Lean Indian patients with non-alcoholic fatty liver disease (NAFLD) have less metabolic risk factors but similar liver disease severity as non-lean patients with NAFLD

Although most patients with NAFLD are obese or overweight, some are lean with normal BMI. Our aim was to assess differences in clinicopathological profile and liver disease severity among lean and non-lean NAFLD. Data of 1040 NAFLD patients over last 10 years was analysed. BMI < 23 kg/m2 categorised lean patients. Non-invasive assessment of steatosis was done by ultrasound and controlled attenuation parameter (CAP) while fibrosis was assessed with FIB-4 and liver stiffness measurement (LSM). FibroScan-AST (FAST) score was used for non-invasive prediction of NASH with significant fibrosis. Histology was reported using NASH-CRN system. 149 (14.3%) patients were lean while 891 (85.7%) patients were non-lean. Diabetes mellitus [25 (16.7%) vs 152 (17.05%), p > 0.99], elevated triglycerides [81 (54.3%) vs 525 (58.9%), p = 0.33] and low HDL [71(47.6%) vs 479(53.7%), p = 0.18] were observed in a similar proportion. Lean patients were less likely to have central obesity [72 (48.3%) vs 788 (88.4%), p < 0.001], hypertension [16 (10.7%) vs 239(26.8%), p < 0.001] and metabolic syndrome [21 (14.09%) vs 290 (32.5%), p < 0.001]. No difference in steatosis assessment was noted using ultrasound (p = 0.55) or CAP (0.11). FAST [0.38 (0.18–0.66) vs 0.39 (0.27–0.73), p = 0.53], FIB-4 [1.08 (0.65–1.91) vs 1.09 (0.66–1.94), p = 0.94] and LSM [6.1 (4.8-7.9) vs 6.2 (4.7–8.6), p = 0.19) were similar. Liver biopsy was available in 149 patients [lean: 19 (12.7%), non-lean: 130 (87.3%)]. There was no difference in the number of patients with NASH [4 (21.05%) vs 20 (15.3%), p = 0.51], significant fibrosis [2 (10.5%) vs 32 (24.6%), p = 0.25] or advanced fibrosis [1 (5.26%) vs 18 (13.84%), p = 0.47]. Although metabolic co-morbidities are less common, there is no difference in liver disease severity among both groups.


INTRODUCTION
Non-alcoholic fatty liver disease (NAFLD) has emerged as a major cause of chronic liver disease worldwide.In India, the burden of NAFLD is alarmingly high with an estimated pooled prevalence of 38.6% compared to a global prevalence of approximately 25.2% [1,2].The histological spectrum of NAFLD broadly comprises of non-alcoholic fatty liver, non-alcoholic steatohepatitis (NASH), NASH with fibrosis, cirrhosis and hepatocellular carcinoma [3].Traditionally, NASH has been considered to be the progressive form of NAFLD.Current knowledge of the natural history of NAFLD overwhelmingly suggests that underlying fibrosis is the principal determinant of clinical outcomes [4].
Although obesity is one of the main risk factors for NAFLD, some patients with NAFLD have normal body mass index (BMI), a phenotype commonly referred to as lean NAFLD.Studies suggest that these patients have less metabolic risk factors compared to overweight or obese patients with NAFLD [5][6][7].However, the severity of underlying liver disease and prognosis of lean NAFLD viz-a-viz NAFLD of the classical phenotype remains a grey area with conflicting evidence.Data from the Indian subcontinent exploring this aspect is extremely scanty [8,9].Granular understanding of the severity of NAFLD in lean patients is crucial for determining their linkage to care and optimising overall management.This is particularly relevant now that NAFLD has been incorporated into the National Program on Prevention and Control of Cancer, Diabetes, Cardiovascular disease and Stroke (NPCDCS) under the aegis of the Ministry of Health and Family Welfare of India [10].In this study, we evaluated the prevalence of normal BMI in a well characterised cohort of Indian patients with NAFLD and explored the differences in clinicopathological profiles and severity of underlying liver disease in lean patients compared to their non-lean counterparts.

