Remaley, Alan T.

Abstract Cholestatic liver diseases, such as primary sclerosing cholangitis and primary biliary cirrhosis, can lead to serum accumulation of lipoprotein X (LpX). LpX is a multilamellar particle high in cholesterol but lacking structural apolipoproteins A1 or B. The absence of ApoB results in no negative feedback on cholesterol biosynthesis and prevents LpX clearance from the liver. While clinical signs and symptoms typically precede laboratory findings, it is possible that in medically complex patients the identification of LpX could be the first observation of cholestatic liver disease. Traditional laboratory methods are insufficient to identify LpX as it is of similar density to low-density lipoprotein (LDL). LpX contains a high concentration of cholesterol which is erroneously reported as LDL-C by routine clinical methods. As LpX is a rare complication of liver disease, clinicians may presume the elevation is a coincidental familial hypercholesterolemia rather than a sequela of liver disease. Currently, lipoprotein gel electrophoresis is the only laboratory method to identify LpX. In this method only performed in specialty lipid laboratories, LpX is readily identified by its unique reverse electrophoretic mobility relative to other lipoproteins. The objective of this study was to characterize lipid panels from LpX-positive samples and develop a suitable mechanism to identify LpX-containing samples with good clinical validity. From 21,377 clinical electrophoresis results reported between Nov 2011 to Nov 2021, LpX was identified in 157 serum samples. Overall, patients with LpX were younger (median 44y vs. 55y, p<0.0001) with significantly higher total cholesterol (812mg/dL vs 190mg/dL, p<0.0001) and lower high density lipoprotein-cholesterol (HDL-C; 3mg/dL vs 45mg/dL, p<0.0001). Data were randomly split (70/30) into training (n=14,964) and testing (n=6,413) cohorts. Receiver operator characteristic curve analysis identified optimal thresholds of 22.5 mg/dL HDL-C (AUC = 0.94) and 378.5 mg/dL total cholesterol (AUC = 0.91), as well as a nonHDL-C/HDL-C ratio of 9.2 (AUC = 0.995). Applying these cutoffs to the testing cohort achieved a sensitivity/specificity of 98%/81% for HDL-C, 96%/87% for total cholesterol, and 98%/94% for nonHDL-C/HDL-C ratio. A multivariate model combining these three parameters showed a sensitivity/specificity of 97%/85%, respectively. In conclusion, low HDL-C, elevated total cholesterol and a ratio of nonHDL-C/HDL-C >9.2 are associated with the presence of LpX. The ratio of nonHDL-C/HDL-C is the most sensitive and specific predictor of LpX. If confirmed in other cohorts, laboratories could include a reporting comment on lipid panels with a nonHDL-C/HDL-C ratio >9.2 cautioning a high suspicion for the presence of LpX and recommending confirmatory testing. It may also be prudent to caution that LDL-C results may not be accurate. In conclusion, identifying patients with high suspicion of LpX based on abnormal lipid panel results may aide in clinical diagnosis, even when an assay to detect LpX is not readily available.

Abstract BACKGROUND Dyslipoproteinemias can be classified by their distinct lipoprotein patterns, which helps determine atherosclerotic cardiovascular disease (ASCVD) risk and directs lipid management but this has required advanced laboratory testing. OBJECTIVE To develop a new algorithm for classifying lipoprotein disorders that only relies on the standard lipid panel. METHODS Lipid thresholds for defining the different lipoprotein phenotypes were derived for Non-High-Density Lipoprotein-Cholesterol (NonHDL-C) and Triglycerides (TG) to be concordant when possible with the current US Multi-Society guidelines for blood cholesterol management. RESULTS The new classification method categorizes patients into all the classical Fredrickson-like phenotypes except for Type III dysbetalipoproteinemia. In addition, a new hypolipidemic phenotype (Type VI) due to genetic mutations in apoB-metabolism is described. The validity of the new algorithm was confirmed by lipid analysis by NMR (N = 11,365) and by concordance with classification by agarose gel electrophoresis/beta-quantification (N = 5504). Furthermore, based on the Atherosclerosis Risk in Communities (ARIC) cohort (N = 14742), the lipoprotein phenotypes differ in their association with ASCVD (TypeV > IIb > IVb > IIa > IVa > normolipidemic) and can be used prognostically as risk enhancer conditions in the management of patients. CONCLUSIONS We describe a clinically useful lipoprotein phenotyping system that is only dependent upon the standard lipid panel. It, therefore, can be easily implemented for increasing compliance with current guidelines and for improving the care of patients at risk for ASCVD.