Abstract
Background
The FibroScan-AST (FAST) score was recently described to detect patients with nonalcoholic steatohepatitis (NASH) having elevated nonalcoholic fatty liver disease (NAFLD) activity score (NAS ≥ 4) and significant fibrosis (≥ F2) on liver biopsy (NASH+ NAS ≥ 4 + F ≥ 2).
Aim
The aim of this study was to validate the FAST score in Indian patients with NAFLD and to derive optimal cut-offs.
Methods
Sixty patients with biopsy-proven NAFLD [men: 38 (63.3%), age 40 (32–52) years] with all parameters for assessing the FAST score within 3 months of liver histology were retrospectively analysed.
Results
Histological NASH was present in 17 patients (28.3%), while 11 (18.3%) patients had NASH + NAS ≥ 4 + F ≥ 2. The area under the curve (AUROC) of the FAST score for discriminating NASH + NAS ≥ 4 + F ≥ 2 was 0.81. Using cut-offs by Newsome et al, the rule-out cut-off (FAST: ≤ 0.35) had a negative predictive value (NPV) of 0.88 [sensitivity: 0.91, specificity: 0.14, negative likelihood ratio (LR): 0.64], while the rule-in cut-off (FAST: ≥ 0.67) had a positive predictive value (PPV) of 0.33 (sensitivity: 0.73, specificity: 0.67, positive LR: 2.22). Fifteen (25%) patients were correctly classified as per histology, while 28 (46.67%) patients fell in the grey zone. On recalculating the optimal cut-offs for our patients, the rule-out cut-off (FAST: ≤ 0.55) had an NPV of 0.95 (sensitivity: 0.90, specificity: 0.45, negative LR: 0.21), while the optimal rule-in cut-off (FAST: ≥ 0.78) had a PPV of 0.70 (sensitivity: 0.64, specificity 0.94, positive LR: 10.39). With these cut-offs, 27 (45%) patients fell in the grey zone and 29 (48.3%) were correctly classified as per histology, performing better than the cut-offs by Newsome et al (P < 0.001).
Conclusion
The FAST score demonstrates good AUROC for detecting NASH with significant fibrosis and inflammation on histology. Cut-offs should be recalibrated based on prevalence of disease.
Lay summary
India has a high burden of NAFLD with an estimated 25 million patients at potential risk for significant liver disease. Liver biopsy remains the gold standard for diagnosing NASH, although its application in routine clinical practice is limited. Noninvasive tests for the simultaneous detection of steatosis, inflammation and fibrosis are thus the need of the hour. The FAST score has been recently suggested for the noninvasive detection of NASH with significant fibrosis (≥ F2) and inflammation (NAS ≥ 4) on liver biopsy. We validated the utility of the FAST score for detecting NASH with significant fibrosis and inflammation on liver biopsy in Indian patients with NAFLD. This noninvasive, easy-to-use and nonproprietary FAST score can correctly classify disease severity in more than 50% patients. However, our results suggest that cut-offs should be recalibrated based on the anticipated prevalence of NASH + NAS ≥ 4 + F ≥ 2 in the given population.
Keywords: nonalcoholic fatty liver disease, NAFLD, steatosis, FibroScan
Abbreviations: ALT, Alanine Aminotransferase; AMA, Antimitochondrial Antibody; ANAs, Antinuclear Antibodies; AST, Aspartate aminotransferase; BMI, Body Mass Index; CAP, Controlled Attenuation Parameter; FAST, FibroScan-AST; FLIP, Fatty Liver Inhibition of Progression; HTN, Hypertension; LKM, Antiliver Kidney Microsomal Antibody; LSM, Liver Stiffness Measurement; NAFLD, Nonalcoholic Fatty Liver Disease; NASH, Nonalcoholic Steatohepatitis; NAS, NAFLD Activity Score; NASH-CRN, NASH Clinical Research Network; SAF, Steatosis, Activity and Fibrosis Score; T2DM, Type 2 Diabetes Mellitus
Nonalcoholic fatty liver disease (NAFLD) is an umbrella term characterised by macrovesicular hepatic steatosis in patients with insignificant or no alcohol intake and absence of secondary causes such as viruses, drug exposure or known genetic causes of fatty liver.1 A subset of these patients develops progressive liver disease characterised by hepatocyte ballooning, lobular inflammation and fibrosis which is termed as nonalcoholic steatohepatitis (NASH). NASH is usually a histological diagnosis, and its distinction from simple steatosis or nonalcoholic fatty liver (NAFL) is very important not only for prognosis (higher disease progression in NASH in comparison with NAFL) but also for therapeutic purposes (pharmacotherapy usually directed towards those with NASH).2,3 Although a large number of noninvasive modalities are available for detecting and assessing the severity of hepatic steatosis and fibrosis, currently there are no validated noninvasive techniques for the simultaneous assessment of all three components of hepatic steatosis, inflammation and fibrosis in patients with NASH.4,5
Recently, Newsome et al6 derived the FibroScan-AST (FAST) score in a cohort of patients with NAFLD from England and validated it in multiple international cohorts. The FAST score demonstrated an impressive AUROC of 0.74–0.95 for detecting patients with NASH, elevated NAFLD activity score (NAS ≥ 4) and significant fibrosis (≥ F2) on liver biopsy. These are the patients who are at high risk of progression and would benefit the most from antifibrotic or antiinflammatory therapy. While the validation cohorts included patients from two Asian countries (China and Malaysia), no Indian cohort was included in their study.6 India has a high prevalence of NAFLD which is responsible for significant liver disease in this country.7,8 The aim of the present study was to validate the FAST score in Indian patients with biopsy-proven NAFLD.