METHODS
In this cross-sectional observational study, data of all adult patients with NAFLD managed prospectively in a real-life fashion over the last 10 years (January 2011 till December 2021) at a tertiary academic centre in north India was collected.NAFLD was diagnosed in the presence of hepatic steatosis inferred on ultrasound, FibroScan [controlled attenuation parameter (CAP) > 248 dB/m] or liver biopsy]in the absence of significant alcohol consumption (less than 20 g/day irrespective of gender) and other etiologies of liver disease including but not limited to chronic viral hepatitis, autoimmune hepatitis, celiac disease, Wilson's disease and haemochromatosis.Patients were included if they had data required for calculating FIB-4 [including age, serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), and platelet count] Patients with decompensated liver cirrhosis or hepatocellular carcinoma were excluded from the study.The study was approved by the institutional ethical committee and has been reported in accordance with the STROBE guidelines [11].
Demographic and clinical details were noted including the presence of metabolic co-morbidities like type 2 diabetes mellitus (T2DM), hypertension (HTN) and dyslipidemia.Anthropometric details including body mass index (BMI) and waist circumference measured with a non-stretchable tape at the midpoint between the top of the iliac crest and the last palpable rib were noted.Data of blood investigations including complete blood counts, liver function test, lipid and glycemic profile at the time of diagnosis were collected.Results of ultrasonography and Fibroscan [Echosens (Paris)] including CAP and liver stiffness measurement (LSM) were also recorded.Histological details were recorded in patients who underwent liver biopsy.
Significant elevation of liver enzymes was defined as elevated AST or ALT more than 2 times of the upper normal limit of 40 U/L.The presence of NASH along with significant fibrosis was inferred non-invasively using the FibroScan-AST (FAST) score in patients in whom AST, CAP and LSM values were available.The FAST score was calculated using the formula: [e^(−1.65 + 1.07*In(LSM) + 2.66*10 −8 *CAP 3   [16,17].

Histologic assessment of liver disease severity
Liver histology was reported by a single central pathologist (AsD) with experience in reporting hepatic histopathology.Hepatic steatosis, ballooning, lobular inflammation and fibrosis were reported in accordance with the NASH clinical research network (NASH-CRN) system [20,21].Definite histological NASH was defined as NAFLD activity score (NAS) ≥ 5 while borderline NASH was defined as NAS of 3-4.Significant and advanced fibrosis were defined as ≥F2 and ≥F3, respectively.

Outcomes
Our primary outcome of interest was the prevalence of normal BMI (<23 kg/m 2 ) among patients with NAFLD.Secondary outcomes included the differences in metabolic profile and liver disease severity between lean and non-lean patients with NAFLD.

Sample size calculation
In a recently published, large multicentric Indian study of 3553 patients with NAFLD, 10.6% of the patients were lean with a normal BMI [22].In order to evaluate a similar proportion of patients at a confidence level of 99% with precision of estimate 0.025, the minimum required sample size was calculated as 1011 patients.

Statistical analysis
GraphPad Prism v9.3.1 (GraphPad Software, San Diego, California USA) was used for statistical analysis.Quantitative data was expressed as mean ± standard deviation or median (interquartile range).Normality of data was assessed using D'Agostino-Pearson test.Nominal and ordinal data were expressed as proportions and percentages.Quantitative data were compared using t-test or Mann-Whitney test as applicable Qualitative data was assessed using Fisher's exact test or chi square test for trend as applicable.Multivariate logistic regression analysis for predictors of advanced fibrosis was done by incorporating the variables of age, BMI, abdominal obesity, T2DM, HTN, dyslipidemia and metabolic syndrome in a stepwise model where variables were entered into the model if p < 0.05 and removed if p > 0.1.Multicollinearity was assessed using variance inflation factors (VIF) with VIF > 10 being considered as suggestive of strong multicollinearity.All statistical tests were performed two sided with p ≤ 0.05 being considered as statistically significant.

