Abstract
BACKGROUND: Abnormal circulating lipid levels have been suggested in relation to lymphoid malignancy (LM) risk. METHODS: We studied UK Biobank participants (n = 403,625) with serum data for cholesterol (total [TC], high-density lipoprotein [HDL], direct low-density lipoprotein [LDL]), triglycerides (TG), and apolipoproteins A1 and B (ApoA1, ApoB). We conducted principal component (PC) analysis and multivariate Cox regression models to estimate hazard ratio (HR) overall, by lipid-lowering drug use and follow-up interval. RESULTS: During an average of 10.5 years of follow-up, 3006 incident LMs occurred (including 667 multiple myelomas [MM], 2193 non-Hodgkin lymphomas [NHL]). Among medication non-users, most lipid levels were inversely associated with risk of most endpoints (HR(Q4vsQ1)range: 0.37 to 0.79), especially closer to diagnosis. In contrast LDL/HDL ratio and PC1 (highly loaded in LDL and ApoB) were consistently positively associated with chronic/small lymphocytic leukaemia risk in each follow-up period and with NHL and B-cell NHL risk within 5 years. Further, LD, ApoB and TG levels were positively associated with MM risk after 10+ years (HR(1-SD)range = 1.26 to 1.60). CONCLUSION: Lipid depletion closer to LM diagnosis might reflect cancer cell metabolism and warrants further work examining individuals with precursor conditions. The MM-specific long-term risk might reflect the known MM-obesity association.