Methods
In a retrospective analysis of the prospectively collected data, the FAST score was applied to 60 patients with biopsy-proven NAFLD managed at a tertiary care centre in north India over a period of five years. Patients were included in the study if they had all the parameters required to assess the FAST score including serum aspartate aminotransferase (AST), FibroScan with controlled attenuation parameter (CAP) and liver stiffness measurement (LSM) values within 3 months of liver histology. The study had the approval of the institute's ethics committee and waiver for informed consent from the patients. The study has been reported in accordance with the Standards for Reporting Diagnostic accuracy studies (STARD) guidelines.
Adult patients with no or low alcohol consumption <20 g/day (history confirmed from 2 reliable family members) were included in the study. All patients had sonographic evidence of hepatic steatosis and raised liver enzymes [aspartate aminotransferase (AST), alanine aminotransferase (ALT) ≥ 40 IU/L] with negative viral markers (hepatitis B surface antigen and anti-hepatitis C antibody), negative autoimmune markers (antinuclear antibody, antismooth muscle antibody, antiliver kidney microsomal antibody and antimitochondrial antibody), normal serum ceruloplasmin levels and normal iron profiles (serum iron, total iron-binding capacity, ferritin and transferrin saturation). Finally, the diagnosis of NAFLD was confirmed on liver biopsy (macrovesicular steatosis in >5% hepatocytes on liver histology). Patients with past history of hepatic decompensation were excluded from the study.
Clinical Assessment
Data on clinical assessment included a thorough history of the patients’ symptoms, history of alcohol intake, type 2 diabetes mellitus (T2DM), hypertension, dyslipidemia and overweight/obesity, along with a comprehensive general physical and systemic examination, including height, weight, body mass index (BMI) and waist circumference. The waist circumference was measured at the midpoint between the lower margin of the last palpable rib and the top of the iliac crest, using a stretch-resistant tape. All patients had details regarding routine blood tests including a complete haemogram, liver function test, lipid profile and viral markers. Further evaluation included serum autoimmune markers, iron profile studies, serum ceruloplasmin and celiac disease workup. Details of ultrasonography and transient elastography [Fibroscan by Echosens (Paris)] were analysed in all the patients. Finally, the details of liver biopsy (carried out within 3 months of the noninvasive assessment, which confirmed the diagnosis and severity of NAFLD), were recorded in all patients.
Metabolic syndrome was defined as having presence of at least three of the five parameters set by the National Cholesterol Education Program Adult Treatment Panel III criteria for diagnosing metabolic syndrome with modified waist circumference cut-offs for Indians (≥ 90 cm in men and ≥ 80 cm in women).9, 10, 11 Similarly, overweight and obesity were defined as per the Asia-Pacific cut-offs as BMI of 23.0–24.9 kg/m2 and more than or equal to 25 kg/m2, respectively.10,11 Hypertriglyceridemia was defined as serum triglyceride level of more than 150 mg/dl, dysglycemia as fasting glucose level of more than or equal to 100 mg/dl or patients on antidiabetic drugs, low high-density lipoprotein (HDL) as HDL level less than 40 mg/dl (male) and 50 mg/dl (female) and hypertension as blood pressure more than or equal to 130/85 mm of Hg or patients already on antihypertensive drugs.9
Noninvasive Assessment of Hepatic Steatosis, Necroinflammation and Fibrosis
Increased hepatic parenchymal echogenicity, blurring of intrahepatic vascular structures and deep signal attenuation were taken as ultrasonographic evidence of presence and severity of hepatic steatosis. Hepatic steatosis was also assessed on FibroScan and based on CAP values; steatosis was graded as absent (S0: < 248 dB/m), mild (S1: 248 to < 268 dB/m), moderate (S2: 268 to < 280 dB/m) or severe (S3: > 280 dB/m).12
Fibrosis was assessed noninvasively using both simple composite scores [Fibrosis-4 (FIB-4) and NAFLD fibrosis score (NFS)] and transient elastography (FibroScan). FIB-4 cut-offs of <1.3 and >3.25 were used for ruling-out and ruling-in advanced fibrosis ( ≥ F3),13 respectively. For NFS, the corresponding cut-off values used were −1.455 and 0.676, respectively.14 Based on LSM values, fibrosis was staged as F0–F1 (<7 kPa), ≥ F2 (7–8.7 kPa), ≥ F3 (8.7–10.3 kPa) and F4 (≥10.3 kPa).15 For FibroScan, an XL probe was used as per the automatic probe selection tool if the probe to liver capsule distance displayed in real time was >25 mm in accordance with the manufacturer's recommendations (https://www.echosens.com/fibroscanprocedure/). The XL probe was also used if there was failure to obtain 10 valid readings with automatically suggested M probe.
