Abstract
Background: Hepatic de novo lipogenesis (DNL) is ideally measured in very low-density lipoprotein (VLDL)-triacylglycerol (TAG). In the fasting state, the majority of plasma TAG typically represents VLDL-TAG; however, the merits of measuring DNL in total plasma TAG have not been assessed. This study aimed to assess the performance of DNL measured in VLDL-TAG (DNL(VLDL-TAG)) compared to that measured in total plasma TAG (DNL(Plasma-TAG)).Methods: Using deuterated water, newly synthesised palmitate was determined in fasting plasma VLDL-TAG and total TAG in 63 subjects taking part in multiple studies resulting in n = 123 assessments of DNL (%new palmitate of total palmitate). Subjects were split into tertiles to investigate if DNL(Plasma-TAG) could correctly classify subjects having 'high' (top tertile) and 'low' (bottom tertile) DNL. Repeatability was assessed in a subgroup (n = 16) with repeat visits.Results: DNL(VLDL-TAG) was 6.8% (IQR 3.6-10.7%) and DNL(Plasma-TAG) was 7.5% (IQR 4.0%-11.0%), and the correlation between the methods was r(s) = 0.62 (p < 0.0001). Bland-Altman plots demonstrated similar performance (mean difference 0.81%, p = 0.09); however, the agreement interval was wide (-9.6% to 11.2%). Compared to DNL(VLDL-TAG), 54% of subjects with low DNL were correctly classified, whilst 66% of subjects with high DNL were correctly classified using DNL(Plasma-TAG). Repeatability was acceptable (i.e. not different) at the group level, but the majority of subjects had an intra-individual variability over 25%.Conclusion: DNL in total plasma TAG performed similarly to DNL in VLDL-TAG at the group level, but there was large variability at the individual level. We suggest that plasma TAG could be useful for comparing DNL between groups.