Skip to main content
Fig. 3 | Lipids in Health and Disease

Fig. 3

From: Lipidomic risk score independently and cost-effectively predicts risk of future type 2 diabetes: results from diverse cohorts

Fig. 3

Validation of the LRS in the AusDiab cohort. a Direct application of the LRS derived from SAFHS cohort to the participants from AusDiab cohort. Left panel shows incidence rate ratio for future T2D associated with one standard deviation change in the LRS. The results are from three Poisson regression models that used duration of follow-up as an exposure variable: U – unadjusted; C – adjusted for clinical covariates; P – adjusted for presence of prediabetes. The clinical covariates used for adjustment in the multivariate models were: age, sex, systolic and diastolic blood pressure, body mass index, total cholesterol, HDL cholesterol, serum triglycerides, family history of diabetes, and use of anti-hypertensive and lipid-lowering drugs. The Right Panel shows three bar charts each of which depicts improvement in the indicated parameter upon addition of the LRS to the indicated regression model. Rotated numbers at the top of the bars are p-values. ΔAIC, improvement in Akaike Information Criterion; IDI, integrated discrimination improvement; NRI, continuous net reclassification index. b LRS recalibrated for AusDiab. Based on the results of Poisson regression analyses in the AusDiab, a recalibrated LRS was calculated as follows: 0.2259*i(Cer(d18:0/18:0))-0.2397*i(LPC 22:1) + 0.3267*i(TG(16:0_18:0_18:1)) where i(L) represents the inverse-normalized plasma concentration of lipid species L. Key to the plots in (Panel b) is the same for those in (Panel a)

Back to article page