Liver Histology
All patients had undergone percutaneous liver biopsy with or without ultrasound guidance, and grades of hepatic steatosis, inflammation and fibrosis were assessed by an NASH clinical research network system and steatosis, activity and fibrosis (SAF) score using the fatty liver inhibition of progression (FLIP) algorithm by an experienced pathologist with expertise in reporting liver histology.16,17 NASH was defined using the FLIP definition (≥1 point for steatosis, ballooning and lobular inflammation, each).18
FAST Score
The FAST score was calculated using [e ˆ(– 1·65 + 1·07 × In(LSM) + 2·66∗10−8 × CAP3 – 63.3 × AST−1)]/[1 + e ˆ (– 1·65 + 1·07 × In(LSM) + 2·66∗10−8 × CAP3 – 63.3 × AST−1)] as carried out by Newsome et al.6 The primary objective of the study was to assess the discriminatory ability of the FAST score to detect patients with NASH, NAS ≥4 and fibrosis stage ≥2 (i.e., NASH + NAS ≥ 4 + F ≥ 2). The secondary objective was to define cut-offs for ruling-in and ruling-out NASH + NAS ≥4 + F ≥ 2 in our cohort of patients with NAFLD.
Sample Size Calculation
A biopsy study in 1000 Indian patients with NAFLD showed the presence of histological NASH in 62% patients.19 A sample size of 60 was calculated to estimate a similar proportion of NASH with a precision of estimate of 0.10 and a confidence level of 0.90.
Statistical Analysis
Statistical analysis was performed using Microsoft Excel 365 (Microsoft, USA) and MedCalc Statistical Software, version 19.3 (MedCalc Software Ltd, Ostend, Belgium). Data are expressed as mean ± standard deviation or median (interquartile range), as applicable. Discriminatory ability of the FAST score for detecting NASH + NAS ≥ 4 + F ≥ 2 was assessed using the area under the curve (AUROC). Confidence intervals (CI) for AUROC were derived using binomial extract. Rule-out and rule-in cut-offs were derived from the receiver operating characteristics (ROC) curve using sensitivity ≥90% and specificity ≥90%, respectively, with optimal likelihood ratios (LRs). Comparison of AUROCs was carried out using paired methods by the DeLong test. Comparison of cut-offs for classifying patients as per histology was carried out using the Chi-square test for trend. For hypothesis testing, P value < 0.05 was considered statistically significant.
Results
Sixty patients (men: 38 [63.3%], median age 40 [32–52] years) with biopsy-proven NAFLD managed in the last five years who met the inclusion criteria were included in the study. The flow diagram of the study is shown in Supplementary Figure 1. The demographic and clinical profile of these patients is shown in Table 1. Using Indian BMI cut-offs, eight (13.3%) patients were overweight (BMI 23–24.9 kg/m2) and 42 (70%) patients were obese (BMI ≥25 kg/m2). Twenty-six (43.3%) patients had an abnormal waist circumference suggestive of central obesity. T2DM, hypertension and metabolic syndrome were present in 11 (18.3%), 10 (16.6%) and 22 (36.7%) patients, respectively.
Table 1.
Baseline Demographic and Clinical Profile of Patients With NAFLD.
| Characteristic | n = 60 |
|---|---|
| Age (years) | 40 (32–52)a |
| Male | 38 (63.3%) |
| BMI (kg/m2) | 28.2 (24.2–31.6)a |
| Waist circumference (cm) | 94 (88–102)a |
| Type 2 diabetes mellitus | 11 (18.3%) |
| Hypertension | 10 (16.6%) |
| AST (IU/L) | 89.5 (60.6–115.5)a |
| ALT (IU/L) | 74.25 (51.75–91.75)a |
| Albumin (g/L) | 4.2 (3.8–4.6)a |
| Platelet count (X109/L) | 218.5 (159.3–280.8)a |
| Fasting glucose (mg/dL) | 97 (87.3–106)a |
| Serum triglyceride (mg/dL) | 166.5 (126–215.2)a |
| Total cholesterol (mg/dL) | 182 (155.8–212.3)a |
| HDL (mg/dL) | 41.6 (33–51.2)a |
AST, aspartate aminotransferase; ALT, alanine aminotransferase; HDL, high-density lipoprotein; BMI, body mass index.
Data expressed as median (interquartile range).
Noninvasive Assessment of Hepatic Steatosis, Inflammation and Fibrosis
All patients had evidence of hepatic steatosis on ultrasound. The details of noninvasive assessment of steatosis and fibrosis are shown in Table 2. More than half of the patients (58.3%) in our cohort had severe steatosis (S3) using CAP values obtained by FibroScan.
Table 2.
Noninvasive Assessment of Hepatic Fibrosis and Steatosis in Patients With NAFLD (n = 60).
| Composite fibrosis scores | FIB-4 | 1.8 (1.16–2.98)a |
|
18 (30%) | |
|
29 (48.3%) | |
|
13 (21.7%) | |
| NFS | 0.59 (−0.52–1.57)a | |
|
4 (6.7%) | |
|
28 (46.7%) | |
|
28 (46.6%) | |
| FibroScan | LSM | 6.7 (5–9.3)a |
|
31 (51.6%) | |
|
10 (16.7%) | |
|
6 (10%) | |
|
13 (21.7%) | |
| CAP | 289 (276.5–324.3)a | |
|
4 (6.7%) | |
|
9 (15%) | |
|
12 (20%) | |
|
35 (58.3%) |
NFS, NAFLD fibrosis score; LSM, liver stiffness measurement; CAP, controlled attenuation parameter; NAFLD, nonalcoholic fatty liver disease.
Data expressed as median (interquartile range).
Although all patients had elevated hepatic transaminases (AST, ALT ≥ 40 IU/L), significant elevation (>2 times the upper limit of normal) of AST and ALT was seen in 35 (58.3%) and 26 (43.3%) patients, respectively.
