Fatty Liver in India: 42.5%, Even in People Who Aren't Overweight

Most Indians who find out they have fatty liver were not being investigated for it. A scan ordered for something else showed a bright liver. The figure usually quoted alongside that finding is 42.5%.

The part that surprises people most is that a meaningful share of Indian cases occur in people whose weight and BMI look entirely normal. What follows is what the 42.5% is an estimate of, how much of the disease sits at a normal BMI, what each test on the pathway can and cannot settle, and why your report may use one of three different names for the same condition.

Key Takeaways:

• The 42.5% figure is a weighted average of 37 separate studies with 99.9% heterogeneity. Three different meta-analyses give 42.5%, 38.6% and 26.9%, all from the same condition, measured differently.
• Lean MASLD (BMI below 23 kg/m²) accounts for roughly one patient in four at Indian liver centres. Lean Asian Indians carry two to three times the insulin resistance and twice the hepatic fat of matched Caucasian individuals at the same BMI.
• A normal ultrasound does not exclude milder disease. Normal liver enzymes do not exclude fatty liver. Both are stated explicitly in the guidelines.
• FIB-4 below 1.3 has 85 to 95% negative predictive value for advanced fibrosis, but the score is unreliable below age 35 and loses specificity above 65. No India-specific validation exists.
• Indian reports may say NAFLD, MASLD or MAFLD for the same condition. The name changed internationally in June 2023. India has not formally adopted the change.

The figure comes from a 2025 systematic review and meta-analysis in the Journal of Gastroenterology and Hepatology by James and colleagues at Queen Mary University of London, covering the epidemiology of metabolic dysfunction-associated steatotic liver disease in South Asian ethnicities. The team searched four databases from January 2002 to November 2023 and pooled 62 studies covering 40,745 participants from India, Sri Lanka, Pakistan and Bangladesh. Fifty-three of those studies were India-based, and the 37 of them that reported prevalence produced the India-specific pooled estimate of 42.5%, with a 95% confidence interval running from 31.7% to 53.3%.

Nobody measured 42.5% of India. The number is a weighted average of 37 separate studies conducted in different settings, on different populations, using different diagnostic methods.

The authors report heterogeneity of 99.9% for that estimate, meaning the underlying studies do not agree with each other at all. The confidence interval is the honest version of the headline: somewhere between roughly a third and just over half.

The urban and rural figures that usually travel with the 42.5% deserve more caution than they get. The 47.1% urban and 18.5% rural estimates are South Asia-wide, not India-only breakdowns, and they rest on a lopsided base of 33 urban studies against only 5 rural ones.

Egger's test found strong evidence of publication bias for the rural estimate. The gap is best read as suggestive of a heavier urban metabolic risk-factor burden, not as a measured contrast between an Indian city and an Indian village.

A reader who searches this will meet more than one number, and the differences are not errors.

• James et al., 2025: 53 India studies, 37 reporting prevalence: pooled estimate 42.5% (CI 31.7 to 53.3). Urban 47.1% vs rural 18.5%, South Asia-wide.
• Shalimar et al., 2022: 50 studies, 23,581 adults: pooled estimate 38.6% (CI 32 to 45.5). Hospital-based 40.8% vs. community 28.2%.
• South Asia meta-analysis, 2023: 22 studies, 15,455 participants: pooled estimate 26.9% (CI 18.9 to 35.8). Urban 32.9% vs rural 22.6%, difference not statistically significant.

The 2022 Indian analysis also found 52.8% prevalence in high-risk groups against 28.1% in average-risk ones, which is the same signal the setting split carries: where a study recruited decides much of what it found. The 2023 South Asia analysis found its urban-rural difference was not statistically significant, which sits awkwardly beside the gap quoted from the 2025 paper.

In Indian adults with type 2 diabetes, a 2026 systematic review of 18 Indian studies covering 81,364 participants pooled prevalence at 56.9%, and reported 55.6% by ultrasonography against 64.3% by FibroScan, so the choice of machine moved the answer by nearly nine points. Indian adolescents sit at the other extreme, estimated at 3.57% in a Global Burden of Disease-based analysis. That is an order of magnitude below the adult figures, and the literature offers no clean explanation for the gap.

