LAUSR

BACKGROUND Cardiovascular risk factors control is fluctuating, tends to change over time, and is potentially impacted by multifactorial interactions. Currently, the existence of risk factors, rather than their variability or interplay with one another, is used to define the population at risk. The association between variability of risk factors and cardiovascular morbidity and mortality risk among patients with T2DM remains debatable. METHODS Using registry-derived data, we identified 29,471 people with T2D, without CVD at baseline, and with at least five measurements of risk factors. Variability for each variable was expressed as quartiles of the standard deviation during three years (exposure). The incidence of myocardial infarction, stroke, and all-cause mortality was assessed during 4.80 (2.40-6.70) years following the exposure phase. The association between measures of variability and the risk of developing the outcome was investigated through multivariable Cox proportional-hazards regression analysis with stepwise variable selection. Then, the recursive partitioning and amalgamation (RECPAM) algorithm was used to explore the interaction among the variability of risk factors associated with the outcome. RESULTS An association between the variability of HbA1c, body weight, systolic blood pressure, and total cholesterol with the outcome considered was found. Among the 6 classes of risk identified by RECPAM, patients with a high variability of both body weight and blood pressure had the highest risk (Class 6, HR = 1.81; 95% CI 1.61-2.05) compared with patients with low variability of both body weight and total cholesterol (Class 1, reference), despite a progressive reduction in the mean level of risk factors during successive visits. Individuals with high weight variability but low-moderate systolic blood pressure variability (Class 5, HR = 1.57; 95% CI 1.28-1.68), patients with moderate/high weight variability associated with high/very high HbA1c variability (Class 4, HR = 1.33; 95%CI 1.20-1.49), subjects with moderate/high weight variability and with low/moderate HbA1c variability (Class 3, HR = 1.12; 95%CI 1.00-1.25), as well as those with low weight variability associated with high/very high total cholesterol variability (Class 2, HR = 1.14; 95%CI 1.00-1.30) also showed a significant increase in the risk of event. CONCLUSIONS Combined high variability of two risk factors, particularly body weight and blood pressure, is associated with cardiovascular risk among patients with T2DM. These findings highlight the importance of continuous balancing of multiple risk factors.