Background: The clinical presentation and outcomes of non-alcoholic fatty liver disease (NAFLD)-related hepatocellular carcinoma are unclear when compared with hepatocellular carcinoma due to other causes. We aimed to establish the prevalence, clinical features, surveillance rates, treatment allocation, and outcomes of NAFLD-related hepatocellular carcinoma.

Methods: In this systematic review and meta-analysis, we searched MEDLINE and Embase from inception until Jan 17, 2022, for articles in English that compared clinical features, and outcomes of NAFLD-related hepatocellular carcinoma versus hepatocellular carcinoma due to other causes. We included cross-sectional and longitudinal observational studies and excluded paediatric studies. Study-level data were extracted from the published reports. The primary outcomes were (1) the proportion of hepatocellular carcinoma secondary to NAFLD, (2) comparison of patient and tumour characteristics of NAFLD-related hepatocellular carcinoma versus other causes, and (3) comparison of surveillance, treatment allocation, and overall and disease-free survival outcomes of NAFLD-related versus non-NAFLD-related hepatocellular carcinoma. We analysed proportional data using a generalised linear mixed model. Pairwise meta-analysis was done to obtain odds ratio (OR) or mean difference, comparing NAFLD-related with non-NAFLD-related hepatocellular carcinoma. We evaluated survival outcomes using pooled analysis of hazard ratios.

Findings: Of 3631 records identified, 61 studies (done between January, 1980, and May, 2021; 94 636 patients) met inclusion criteria. Overall, the proportion of hepatocellular carcinoma cases secondary to NAFLD was 15·1% (95% CI 11·9-18·9). Patients with NAFLD-related hepatocellular carcinoma were older (p<0·0001), had higher BMI (p<0·0001), and were more likely to present with metabolic comorbidities (diabetes [p<0·0001], hypertension [p<0·0001], and hyperlipidaemia [p<0·0001]) or cardiovascular disease at presentation (p=0·0055) than patients with hepatocellular carcinoma due to other causes. They were also more likely to be non-cirrhotic (38·5%, 27·9-50·2 vs 14·6%, 8·7-23·4 for hepatocellular carcinoma due to other causes; p<0·0001). Patients with NAFLD-related hepatocellular carcinoma had larger tumour diameters (p=0·0087), were more likely to have uninodular lesions (p=0·0003), and had similar odds of Barcelona Clinic Liver Cancer stages, TNM stages, alpha fetoprotein concentration, and Eastern Cooperative Oncology Group (ECOG) performance status to patients with non-NAFLD-related hepatocellular carcinoma. A lower proportion of patients with NAFLD-related hepatocellular carcinoma underwent surveillance (32·8%, 12·0-63·7) than did patients with hepatocellular carcinoma due to other causes (55·7%, 24·0-83·3; p<0·0001). There were no significant differences in treatment allocation (curative therapy, palliative therapy, and best supportive care) between patients with NAFLD-related hepatocellular carcinoma and those with hepatocellular carcinoma due to other causes. Overall survival did not differ between the two groups (hazard ratio 1·05, 95% CI 0·92-1·20, p=0·43), but disease-free survival was longer for patients with NAFLD-related hepatocellular carcinoma (0·79, 0·63-0·99; p=0·044). There was substantial heterogeneity in most analyses (I²>75%), and all articles had low-to-moderate risk of bias.

Interpretation: NAFLD-related hepatocellular carcinoma is associated with a higher proportion of patients without cirrhosis and lower surveillance rates than hepatocellular carcinoma due to other causes. Surveillance strategies should be developed for patients with NAFLD without cirrhosis who are at high risk of developing hepatocellular carcinoma.

Funding: None.