Before any lean-disease percentage means anything, the threshold behind it has to be stated. Asian-Indian consensus places the lean-to-overweight boundary at a BMI of 23 kg/m², against 25 in Western criteria, and the overweight-to-obese boundary at 25 rather than 30. Someone at a BMI of 24 is therefore lean by Western criteria and overweight by Indian ones, and that single definitional choice swings the reported share of lean cases substantially.

Using the 23 kg/m² boundary, Indian hospital-based studies put lean disease at between 11% and 31.7% of all fatty liver cases at tertiary liver centres. A 2024 Delhi cross-sectional study of 6,146 participants found MASLD in 56.4% of the sample, of which lean cases made up 11.3%. A central-India study combining ultrasound with FibroScan found fatty liver in 20.5% of lean adults, with significant fibrosis in 11% of those.

Across South Asia, non-obese individuals accounted for 43.4% of the total disease population in one pooled analysis. An older Indian study reporting 75% of cases as non-obese circulates in reviews, but its original methodology could not be independently verified and it reads as a single outlying finding.

The ethnic comparison is where the Indian picture separates from the Western one. Reviews put lean fatty liver at between 5% and 34% worldwide, with the highest rates, above 30%, reported in India. Lean, non-diabetic, non-smoking ethnic Asian Indians have been shown to carry two to three times the insulin resistance and twice the hepatic triglyceride content of matched Caucasian, Hispanic, Black and East Asian individuals at the same BMI.

An ultrasound-positive rate of 23% in a population with relatively low BMI but high visceral fat is a more alarming finding than the same rate in an obese population. Three mechanisms recur in the Indian literature:

• More visceral fat at a given BMI
• Lower relative muscle mass, which is a matter for lean mass measurement rather than assumption
• Insulin resistance, which is repeatedly identified as the dominant driver and is present even in lean patients without overt metabolic syndrome

Genetics is the weakest link in that chain. A South Indian study found TM6SF2 significantly associated with disease, a PNPLA3 variant reached significance only in North East Indian ancestry, and a Kerala genome-wide association study of 908 participants found the usual European and East Asian risk variants performed far less strongly. There is no settled South Asian genetic risk profile to report.

This is a real split in the literature, not a case of one side being out of date, and different well-designed studies point in different directions.

Evidence that lean MASLD carries higher mortality: A 2024 updated meta-analysis of 14 studies covering 94,181 patients found lean disease carried a 61% higher all-cause mortality risk, independent of age, sex and cardiometabolic risk factors. A separate meta-analysis found a hazard ratio of 1.34 for all-cause mortality and 2.14 for liver-specific mortality. A US cohort found higher overall mortality in lean patients despite lower rates of both cirrhosis and diabetes.

Evidence that the difference is smaller than it looks: A meta-analysis of 10 cohort studies covering 109,151 patients found no significant difference in all-cause mortality. A histology-focused meta-analysis found lean patients had less severe fibrosis and steatohepatitis. Several cohort analyses find that once age is controlled for, because lean patients tend to be older at diagnosis, the difference disappears entirely.

The caveat that matters most for an Indian reader is that none of this evidence is Indian. Every one of those mortality comparisons comes from US, European or East Asian cohorts, and the Indian lean phenotype differs mechanistically from the Western one, so the transfer cannot be assumed in either direction.

In June 2023, a multisociety Delphi consensus involving 236 panellists from 56 countries replaced non-alcoholic fatty liver disease with metabolic dysfunction-associated steatotic liver disease, and NASH with MASH. Steatotic liver disease became the umbrella term, covering MASLD, alcohol-related liver disease, and a new middle category, MetALD, for people with metabolic dysfunction who also drink moderately.

Under the new framework, a MASLD diagnosis requires hepatic steatosis plus at least one of five cardiometabolic criteria:

• BMI at or above 25 kg/m² (or 23 for Asian populations), or a raised waist circumference
• Fasting glucose at or above 100 mg/dL, or treated diabetes
• Blood pressure at or above 130/85 mmHg, or treatment for it
• Triglycerides at or above 150 mg/dL, or lipid-lowering treatment
• HDL cholesterol below defined sex-specific thresholds

Studies suggest 95% to 99% of the old population is captured by the new definition.