Advanced fibrosis (≥ F3) was predicted to be present in 13 (21.7%) and 28 (46.6%) patients by using FIB-4 and NFS, respectively. FibroScan was carried out using M probe in 15 (25%) patients, while XL probe was used in the rest. There were no failures while using FibroScan. Using LSM values, 19 (31.7%) patients had advanced fibrosis (≥ F3), whereas significant fibrosis (≥ F2) was present in 29 (48.3%) patients. The AUROC of LSM (0.88, 95% CI: 0.78–0.95) for detecting advanced fibrosis on histology was significantly better than that of FIB-4 (0.71, 95% CI: 0.57–0.82; P = 0.04). There was no difference between the AUROCs of LSM and NFS (0.74, 95% CI: 0.61–0.84; P = 0.11) or FIB-4 and NFS (P = 0.64) for detecting advanced fibrosis on biopsy (Supplementary Figure 2).
Liver Histology
All patients had evidence of macrovesicular steatosis in >5% hepatocytes on liver histology. Significant fibrosis (≥ F2) and advanced fibrosis (≥ F3) were present in 21 (35%) and 14 (23.3%) patients, respectively. Histological NASH was present in 17 (28.3%) patients as per the FLIP definition, while 11 (18.3%) patients had NASH + NAS ≥ 4 + F ≥ 2 on biopsy (Table 3).
Table 3.
Histopathological Profile of Patients With NAFLD (n = 60).
| Fibrosis stage | 0 | 16 (26.7%) |
| 1 | 23 (38.3%) | |
| 2 | 7 (11.7%) | |
| 3 | 9 (15%) | |
| 4 | 5 (8.3%) | |
| Ballooning grade | 0 | 42 (70%) |
| 1 | 13 (21.7%) | |
| 2 | 5 (8.3%) | |
| Lobular inflammation grade | 0 | 10 (16.7%) |
| 1 | 33 (55%) | |
| 2 | 15 (25%) | |
| 3 | 2 (3.3%) | |
| Steatosis grade | 0 | 0 |
| 1 | 19 (31.7%) | |
| 2 | 27 (45%) | |
| 3 | 14 (23.3%) | |
| NAS score ≥4 | 38 (63.3%) | |
| NASH | 17 (28.3%) | |
| Significant fibrosis (F ≥ 2) | 21 (35%) | |
| Advanced fibrosis (F ≥ 3) | 14 (23.3%) | |
| NASH + NAS ≥4 + F ≥ 2 | 11 (18.3%) | |
NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NAS, NAFLD activity score.
FAST Score
The FAST score demonstrated good discriminatory performance for NASH + NAS ≥ 4 + F ≥ 2 in our cohort with an AUROC of 0.81 (95% CI: 0.68–0.90; P = 0.001) (Figure 1). The AUROC of the FAST score was significantly better than that of FIB-4 [0.52 (95% CI: 0.39–0.65); P = 0.04] and NFS [0.55 (95% CI: 0.41–0.68); P = 0.03] for detecting NASH + NAS ≥ 4 + F ≥ 2 on biopsy (Figure 2).
Figure 1.
Diagnostic performance of the FAST score for diagnosing NASH + NAS ≥ 4 + F ≥ 2; (A) AUROC curve; (B) sensitivity and specificity at various values of FAST score. NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NAS, NAFLD activity score; FAST, FibroScan-AST.
Figure 2.
AUROC of FIB-4, NFS and FAST scores for diagnosing NASH + NAS ≥4 + F ≥ 2. NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NAS, NAFLD activity score; FAST, FibroScan-AST.
As per the cut-offs suggested by Newsome et al6, the rule-out cut-off (FAST: ≤ 0.35) had a negative predictive value (NPV) of 0.88 (sensitivity: 0.91, specificity: 0.14, negative LR: 0.64), whereas the rule-in cut-off (FAST: ≥ 0.67) had a positive predictive value (PPV) of 0.33 (sensitivity: 0.73, specificity: 0.67, positive LR: 2.22). Twenty-eight (46.67%) patients in our cohort fell in the grey zone using these cut-offs (FAST: 0.35–0.67) (Table 4). Of the 32 patients who were ruled in or ruled out by these cut-offs, 15 (46.9%) patients were correctly classified as per histology. Thus, only 15 (25%) patients of the whole cohort could be correctly classified using the cut-offs derived by Newsome et al.6
Table 4.
Diagnostic Performance of the FAST Score for Diagnosing NASH + NAS ≥ 4 + F ≥ 2 in Indian Patients With NAFLD (n = 60).
| Using cut-offs proposed by Newsome et al6 | ||||||||
|---|---|---|---|---|---|---|---|---|
| Rule-out zone (FAST: ≤0.35) |
Grey zone (FAST: 0.35–0.67) |
Rule-in zone (FAST: ≥0.67) |
||||||
| n (%) | Sensitivity | Specificity | NPV | n (%) | n (%) | Sensitivity | Specificity | PPV |
| 8 (13.3% | 0.91 | 0.14 | 0.88 | 28 (46.7%) | 24 (40%) | 0.73 | 0.67 | 0.33 |
| Using optimal cut-offs derived in this study | ||||||||
|---|---|---|---|---|---|---|---|---|
| Rule-out zone (FAST: ≤0.55) |
Grey zone (FAST: 0.55–0.78) |
Rule-in zone (FAST: ≥0.78) |
||||||
| n (%) | Sensitivity | Specificity | NPV | n (%) | n (%) | Sensitivity | Specificity | PPV |
| 23 (38.3%) | 0.90 | 0.45 | 0.95 | 27 (45%) | 10 (16.7%) | 0.64 | 0.94 | 0.70 |
NPV, negative predictive value; PPV, positive predictive value; FAST, FibroScan-AST; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; NAS, NAFLD activity score.
