Bivariate Mixture Models of Serum Ferritin and Transferrin Saturation Predict Stable Components Measured 12 Years Apart in a Healthy Australian Population

Abstract 5281

Bivariate mixture modeling was used to analyze joint population distributions of transferrin saturation (TS) and serum ferritin concentration (SF) measured in the Hemochromatosis and Iron Overload Screening (HEIRS) Study (McLaren et al.Translational Research 2008 151(2), 97−109). Four components (C1, C2, C3, and C4) were identified with successively age−adjusted increasing means for TS and SF from data contributed by 26,832 African Americans, 12,620 Asians, 12,264 Hispanics, and 43,254 whites. We used data from the Australian “HealthIron” study of genetic and environmental modifiers of hereditary hemochromatosis to validate the mixture model approach for component analysis and to determine whether the component distributions appear to be stable over time or whether an individual's TS and SF values move from one component to another. Between 2004−2006, a sample of participants of northern European descent was selected from the Melbourne Collaborative Cohort Study (n=31,192 (17,951 women) recruited 1990–1994) to participate in the HealthIron study (n=1,438 (783 women)). Of these, 1,052 (579 women) participated in CAPI interviews and clinical examination. We applied the bivariate mixture modeling approach to TS and SF data from HealthIron. Follow−up data were excluded for individuals who were treated after baseline evaluations. The EMMIX program was used to predict component membership for HealthIron participants using the model based on the HEIRS data alone. Component transition probabilities over time were estimated as observed proportions. Analyses were performed separately for baseline and follow−up data. Four components (C1, C2, C3, C4) with successively age−adjusted increasing means for TS and SF, were identified using baseline data from 926 whites (426 men, 500 women) and using follow−up data from 771 whites (341 men, 430 women). At baseline, the largest component, C2, had normal mean TS (28% for women, 35% for men) and SF (139 mg/L for women, 309 μg/L for men), higher than the upper 95% confidence limits for means in analyses of HEIRS data. At follow−up, the largest component, C2, had normal mean TS (29% for women, 33% for men) and SF (124 μg/L for women, 183 μg/L for men) consisting of component proportions 0.64 for women and 0.72 for men. C3 and C4 had progressively higher mean values for TS and SF with progressively lower component proportions. At baseline, C1 had mean TS values of 15% for women (27% for men), and mean SF values of 23 mg/L for women (59 mg/L for men), similar to those found using HEIRS data. At follow−up, C1 had mean TS values of 18% for women and men, and mean SF values of 21 μg/L for women and 20 μg/L for men. Only female C282Y homozygotes for the iron gene, HFE, showed evidence that component transition probabilities shifted significantly over time;19 of 49 (39%) had TS and SF values that were in the same bivariate component at baseline and follow−up (Test of Symmetry, p=0.014). As for analyses of data from the HEIRS Study, a mixture of four components was found in data from HealthIron participants suggesting that the model is transferable from one white population to another, although estimated means within components may differ. The longitudinal aspect of this study illustrates that, with the exception of female C282Y homozygotes, the components of the mixture distributions are largely stable over time.