Sex differences in the association between albumin and all-cause and vascular mortality.

BACKGROUND: Low serum albumin levels are associated with cardiovascular disease and mortality risk. This study evaluated the predictive value of low serum albumin for all-cause-mortality in a large Viennese patient cohort and investigated sex differences in the association between serum albumin and mortality. MATERIALS AND METHODS: Serum albumin concentrations of 285 930 patients, who attended the General Hospital Vienna between 1992 and 2002, were evaluated and linked with the Austrian Death Registry. The median observation period was 7.4 +/- 4.0 years and the death rate was 16.8%. For Cox regression analysis, albumin levels were divided into deciles, the highest category served as reference value. To analyse associations between albumin and mortality independent of liver function, results were adjusted for cholinesterase, which indicates protein synthesis capacity of the liver. RESULTS: Hazard ratios for all-cause-mortality increased linearly with decreasing albumin levels from 1.05 in the 9th to 2.98 in the 1st decile. Adjusted for cholinesterase, the relative risk for mortality was still 1.91 in the lowest category. Compared with women, men had an average 50% increased risk of death in almost every decile, adjusting for cholinesterase reduced the sex difference to a 10-20% higher mortality risk for men. In critically ill patients, hazard ratios for all-cause-mortality ranged from 4.5 in the 9th decile to 9.5 in the lowest albumin category. CONCLUSION: This study demonstrates a strong inverse association between serum albumin and mortality in a large patient cohort. The predictive value of low albumin was remarkably higher in men than in women.

Evidence of a U-shaped association between factor XII activity and overall survival.

INTRODUCTION: The clinical relevance of decreased coagulation factor XII (FXII) plasma activity as a risk factor for both venous and arterial thrombosis is still discussed controversially. The current study evaluated the predictive value of FXII levels for all-cause mortality in a large Viennese patient cohort. PATIENTS AND METHODS: Individuals, whose FXII activity levels were determined for suspected coagulation disorders or thrombophilia screening between 1991-2003 were included in this study (n = 8936, 51% male, 49% female, median age 43 years). Death/survival was determined by record linkage with the Austrian Death Registry. The median observation period was 5 years covering a total of 46 400 person years; the death rate was 17.1%. For Cox regression analysis, FXII plasma activity was divided into 11 categories of 10% steps with the category of > 100% FXII serving as a reference category. RESULTS: With decreasing FXII plasma activity, hazard ratios for all-cause mortality gradually increased linearly from 1.0 in the > 100% category to 1.5 (95% CI: 1.2-1.9) in the 80-90% category to 4.7 (95% CI: 3.4-6.5) in the 10-20% category. Similar results were obtained, when only vascular mortality or death as a result of ischemic heart disease was considered. No significant increase in all-cause mortality (HR: 1.4, 95%CI 0.7-2.8) was observed in the small group of FXII-deficient subjects [0-10% category (n = 58)]. CONCLUSIONS: This study first demonstrates a strong and almost linear association of FXII plasma activity between 90% and 10% with all-cause mortality in a large Viennese patient cohort. Interestingly, mortality rates are not increased when FXII activity is below 10%, resulting in a U-shaped survival curve.