Prevalence and genetics of non-alcoholic fatty liver disease

Non-alcoholic fatty liver disease (NAFLD) is a heterogeneous condition and at least two different forms exist. ‘Obese/Metabolic NAFLD’ characterizes subjects with metabolic syndrome and insulin resistance, whereas NAFLD associated with the common PNPLA3 I148M variant (‘PNPLA3 NAFLD’) is not accompanied by insulin resistance. This thesis was undertaken to elucidate the pathogenetic mechanisms of NAFLD (Studies I and III) and to estimate the prevalence of non-alcoholic steatohepatitis in Finnish subjects (Study II). The study groups consisted of 417 (I) and 372 (III) subjects, as well as 296 bariatric surgery patients from Finland and 2849 subjects from a population-based D2D-sample (II). In Study II, 380 non-bariatric surgery patients who had undergone a liver biopsy for suspected NAFLD from Italy were used as an external validation cohort. In Study I liver fat content was measured by proton magnetic resonance spectroscopy (1H-MRS), in Study II by either biopsy or 1H-MRS, and in Study III in 75% by 1H-MRS and in 25% by biopsy. Clinical characteristics (I, II, III), plasma apoC3 concentrations (I), and serum cytokeratin 18 fragments (II) were determined. In Study III, lipidomic analyses were performed using ultra-performance liquid chromatography coupled to time-of-flight mass spectrometry (UPLC-MS). Individuals were genotyped for rs2854116 and rs2854117 in APOC3 (I) and the known rs738409 in PNPLA3 (I, II, III). In Study II, we developed and validated a ‘NASH score’ in the Finnish and Italian biopsy cohorts, which was then used to predict NASH prevalence in the population-based D2D study. In Study III, the subjects were divided into groups based on PNPLA3 genotype or obesity, and the absolute and relative triacylglyceride concentrations were compared between the subgroups. No difference existed between the APOC3 variant allele (T-455C or C-482T or both) carriers and non-carriers in liver fat or apoC3 concentrations, whereas those with the PNPLA3 GG genotype at rs738409 had a 2.7-fold higher liver fat content than those with the CC genotype (I). The ‘NASH score’ included PNPLA3 genotype, aspartate aminotransferase (AST), and fasting insulin. The area under the ROC for this score was 0.774 (95% CI 0.709-0.839) in Finns and 0.759 (95% CI 0.711-0.807) in Italians (NS) (II). The prevalence of NASH based on this score in the D2D study was 6.0% (95% CI 5.0-6.9%) (II). Sensitivity analysis was performed by a Bayesian model, which gave a NASH population prevalence of 3.6% (95% CI 0.2-7.7%) using the ‘NASH Score’ (II). Absolute and relative deficiency of distinct circulating TAGs was observed in the PNPLA3148MM/148MI as compared with the PNPLA3148II group (III). Genotypes in the obese and ‘non-obese’ groups were similar but the obese subjects were insulin-resistant (III). Liver fat was increased in both obese and PNPLA3148MM/148MI groups (III). Multiple changes in the relative TAG concentrations were observed between obese and ‘non-obese’ groups (III). These closely resembled those between obese subjects with ‘obese NAFLD’ versus the ‘PNPLA3 NAFLD’ (III). Variation in liver fat cannot be explained by genetic variants in APOC3 (I). The population-based prevalence of NASH in Finnish subjects is ~5% (II). The circulating TAG profile depends on the aetiology of NAFLD (III). In ‘PNPLA3 NAFLD’
a relative deficiency of TAGs is observed, supporting the idea that the I148M variant prevents lipolysis rather than stimulates TAG synthesis in the liver (III).