Because of the poor PPV of the Newsome's rule-in cut-off FAST score in our patients, we recalculated the optimal rule-out and rule-in cut-offs for the FAST score in our cohort of patients. We found an optimal FAST score rule-out cut-off of ≤0.55 with an NPV of 0.95 (sensitivity: 0.90, specificity: 0.45, negative LR: 0.21) and an optimal rule-in cut-off for FAST score of ≥0.78 with a PPV of 0.70 (sensitivity: 0.64, specificity 0.94, positive LR: 10.39). However, using these cut-offs also, 27 (45%) patients continued to remain in the grey zone (FAST: 0.55–0.78) (Table 4). But, of the 33 patients who were ruled in or ruled out by the cut-offs derived by us, 29 (87.9%) patients were correctly classified with respect to histologic findings. Thus, 48.3% of the patients in our study cohort could be correctly classified using the cut-offs derived by us. Overall, the cut-offs derived by us performed significantly better than those of Newsome et al6 in classifying patients with respect to the histologic findings of NASH + NAS ≥ 4 + F ≥ 2 (P < 0.001) (Table 5).
Table 5.
Accuracy of FAST Score Cut-offs for Correctly Classifying Indian Patients With NAFLD (n = 60).
| Classification | Using cut-offs by Newsome et al6 | Using optimal cut-offs derived in this study | P value |
|---|---|---|---|
| Correctly classified [true negative + true positive] | 15 (25%) [7 + 8] | 29 (48.3%) [22 + 7] | <0.001 |
| Grey zone | 28 (46.7%) | 27 (45%) | |
| Wrongly classified [false negative + false positive] | 17 (28.3%) [1 + 16] | 4 (6.7%) [1 + 3] |
NAFLD, nonalcoholic fatty liver disease; FAST, FibroScan-AST.
Predictive values depend on the prevalence of disease in the given population. This is important as the prevalence of NASH + NAS ≥ 4 + F ≥ 2 in the community is likely to be lower than that in our cohort and in the cohorts studied by Newsome et al.6 We thus calculated the predicted PPV and NPV of our rule-in and rule-out cut-offs at different anticipated prevalence of NASH + NAS ≥ 4 + F ≥ 2 (Supplementary Table 1). The predicted NPVs were robust (>0.8) over a wide range of anticipated prevalence (5–50%) and excellent (>0.9) at anticipated prevalence of less than 30%. However, the predicted PPVs were suboptimal at low prevalence of NASH + NAS ≥ 4 + F ≥ 2 (Supplementary Table 1).
Discussion
The prevalence of NAFLD in India ranges from 9% to 53% with a rural-urban divide.7,8 These figures are likely to increase in the near future due to increased urbanization, changing lifestyle and increasing incidence of obesity/overweight and diabetes mellitus in India.20 Moreover, Indians are more susceptible to insulin resistance and NAFLD even at a lower BMI.21 NAFLD has also emerged as one of the leading causes of cryptogenic cirrhosis and hepatocellular carcinoma in India.22, 23, 24 Evidence suggests that there are certain subtle differences between Indian patients with NAFLD and their Western counterparts. Indian patients are less likely to have morbid obesity, diabetes, hypertension or metabolic syndrome compared with Western patients with NAFLD.25,26 Early data suggesting that liver histology findings are relatively innocuous in Indian patients with NAFLD with mild-moderate degrees of inflammation and fibrosis have been challenged by the recent interim analysis of an ongoing real-life, multicentric observational study (Indian Consortium on NAFLD) and a histological study of a large number of patients showing presence of histological NASH in more than 60% of patients and advanced fibrosis in 35% patients.19,25, 26, 27, 28 Given the huge population of India, NAFLD is likely to snow ball into a public health problem with significant impact on the limited health resources in the country. An Indian position paper on NAFLD estimated that there are at least 25 million patients with NAFLD in the country with a potential risk for significant liver disease.7 It is of vital importance that we develop improved techniques that allow easier identification of the subset of patients who are likely to have progressive disease.
