Bradley Tucker1, Sonia Sawant2, Hannah McDonald3, Kerry-Anne Rye4, Sanjay Patel5, Kwok Leung Ong4, Blake J Cochran6. 1. Heart Research Institute, Sydney, Australia; Faculty of Medicine and Health, University of Sydney, Sydney, Australia; School of Medical Sciences, UNSW, Sydney, Australia. 2. Royal Prince Alfred Hospital, Sydney, Australia. 3. Bankstown-Lidcombe Hospital, Sydney, Australia. 4. School of Medical Sciences, UNSW, Sydney, Australia. 5. Heart Research Institute, Sydney, Australia; Faculty of Medicine and Health, University of Sydney, Sydney, Australia; Royal Prince Alfred Hospital, Sydney, Australia. 6. School of Medical Sciences, UNSW, Sydney, Australia. Electronic address: b.cochran@unsw.edu.au.
Abstract
BACKGROUND AND AIMS: There is some evidence of a cross-sectional, and possibly causal, relationship of lipid levels with leukocyte counts in mice and humans. This study investigates the cross-sectional and longitudinal relationship of blood lipid and lipoprotein levels with leukocyte counts in the UK Biobank cohort. METHODS: The primary cross-sectional analysis included 417,132 participants with valid data on lipid measures and leukocyte counts. A subgroup analysis was performed in 333,668 participants with valid data on lipoprotein(a). The longitudinal analysis included 9058 participants with valid baseline and follow-up data on lipid and lipoprotein levels and leukocyte counts. The association of lipid and lipoprotein levels with leukocyte counts was analysed by multivariable linear regression. RESULTS: Several relationships were significant in both cross-sectional and longitudinal analysis. After adjustment for demographic, socioeconomic and other confounding factors, a higher eosinophil count was associated with lower HDL cholesterol and apolipoprotein A-I concentration (p < 0.001). Higher triglycerides levels were associated with higher total leukocyte, basophil, eosinophil, monocyte and neutrophil counts (all p < 0.01). A higher lymphocyte count was associated with a higher apolipoprotein B level (p < 0.001). In the longitudinal analysis, lipoprotein(a) was inversely associated with basophil count in men but not women (p < 0.001). CONCLUSIONS: Triglyceride levels demonstrate a robust positive association with total and differential leukocyte counts suggesting they may be directly involved in leukogenesis. However, unlike in murine models, the remainder of these relationships is modest, which suggests that cholesterol and lipoproteins are minimally involved in leukogenesis in humans.
BACKGROUND AND AIMS: There is some evidence of a cross-sectional, and possibly causal, relationship of lipid levels with leukocyte counts in mice and humans. This study investigates the cross-sectional and longitudinal relationship of blood lipid and lipoprotein levels with leukocyte counts in the UK Biobank cohort. METHODS: The primary cross-sectional analysis included 417,132 participants with valid data on lipid measures and leukocyte counts. A subgroup analysis was performed in 333,668 participants with valid data on lipoprotein(a). The longitudinal analysis included 9058 participants with valid baseline and follow-up data on lipid and lipoprotein levels and leukocyte counts. The association of lipid and lipoprotein levels with leukocyte counts was analysed by multivariable linear regression. RESULTS: Several relationships were significant in both cross-sectional and longitudinal analysis. After adjustment for demographic, socioeconomic and other confounding factors, a higher eosinophil count was associated with lower HDL cholesterol and apolipoprotein A-I concentration (p < 0.001). Higher triglycerides levels were associated with higher total leukocyte, basophil, eosinophil, monocyte and neutrophil counts (all p < 0.01). A higher lymphocyte count was associated with a higher apolipoprotein B level (p < 0.001). In the longitudinal analysis, lipoprotein(a) was inversely associated with basophil count in men but not women (p < 0.001). CONCLUSIONS: Triglyceride levels demonstrate a robust positive association with total and differential leukocyte counts suggesting they may be directly involved in leukogenesis. However, unlike in murine models, the remainder of these relationships is modest, which suggests that cholesterol and lipoproteins are minimally involved in leukogenesis in humans.