Elevated Levels of Circulating γ‐Tocopherol as a Surrogate Marker of Mortality Risk in Populations of Adult Men Stratified by Health Index Score

Background Previous research has demonstrated that circulating levels of γ‐tocopherol (γT) increase in response to nutrient deficiencies, poor diet, and conditions associated with increased risk of death. Because of its novel chemistry, its unique response to inflammation i n vivo , and previously described relation to adherence to the DASH diet,* the utility of elevated levels of γT as a surrogate population‐level marker of mortality risk in men was investigated. Approach A Health Index (HI) was developed utilizing seven physiological and demographic variables for which evidence of an association with poor health outcomes exists ( Table 1). Levels considered to be healthy were assigned a value of one whereas values considered to be unhealthy were assigned a value of zero and intermediate levels a value of 0.5 ( Table 1). Plasma values and demographic data from control subjects of a prostate cancer study described previously** (637 control men**; mean age at blood draw = 68.64 years) from the Multiethnic Cohort, for which there was no missing data, were used to determine a HI score (range from 0, unhealthy – 7, healthy) for each individual. The relation of the index score to mortality and γT was assessed by determining odds ratios for all‐cause mortality and hyper‐γ‐tocopherolemia (γT ≥2 μg/ml) for six stratified groups of men with HI scores varying from 0 to 7. Odds ratios and p values were calculated by two‐sided Fisher's exact test. Results As shown in Table 2, individuals with the lowest HI scores (<2) were significantly more likely to die than those with HI scores >5 (OR = 7.00; p <0.0001) during the period of observation (median = 6.3 years). The group with the lowest HI score also had a significantly greater proportion of men found to be hyper‐γ‐tocopherolemic, OR = 6.92 (p< 0.0001). Intermediate HI scores were generally associated with intermediate odds for both death and hyper‐γ‐tocopherolemia. No significant differences in age were observed between groups in Table 2. Additional analysis of 317 matched subjects that subsequently developed prostate cancer in the same cohort study yielded similar results, consistent with the observations on the control subjects. Assessment of years lost/100 years of observation showed control men with HI scores of <3 lost 7.98 years/100 years observed compared to 2.03/100 in those scoring HI >5 (p<0.0001). Figure 1 shows the change in mean plasma γT as a function of HI score. Conclusions Hyper‐γ‐tocopherolemia may be a useful surrogate marker for comparing the risk of premature mortality between older male populations in cohort studies and for intervention trials designed to lower mortality risk. γT provides a population marker for which it is predicted that interventions which alter mortality risk may have a direct impact on circulating γT levels. Further studies utilizing larger populations, women, additional risk factors, and modifiable health behaviors are needed to verify the general utility of hyper‐γ‐tocopherolemia as a biomarker of a population's mortality risk. Support or Funding Information This work was supported in part by National Cancer Institute grantsR03 CA132149, P01 CA33619, R37 CA54281, S10 RR020890, and P30CA71789. 1 Selected parameters used to create a Health Index (HI) score.Risk Factor Values Health Index ScoreSmoking Status Nonsmoker 1.0 Former 0.5 Current 0BMI (kg/m 2 ) 20–25 1.0 25–30 0.5 <20 or >30 0Dietary Vitamin E (mg/day) >15 1.0 8–15 0.5 <8 0Plasma pro‐Vitamin A Carotenoids (ng/ml) >450 1.0 300–450 0.5 <300 0Plasma Vitamin D (nM) >80 1.0 65–80 0.5 <65 0Plasma CRP (mg/L) <1 1.0 1–3 0.5 >3 0Plasma α‐Tocopherol (μg/ml) 12–18 1.0 >18 0.5 <12 02 Association of HI score with mortality and hyper‐γT. Subjects were assigned composite HI scores based on Table 1. Odds ratios for all‐cause mortality and hyper‐γT as categorical variables were determined for each stratified group (<2; 2.5‐3; 3.5–4; 4.5; 5; >5.5).Mean HI Score Subjects (N) Deaths (N) Hyper‐γ‐T (N) Mortality (OR) p Value Hyper‐γ‐T (OR) p value1.60 57 12 37 7.00 0.0008 6.92 <0.0001 2.78 117 12 65 3.00 0.069 4.68 <0.0001 3.77 182 10 77 1.53 NS 2.74 0.0002 4.50 87 1 27 0.31 NS 1.68 NS 5.00 85 3 20 0.96 NS 1.15 NS 5.86 109 4 23 1.0 (Ref) ‐ 1.0 (Ref) ‐