One of the important constraints in understanding NAFLD is the absence of a disease-specific biomarker. Liver biopsy is the only method for detecting ballooning and lobular inflammation which is paramount for differentiating NASH from NAFL. This distinction is important not only prognostically (NASH is generally considered the progressive form of NAFLD) but also in selecting patients for therapeutic clinical trials.2 Evidence suggests that underlying necroinflammatory activity promotes ongoing hepatic injury and fibrosis is the chief determinant of histologic and clinical outcomes.29,30 Liver biopsy remains the only method for simultaneous assessment of steatosis, inflammation and fibrosis. However, liver biopsy is an invasive procedure and is associated with a definite albeit small risk of major complications (1–3%) and mortality (0.01%).31 Moreover, the biopsy procedure samples <0.002% of the total liver volume, and sampling variability is a concern.32 Composite scores such as FIB-4 and NFS have been validated for predicting fibrosis in patients with NAFLD and showed good AUROCs for detecting advanced fibrosis (≥F3) on biopsy in our cohort. However, these scores cannot measure steatosis or necroinflammatory activity. As expected, FIB-4 and NFS showed poor accuracy for detecting NASH + NAS ≥ 4 + F ≥ 2. FibroScan not only assesses fibrosis (LSM) but also grades steatosis using CAP. By incorporating serum AST with FibroScan-derived parameters of LSM and CAP, the FAST score allows the simultaneous assessment of steatosis (CAP), inflammation (AST) and fibrosis (LSM).6 Given the similarities in the scoring parameters, the FAST score can be looked upon as a noninvasive counterpart of the biopsy-based SAF score.18 Similar to the observations of Newsome et al6, the FAST score showed good discrimination with an AUROC of 0.81 for the noninvasive identification of patients with NASH, significant activity and fibrosis (NASH + NAS ≥ 4 + F ≥ 2) in our cohort of Indian patients with biopsy-proven NAFLD. These are the patients with high risk of progressive disease who are likely to be benefitted by the emerging therapies in NASH. Our findings are further corroborated by a recent study, reporting an AUROC of 0.79 with the FAST score for detecting NASH + NAS ≥4 + F ≥ 2 in Indian patients with NAFLD.33
Using the cut-offs proposed by Newsome et al6 in our cohort, we found that the rule-out cut-off (FAST: ≤0.35) had an NPV of 0.88 with a negative LR of 0.64.6 However, the rule-in cut-off (FAST: ≥0.67) had a PPV of only 0.33 with a positive LR of 2.23. We attribute the suboptimal performance of the rule-in cut-offs derived by Newsome et al6 to the low prevalence of NASH + NAS ≥ 4 + F ≥ 2 (18.3%) in our cohort. In the study by Newsome et al6, similar observations were made in the China Wenzhou, French bariatric surgery and Malaysian validation cohorts where the prevalence of NASH + NAS ≥ 4 + F ≥ 2 was 9%, 15% and 20%, respectively. The PPVs of the rule-in cut-off (FAST: ≥0.67) in these 3 cohorts were 0.33, 0.63 and 0.54, respectively.6 Similarly, Anand et al33 recently reported poor PPV (0.43) and positive LR (2.53) with these cut-offs in their cohort of Indian patients with NAFLD where the proportion of NASH + NAS ≥ 4 + F ≥ 2 was 23%.
Any predictive model needs to be calibrated in accordance with the population under study to account for spectrum bias. It is well known that the prevalence of the disease affects the predictive values.34 We thus recalculated the optimal cut-off values in our cohort of patients. The optimal rule-out (FAST: ≤0.55) and rule-in (FAST: ≥0.78) cut-offs were associated with an NPV of 0.95 and a PPV of 0.70, respectively. The FAST score demonstrated good performance characteristics at these cut-offs with a negative LR of 0.21 (rule-out cut-off) and a positive LR of 10.39 (rule-in cut-off). Using these cut-offs, 48.3% of all the patients in our cohort and 87.9% of the patients lying outside the grey zone could be correctly classified with respect to their histologic findings.
Elastography usually performs better in ruling-out rather than in ruling-in significant fibrosis.35,36 Similarly, scrutiny of the NPV and PPV values observed by us and Newsome et al6 suggest that the FAST score may be better at ruling out patients with active necroinflammation and significant fibrosis. Supplementary Table 1 shows robust NPVs with the rule-out cut-off over a wide range of prevalence of NASH + NAS ≥ 4 + F ≥ 2, while the PPV of the rule-in cut-off is suboptimal at low prevalence. Thus, biopsy can be confidently avoided in patients falling in the rule-out zone of the FAST score. However, management of patients falling in the rule-in zone of the FAST score should be individualised, and liver biopsy should ideally be considered to confirm the presence of active necroinflammation and significant fibrosis (particularly in areas of low prevalence). This makes it ideally suited for primary healthcare settings by allowing proper vetting and need for referral to specialists (gastroenterologist or hepatologist) which is limited in India. The utility of the FAST score for monitoring response to treatment is an underexplored grey area. Preliminary evidence from paired biopsies in a cohort of 48 patients who underwent bariatric surgery suggested a good correlation between FAST score and improvement in NAS on histology at 1 year after surgery.33
One of the vexing issues with predictive models having two different threshold values is the problem of the grey zone. Indeed, 45% of our patients fell in the grey zone despite using optimised cut-offs. Tackling these patients will require the consideration of several factors including proximity to cut-offs (rule-in or rule-out), age and presence of metabolic risk factors. Repeating the FAST score after a reasonable interval or an alternative noninvasive test may be a rational option in a subgroup of such patients. However, we anticipate that a substantial proportion of these patients will require liver biopsy to accurately portray their disease severity in the spectrum of NAFLD.
The FAST score has several advantages which make it suitable for Indian settings. It is noninvasive and easy to use. Importantly, it is nonproprietary. However, it requires the use of a specific machine (FibroScan) which has important cost implications. However, compared with other elastographic techniques, performing FibroScan requires comparatively little training of the operator.35,36 The FAST score can correctly classify disease severity in more than 50% patients, thereby avoiding invasive liver biopsy.