Indiahas not moved with it. INASL stated in its 2023 guidance paper that the change "may create confusion not only among hepatologists but also among physicians and non-hepatologists", and that it continues to recommend NAFLD and NASH terminology within India until adoption is uniform.

A separate 2023 INASL paper had earlier proposed MAFLD as an interim term. No ICMR position statement on nomenclature was located, which is a genuine gap. As of 2026, an Indian report may reasonably carry any of the three names, and a change in the word is not a change in the condition.

Fatty liver diagnosis in India involves several tests, each serving a different purpose and carrying its own limitations. The diagnostic pathway typically begins with ultrasound, followed by blood-based fibrosis assessment and further investigations where needed.

The main tests include:

• Ultrasound: The default first test in India, used alone in 28 of 44 Indian prevalence studies. It reliably detects fat only when liver fat content reaches roughly 20% to 30%, so a normal scan does not exclude milder disease.
• ALT and AST: Liver enzymes widely tested alongside ultrasound. Normal enzyme levels do not exclude fatty liver. The 2025 meta-analysis explicitly excluded studies diagnosing the condition on enzymes alone.
• FIB-4: A blood-based fibrosis risk score calculated from age, AST, ALT and platelet count. A score below 1.3 has an 85% to 95% negative predictive value for advanced fibrosis. Scores between 1.3 and 2.67 are indeterminate, while scores above 2.67 warrant hepatology referral.
• Transient elastography: Measures liver fat and stiffness together. It was preferred by 91.6% of 609 surveyed Indian gastroenterologists and hepatologists. Stiffness values range from below 7 kPa for no or mild fibrosis to above 11.6 kPa for cirrhosis, although thresholds vary by study and equipment.
• Biopsy: Used when non-invasive test results are indeterminate or discordant. Only 25.3% of clinicians in the same survey reported using it.

Two important limitations apply to these assessments:

• FIB-4 has age boundaries. It is unreliable below age 35, where younger age lowers the score. Above age 65, AASLD 2023 and EASL 2024 guidance use a modified threshold of 2.0 instead of 1.3 to preserve specificity.
• India-specific validation is lacking. The FIB-4 cut-offs and elastography thresholds were derived from Western, Japanese or East Asian cohorts. Indian clinicians therefore use thresholds established in other populations.

FIB-4 typically comes before elastography to help identify people at low risk of advanced fibrosis and avoid unnecessary additional testing.

According to INASL's 2023 position, most fatty liver cases in India are discovered incidentally during abdominal ultrasounds performed for other reasons. Screening is recommended for people with type 2 diabetes, obesity or metabolic syndrome rather than the general population. No standalone ICMR screening guideline for fatty liver was located.

A fatty liver finding is a marker of metabolic risk more than a verdict on the liver itself. In Western cohort data, it is associated with:

• Roughly double the risk of incident type 2 diabetes over a median five years, rising with more severe steatosis
• A clear difference in 10-year all-cause mortality by fibrosis stage: 7.7% at stages F0 to F2 against 41.5% at cirrhosis
• Advanced fibrosis in 14.0% of South Asian patients with biopsy data, higher than comparable UK and US estimates (though this may reflect which patients get biopsied)

These are associations from Western and mixed-ethnicity cohorts, not causal chains established in Indian populations.

None of those absolute risks were measured in Indian cohorts. For liver cancer, the 2025 meta-analysis found only one South Asian study reporting an annual rate, too few to pool.

The useful response to a grade 1 finding is not to look up a survival curve. Take the report to a registered medical practitioner and establish two things: which of the five cardiometabolic criteria are actually met in your case, and whether a fibrosis risk score has been calculated rather than inferred from the steatosis grade.

If a normal-weight reader takes one thing from the Indian data, it is that the scan finding does not have to be explained away by the number on the scale.

Found an incidental fatty liver on a scan? A Voy clinician can read the finding alongside your metabolic markers. Book an assessment.

This article is for general information and education only and is not medical advice, diagnosis, or treatment. GLP-1 and other medications referenced are prescription-only and are appropriate only for certain people under the supervision of a qualified clinician. Do not start, stop, or change any medication based on this article. Please consult a registered medical practitioner about your individual circumstances. Information reflects what was available at the time of review and may change.