RESULTS
One thousand and forty patients with NAFLD met the eligibility criteria and were included in this study.The average age of the patients was 40.9 ± 11.34 years and majority were males (58%).One hundred and forty-nine (14.3%) patients were lean while 891 (85.7%) patients were non-lean [overweight: 194 (18.6%), obese: 697 (67%) as shown in Fig. 1.Age of lean and non-lean patients were similar [38.5 ± 12 years vs 41.46 ± 11.1 years, p = 0.10].There was no difference in gender distribution among lean and non-lean patients as shown in Table 1.
Liver disease severity in lean and non-lean NAFLD on noninvasive assessment (Table 2) Hepatobiliary ultrasound parameters and data for calculation of FIB-4 was available in all patients.FibroScan parameters of CAP and LSM were available in 197 (54 lean and 143 non-lean) and 333 (82 lean and 251 non-lean) patients, respectively.

Risk factors for advanced fibrosis
As LSM and liver histology were not available in all patients, FIB-4 was used for assessment of risk factors for advanced fibrosis.BMI was not a significant predictor for ruling out (FIB-4 < 1.3) or ruling in (FIB-4 > 2.67) advanced fibrosis as shown in Tables 4 and 5

DISCUSSION
Prevalence of patients with normal BMI in our cohort of Indian patients with NAFLD was 14.3%.Globally, the prevalence of lean NAFLD among patients with NAFLD has been estimated to be around 20% [5,6].In Indians, the pooled proportion of lean individuals among patients with NAFLD is 16.97% while the pooled prevalence of lean NAFLD in the general community is 6.5%.The reported proportion of lean patients in previous Indian studies on NAFLD ranges from 6.8% to 31.7% [22][23][24].A part from regional variations, the wide range in the reported proportion of lean patients in Indian studies is also partially attributable to differences in the modalities for the diagnosis of NAFLD and BMI cut-off used for defining lean patients.None the less, the available evidence suggests that although uncommon, the actual burden of lean NAFLD in India is still likely to be substantial given the extremely high prevalence of NAFLD in India.Larger, multicentric studies are required to better estimate the true burden of lean NAFLD.Further, community-based studies are the need of the hour to assess the prevalence of lean NAFLD among the general population.
Recent western studies have suggested that lean patients with NAFLD are younger than their obese counterparts [25,26].However, we did not observe any difference in age between our lean and non-lean patients.Similar findings have been reported in two other studies from north and south India [8,27].This suggests that Indian lean patients are detected with NAFLD at an older age than western patients.It could also be indicative of a referral bias due to the common perception among clinicians that NAFLD in individuals with normal body weight is mild.Bulk of the available evidence suggests that metabolic comorbidities like central obesity, hypertension, T2DM, and metabolic syndrome are less common in lean patients with NAFLD compared to obese patients [5][6][7].Similarly, our patients with lean NAFLD were less likely to have central obesity, hypertension, and metabolic syndrome compared to non-lean patients.In our cohort, elevated triglycerides and low HDL were observed in a similar proportion of lean and non-lean patients with Novelties is line with the findings of a meta-analysis that reported no difference in elevated triglycerides or dyslipidemia among lean and obese patients [5].This is an intriguing observation.Similar risk of dyslipidemia among lean and non-lean patients may be indicative of increased adipose tissue insulin resistance with consequential increase in peripheral lipolysis and serum lipid profile derangements despite the quantitatively lower total body fat mass in lean patients [8].Contrary to the available evidence, we did not observe any difference in T2DM among lean and non-lean patients.This is difficult to explain and may be related to the low prevalence of T2DM in our cohort.