In conclusion, results of our study validate the utility of the FAST score for detecting patients with NASH, significant fibrosis (≥F 2) and inflammation (NAS ≥ 4) on liver biopsy in Indian patients with NAFLD. However, our results suggest that cut-offs should be recalibrated based on the anticipated prevalence of NASH + NAS ≥ 4 + F ≥ 2 in the given population. Moreover, a substantial number of patients may still fall in the grey zone and may continue to require liver biopsy.
Credit authorship contribution statement
A.D. contributed to statistical analysis, histopathology reviews, conceptualization and critical revision; A.K., A.D. and S.M. contributed to manuscript writing; M.M. N.V., M.P. and S.T. contributed to data collection; V.S. contributed to critical revision.
Conflicts of interest
The authors have none to declare.
Funding
None.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.jceh.2021.06.008.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Flow-diagram of study.
Supplementary Figure 2.
Comparison of AUROCs of non-invasive assessment for detecting advanced fibrosis (≥F3) on biopsy.
References
- 1.Chalasani N., Younossi Z., Lavine J.E., et al. The diagnosis and management of nonalcoholic fatty liver disease: practice guidance from the American Association for the Study of Liver Diseases. Hepatology. 2018;67:328–357. doi: 10.1002/hep.29367. [DOI] [PubMed] [Google Scholar]
- 2.De A., Duseja A. Natural history of simple steatosis or nonalcoholic fatty liver. J Clin Exp Hepatol. 2020;10:255–262. doi: 10.1016/j.jceh.2019.09.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Konerman M.A., Jones J.C., Harrison S.A. Pharmacotherapy for NASH: current and emerging. J Hepatol. 2018;68:362–375. doi: 10.1016/j.jhep.2017.10.015. [DOI] [PubMed] [Google Scholar]
- 4.Manuel Echevarría J., León P., Pozo F., Avellón A. EASL-ALEH Clinical Practice Guidelines: non-invasive tests for evaluation of liver disease severity and prognosis. J Hepatol. 2015;63:237–264. doi: 10.1016/j.jhep.2015.04.006. [DOI] [PubMed] [Google Scholar]
- 5.Castera L., Friedrich-Rust M., Loomba R. Noninvasive assessment of liver disease in patients with nonalcoholic fatty liver disease. Gastroenterology. 2019;156 doi: 10.1053/j.gastro.2018.12.036. 1264-1281.e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Newsome P.N., Sasso M., Deeks J.J., et al. FibroScan-AST (FAST) score for the non-invasive identification of patients with non-alcoholic steatohepatitis with significant activity and fibrosis: a prospective derivation and global validation study. Lancet Gastroenterol Hepatol. 2020;5:362–373. doi: 10.1016/S2468-1253(19)30383-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chalmers J., Ban L., Leena K.B., et al. Cohort profile: the Trivandrum non-alcoholic fatty liver disease (NAFLD) cohort. BMJ Open. 2019;9 doi: 10.1136/bmjopen-2018-027244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Duseja A., Najmy S., Sachdev S., et al. High prevalence of non-alcoholic fatty liver disease among healthy male blood donors of urban India. JGH Open. 2019 Apr 1;3:133–139. doi: 10.1002/jgh3.12117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Third report of the National Cholesterol Education Program (NCEP) expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment Panel III) final report. Circulation. 2002;106:3143–3421. [PubMed] [Google Scholar]
- 10.Misra A., Misra R., Wijesuriya M., Banerjee D. The metabolic syndrome in South Asians: continuing escalation & possible solutions. Indian J Med Res. 2007;125:345–354. [PubMed] [Google Scholar]
- 11.Duseja A., Singh S.P., Saraswat V.A., et al. Non-alcoholic fatty liver disease and metabolic syndrome-position paper of the Indian National Association for the study of the liver, endocrine society of India, Indian college of cardiology and Indian society of gastroenterology. J Clin Exp Hepatol. 2015;5:51–68. doi: 10.1016/j.jceh.2015.02.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Karlas T., Petroff D., Sasso M., et al. Individual patient data meta-analysis of controlled attenuation parameter (CAP) technology for assessing steatosis. J Hepatol. 2017;66:1022–1030. doi: 10.1016/j.jhep.2016.12.022. [DOI] [PubMed] [Google Scholar]
- 13.Shah A.G., Lydecker A., Murray K., et al. Comparison of noninvasive markers of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol. 2009 Oct;7:1104–1112. doi: 10.1016/j.cgh.2009.05.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Angulo P., Hui J.M., Marchesini G., et al. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology. 2007;45:846–854. doi: 10.1002/hep.21496. [DOI] [PubMed] [Google Scholar]
- 15.Wong V.W.S., Vergniol J., Wong G.L.H., et al. Diagnosis of fibrosis and cirrhosis using liver stiffness measurement in nonalcoholic fatty liver disease. Hepatology. 2010;51:454–462. doi: 10.1002/hep.23312. [DOI] [PubMed] [Google Scholar]
- 16.Kleiner D.E., Brunt E.M., Van Natta M., et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. 2005;41:1313–1321. doi: 10.1002/hep.20701. [DOI] [PubMed] [Google Scholar]
