LAUSR

Background: Dyslipidaemia is associated with increased cancer risk. However, the prognostic value of visit-to-visit lipid variability (VVLV) is unexplored in this regard. Objective: To investigate the associations between VVLV and the risk of incident cancer. Design: Retrospective cohort study. Setting: Family medicine clinics. Patients: Adults attending a family medicine clinic in Hong Kong during 2000-2003, excluding those with <3 tests for low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides, and total cholesterol (TC) each, those with prior cancer diagnosis, and those with <1 year of follow-up. Measurements: Visit-to-visit LDL-C, HDL-C, TC, and triglycerides variabilities were measured by the coefficient of variation (CV). Patients were followed up until 31st December 2019 for the primary outcome of incident cancer. Results: Altogether, 69,186 patients were included (26,679 males (38.6%); mean age 60+/-13 years; mean follow-up 16+/-3 years); 7958 patients (11.5%) had incident cancer. Higher variability of LDL-C, HDL-C, TC, and TG was associated with higher risk of incident cancer. Patients in the third tercile of the CV of LDL-C (adjusted hazard ratio (aHR) against first tercile 1.06 [1.00, 1.12], p=0.049), HDL-C (aHR 1.37 [1.29, 1.44], p<0.001), TC (aHR 1.10 [1.04, 1.17], p=0.001), and TG (aHR 1.11 [1.06, 1.18], p<0.001) had the highest risks of incident cancer. Among these, only HDL-C variability remained associated with the risk of incident cancer in users of statins/fibrates. Limitations: Due to the observational nature of this study, there may be residual and unmeasured confounders. Patient data could not be individually adjudicated, implying that coding errors may be possible.