Serum retinyl esters are positively correlated with analyzed total liver vitamin A reserves collected from US adults at time of death

Abstract

Background: Minimal human data exist on liver vitamin A (VA) compared with serum biomarkers. Cutoffs of 5% and 10% total serum VA as retinyl esters (REs) suggest a VA intoxication diagnosis.

Objectives: We compared total liver VA reserves (TLRs) with the percentage of total serum VA as REs to evaluate hypervitaminosis with the use of US adult autopsy samples. Secondary objectives evaluated serum retinol sensitivity, TLRs among lobes, and hepatic α-retinol concentrations, an α-carotene cleavage product.

Design: Matched serum and liver samples were procured from cadavers (n = 27; mean ± SD age: 70.7 ± 14.9 y; range: 49-101 y). TLRs and α-REs were quantified by ultra-performance liquid chromatography. Pearson correlations showed liver and serum associations. Sensitivity and specificity were calculated for >5%, 7.5%, and 10% total serum VA as REs to predict TLRs and for serum retinol <0.7 and 1 μmol/L to predict deficiency.

Results: Serum RE concentrations were correlated with TLRs (r = 0.497, P < 0.001). Nine subjects (33%) had hypervitaminosis A (≥1.0 μmol VA/g liver), 2 of whom had >7.5% total serum VA as REs; histologic indicators corroborated toxicity at 3 μmol/g liver. No subject had >10% total serum VA as REs. Serum retinol sensitivity to determine deficiency (TLRs <0.1 μmol VA/g) was 83% at 0.7 and 1 μmol/L. Hepatic α-retinol was positively correlated with age (P = 0.047), but removing an outlier nullified significance.

Conclusions: This study evaluated serum REs as a biomarker of VA status against TLRs (gold standard), and abnormal histology suggested that 7.5% total serum VA as REs is diagnostic for toxicity at the individual level in adults. The long-term impact of VA supplements and fortificants on VA status is currently unknown. Considering the high prevalence of hypervitaminotic TLRs in this cohort, and given that many countries are adding preformed VA to processed products, population biomarkers diagnosing hypervitaminosis before toxicity are urgently needed. This trial was registered at clinicaltrials.govas NCT03305042.

Figures

FIGURE 1

CONSORT chart for sample requests and procurement from cadavers (n = 27). Samples were requested over a 4-y period. Equal numbers from each age range and sex were requested. A total of 13 male and 14 female donors were received. Larynx data will be published elsewhere. CONSORT, Consolidated Standards of Reporting Trials.

FIGURE 2

Representative histology of liver biopsies from human donors in VA-deficient, VA-sufficient, and hypervitaminotic/toxic statuses. Tissue sections were stained with hematoxylin and eosin (left panels) and Masson's trichrome (right panels). The VA-deficient donor had a pathologist-confirmed diagnosis of hepatic cirrhosis. Insets show polynuclear inflammatory cell infiltrate and bilirubin deposit (yellow) in the VA-deficient condition and lipid deposits (white) and stellate cell hypertrophy in the hypervitaminotic A condition; black arrows indicate dense collagen deposition (blue) associated with advanced cirrhosis in the VA-deficient condition; black arrowheads indicate normal collagen deposition associated with connective tissue of the portal triad in the VA-sufficient condition and the central vein in the hypervitaminosis A condition. Scale bar, 50 µm (insets, 25 µm). CV, central vein; pt, portal trial; VA, vitamin A.

FIGURE 3

Scatterplots of relations with total liver vitamin A reserves (n = 27). (A) Serum retinyl esters compared with total liver vitamin A reserves (r = 0.497, P = 0.008). (B) Serum retinol concentrations compared with total liver vitamin A reserves represented by the black linear regression line (r = 0.00, P = 0.95). The gray lines are the generally accepted cutoffs for vitamin A deficiency. (C) Percentage total serum vitamin A as retinyl esters compared with total liver vitamin A reserves (r = 0.43, P < 0.001). The solid gray lines are the current cutoffs for hypervitaminosis A (≥1.0 μmol/g liver) and the cutoffs of 5% used for children (19) and 10% used for adults (1) for percentage total serum vitamin A as retinyl esters. The dashed lines are those proposed by the current analyses to indicate vitamin A toxicity in these adult cadavers (i.e., 7.5% total serum vitamin A as retinyl esters and 3 μmol/g liver). Linear regression was used to assess relations.