- 17.Bedossa P., Poitou C., Veyrie N., et al. Histopathological algorithm and scoring system for evaluation of liver lesions in morbidly obese patients. Hepatology. 2012;56:1751–1759. doi: 10.1002/hep.25889. [DOI] [PubMed] [Google Scholar]
- 18.Hagström H., Nasr P., Ekstedt M., et al. SAF score and mortality in NAFLD after up to 41 years of follow-up. Scand J Gastroenterol. 2017;52:87–91. doi: 10.1080/00365521.2016.1230779. [DOI] [PubMed] [Google Scholar]
- 19.Rastogi A., Shasthry S.M., Agarwal A., et al. Non-alcoholic fatty liver disease – histological scoring systems: a large cohort single-center, evaluation study. APMIS. 2017;125:962–973. doi: 10.1111/apm.12742. [DOI] [PubMed] [Google Scholar]
- 20.Mitra S., De A., Chowdhury A. Epidemiology of non-alcoholic and alcoholic fatty liver diseases. Transl Gastroenterol Hepatol. 2020;5 doi: 10.21037/tgh.2019.09.08. 16–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Das K., Das K., Mukherjee P.S., et al. Nonobese population in a developing country has a high prevalence of nonalcoholic fatty liver and significant liver disease. Hepatology (Baltimore, Md) 2010 May;51:1593–1602. doi: 10.1002/hep.23567. [DOI] [PubMed] [Google Scholar]
- 22.Nayak N.C., Vasdev N., Saigal S., Soin A.S. End-stage nonalcoholic fatty liver disease: evaluation of pathomorphologic features and relationship to cryptogenic cirrhosis from study of explant livers in a living donor liver transplant program. Hum Pathol. 2010;41:425–430. doi: 10.1016/j.humpath.2009.06.021. [DOI] [PubMed] [Google Scholar]
- 23.Duseja A., Sharma B., Kumar A., et al. Nonalcoholic fatty liver in a developing country is responsible for significant liver disease. Hepatology. 2010;52:2248–2249. doi: 10.1002/hep.23838. [DOI] [PubMed] [Google Scholar]
- 24.Tohra S., Duseja A., Taneja S., et al. Experience with changing etiology and non-transplant curative treatment modalities for hepatocellular carcinoma in a real-life setting – a retrospective descriptive analysis. J Clin Exp Hepatol. 2021;11(6):682–690. doi: 10.1016/j.jceh.2021.02.002. Epub 2021 Feb 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Duseja A., Das A., Das R., et al. The clinicopathological profile of Indian patients with nonalcoholic fatty liver disease (NAFLD) is different from that in the west. Dig Dis Sci. 2007 Aug 7;52:2368–2374. doi: 10.1007/s10620-006-9136-y. [DOI] [PubMed] [Google Scholar]
- 26.Duseja A., Das A., Dhiman R.K., et al. Indian patients with nonalcoholic fatty liver disease presenting with raised transaminases are different at presentation. World J Gastroenterol. 2007;13:649–650. doi: 10.3748/wjg.v13.i4.649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Madan K., Batra Y., Gupta S.D., et al. Non-alcoholic fatty liver disease may not be a severe disease at presentation among Asian Indians. World J Gastroenterol. 2006;12:3400–3405. doi: 10.3748/wjg.v12.i21.3400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Duseja A., Singh S.P., Mehta M., et al. Changing clinicopathological profile of nonalcoholic fatty liver disease in India - interim results of the ICON-D (Indian Consortium on NAFLD) Study. Hepatol Int. 2021 (Abstract), In press. [Google Scholar]
- 29.Angulo P., Kleiner D.E., Dam-Larsen S., et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology. 2015;149 doi: 10.1053/j.gastro.2015.04.043. 389-397.e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ekstedt M., Hagström H., Nasr P., et al. Fibrosis stage is the strongest predictor for disease-specific mortality in NAFLD after up to 33 years of follow-up. Hepatology. 2015;61:1547–1554. doi: 10.1002/hep.27368. [DOI] [PubMed] [Google Scholar]
- 31.Bravo A.A., Sheth S.G., Chopra S. Liver biopsy. N Engl J Med. 2001;344:495–500. doi: 10.1056/NEJM200102153440706. [DOI] [PubMed] [Google Scholar]
- 32.Ratziu V., Charlotte F., Heurtier A., et al. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology. 2005;128:1898–1906. doi: 10.1053/j.gastro.2005.03.084. [DOI] [PubMed] [Google Scholar]
- 33.Anand A., Elhence A., Vaishnav M., et al. FibroScan–aspartate aminotransferase score in an Asian cohort of non-alcoholic fatty liver disease and its utility in predicting histological resolution with bariatric surgery. J Gastroenterol Hepatol. 2021;36:1309–1316. doi: 10.1111/jgh.15358. [DOI] [PubMed] [Google Scholar]
- 34.Poynard T., Muntenau M., Morra R., et al. Methodological aspects of the interpretation of non-invasive biomarkers of liver fibrosis: a 2008 update. Gastroenterol Clin Biol. 2008;32(6 suppl 1):8–21. doi: 10.1016/S0399-8320(08)73990-3. [DOI] [PubMed] [Google Scholar]
- 35.Friedrich-Rust M., Poynard T., Castera L. Critical comparison of elastography methods to assess chronic liver disease. Nat Rev Gastroenterol Hepatol. 2016;13:402–411. doi: 10.1038/nrgastro.2016.86. [DOI] [PubMed] [Google Scholar]
- 36.Kapur S., Kalra N., Bhatia A., et al. Comparison of elastography point quantification with transient elastography in patients with chronic viral hepatitis and nonalcoholic fatty liver disease: a pilot study. J Clin Exp Hepatol. 2021;11:21–29. doi: 10.1016/j.jceh.2020.06.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Flow-diagram of study.