic fatty liver disease. Am. J. Clin. Pathol. 2007;128(5):837–847. doi: 10.1309/RTPM1PY6YGBL2G2R. [DOI] [PubMed] [Google Scholar]
- 31.Brunt E.M. Pathology of nonalcoholic fatty liver disease. Nat. Rev. Gastroenterol. Hepatol. 2010;7(4):195–203. doi: 10.1038/nrgastro.2010.21. [DOI] [PubMed] [Google Scholar]
- 32.Argo C.K., Northup P.G., Al-Osaimi A.M., Caldwell S.H. Systematic review of risk factors for fibrosis progression in non-alcoholic steatohepatitis. J. Hepatol. 2009;51(2):371–379. doi: 10.1016/j.jhep.2009.03.019. [DOI] [PubMed] [Google Scholar]
- 33.Giannini E., Botta F., Fasoli A. Progressive liver functional impairment is associated with an increase in AST/ALT ratio. Dig. Dis. Sci. 1999;44(6):1249–1253. doi: 10.1023/a:1026609231094. [DOI] [PubMed] [Google Scholar]
- 34.Fan J.G., Zhu J., Li X.J. Fatty liver and the metabolic syndrome among Shanghai adults. J. Gastroenterol. Hepatol. 2005;20(12):1825–1832. doi: 10.1111/j.1440-1746.2005.04058.x. [DOI] [PubMed] [Google Scholar]
- 35.Tilg H., Moschen A.R. Evolution of inflammation in nonalcoholic fatty liver disease: the multiple parallel hits hypothesis. Hepatology. 2010;52(5):1836–1846. doi: 10.1002/hep.24001. [DOI] [PubMed] [Google Scholar]
- 36.Peverill W., Powell L.W., Skoien R. Evolving concepts in the pathogenesis of NASH: beyond steatosis and inflammation. Int. J. Mol. Sci. 2014;15(5):8591–8638. doi: 10.3390/ijms15058591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Qureshi K., Clements R.H., Abrams G.A. The utility of the “NAFLD fibrosis score” in morbidly obese subjects with NAFLD. Obes. Surg. 2008;18(3):264–270. doi: 10.1007/s11695-007-9295-8. [DOI] [PubMed] [Google Scholar]
- 38.Wong V.W., Wong G.L., Chim A.M. Validation of the NAFLD fibrosis score in a Chinese population with low prevalence of advanced fibrosis. Am. J. Gastroenterol. 2008;103(7):1682–1688. doi: 10.1111/j.1572-0241.2008.01933.x. [DOI] [PubMed] [Google Scholar]
- 39.Perez-Gutierrez O.Z., Hernandez-Rocha C., Candia-Balboa R.A. Validation study of systems for noninvasive diagnosis of fibrosis in nonalcoholic fatty liver disease in Latin population. Ann. Hepatol. 2013;12(3):416–424. [PubMed] [Google Scholar]
- 40.Yoneda M., Imajo K., Eguchi Y. Noninvasive scoring systems in patients with nonalcoholic fatty liver disease with normal alanine aminotransferase levels. J. Gastroenterol. 2013;48(9):1051–1060. doi: 10.1007/s00535-012-0704-y. [DOI] [PubMed] [Google Scholar]