We used multiple non-invasive techniques for assessing the various histologic facets of NAFLD to amplify the robustness of our inferences.Biochemical liver function perturbations as assessed by AST and ALT were similar between the lean and non-lean patients in our cohort.We did not observe any difference in the severity of steatosis on non-invasive assessment by ultrasound and CAP.Non-invasive detection of NASH, the diagnosis of which has both prognostic and therapeutic implications, has long been an unmet met.The FAST score encompasses measures of steatosis (CAP), inflammation (AST) and fibrosis (LSM) and may be considered a non-invasive analogue of the histological steatosis, activity and fibrosis (SAF) score [16].It was recently shown by Newsome et al. to have a high discriminatory ability for the detection of patients with NASH and significant fibrosis and has been validated in multiple geographic regions including India [16,17,28].There was no difference in FAST scores among the lean and non-lean patients in our cohort.With respect to fibrosis, FIB-4 and LSM were similar among lean and non-lean patients and there was no difference in the proportion of patients in whom advanced fibrosis was ruled out or ruled in using both these modalities.Our observations strongly suggest that liver disease in lean patients with NAFLD is not milder than that in their non-lean counterparts.
Our observations on non-invasive assessment were further corroborated on histology although it was available in only 149 patients.We observed no difference in steatosis, lobular inflammation, ballooning, NAS or fibrosis stages among lean and non-lean patients.Further, there was no difference in the number of patients with NASH, significant or advanced fibrosis between these two groups.Although some studies have suggested that lean or non-obese patients NAFLD have relatively mild disease on histology, there are several reports to the contrary [5,8,9,25,[29][30][31][32][33][34][35].Two other studies from the Indian subcontinent have assessed liver histology in lean or non-obese patients with NAFLD and both reported that these patients do not have a lower risk of NASH or advanced fibrosis [8,9].
There are few important caveats in interpreting the available evidence on the severity of liver disease in lean patients with NAFLD.There is a variation in the BMI cut-off used for defining  where an apparent link between BMI and clinical events like T2DM, hepatocellular carcinoma and cardiovascular events disappeared after adjustment for age [26].Previous longitudinal studies have also demonstrated that NAFLD in lean patients can progress to severe liver disease and clinical events may occur without substantial increase in BMI on follow-up [26,35].Taken together, the evidence suggests that a BMI-guided approach to risk stratification in patients with NAFLD may not be prudent.Further long-term follow-up studies are needed to get a better understanding of the natural history and prognosis in lean patients with NAFLD.We acknowledge the limitations of our study including the unavailability of CAP, LSM and histological data in all patients.This reflects the constraints of real-world clinical practice as it is not feasible to perform liver biopsy or FibroScan in all patients due to logistic and ethical concerns.In our institute, CAP function in FibroScan was incorporated in 2015.Hence, CAP parameter was not available in patients who presented prior 2015.Assessment of insulin resistance and measurement of other markers of metabolic dysfunction like serum uric acid and hs-CRP were not done as these are not a part of the routine evaluation and management of patients of NAFLD at our institute.Lastly, the inclusion of a control group of patients without NAFLD would have added further granularity to our observations and inferences.
In conclusion, lean patients with NAFLD are less likely to have abdominal obesity, hypertension and metabolic syndrome compared to their non-lean counterparts.However, there is no difference in biochemical parameters, non-invasive assessment of liver disease severity, and the histologic presence of NASH and significant or advanced fibrosis among lean and non-lean Indian patients with NAFLD.

Table 1 .
Demographic and metabolic profile of lean and non-lean patients with NAFLD (n = 1040).

Table 2 .
Non-invasive assessment of liver disease severity in lean and non-lean patients with NAFLD (n = 1040).
AST aspartate aminotransferase, ALT alanine transaminase, CAP controlled attenuation parameter, LSM liver stiffness measurement.A.De et al.

Table 4 .
[36]ictors for ruling out advanced fibrosis (FIB-4 < 1.3) on univariate and multivariate analysis (n = 1040).heterogeneitytothe evidence.As such this BMI based categorisation into lean, overweight and NAFLD is partly arbitrary.Hence, we assessed the risk of advanced fibrosis, the main determinant of clinical outcomes, with BMI as a continuous measure.We did not find BMI to be a significant predictor for ruling out or ruling in advanced fibrosis as assessed by FIB-4 scores thereby suggesting that BMI does not impact the risk of advanced fibrosis.Age is another confounding factor in assessing liver disease severity in lean patients.On the one hand there is a correlation between age and BMI while on the other age is a known risk factor for fibrosis and poor clinical outcomes[36].