Higher serum β2-microglobulin levels are associated with better survival in chronic hemodialysis patients: a reverse epidemiology

Aims: beta(2)-Microglobulin (beta(2)-M) has been considered a surrogate marker of putative mid-molecular weight (MW) uremic toxins, compounds difficult to dialyze by low-flux dialysis membranes. This study was performed to evaluate the relationship between serum beta(2)-M and survival of chronic hemodialysis (CHD) patients and the association of beta(2)-M levels and factors associated with mortality. Methods: Part I of this study is a retrospective cohort evaluation that determined the relationship between beta(2)-M and mortality, and Part II is a cross-sectional study that evaluated the relationship between beta(2)-M and factors associated with mortality. Laboratory parameters, including beta(2)-M, albumin, prealbumin, creatinine, blood urea nitrogen (BUN), high-sensitivity C-reactive protein (hs-CRP), lipid battery, KT/V, and normalized protein nitrogen appearance (nPNA), were reviewed in Part I and measured in Part II. Clinical and demographic data, including age, sex, duration of hemodialysis, presence of cardiovascular disease, and presence of diabetes mellitus, were also recorded. Results: Part I: During the follow-up period of 5 years, there were 95 all-cause deaths among the 289 patients. Comparison of survivors and non-survivors indicated that serum beta(2)-M was higher in survivors (36.8 +/- 12.3 vs. 32.6 +/- 13.2 mu g/ml, p = 0.009). Kaplan-Meier analysis indicated that all-cause mortality in the lower beta(2)-M group was significantly higher compared to the higher beta(2)-M group (p < 0.0001). Multivariate Cox regression analyses indicated elevated beta(2)-M levels were significantly associated with lower mortality rate (relative risk: 0.608; 95% CI: 0.37 to 0.99; p = 0.046). Part II: Themean serum beta(2)-M concentration was 37.1 +/- 14.4 mu g/ml. Univariate analysis indicated that beta(2)-M was positively correlated with nPNA, duration of HD, BMI, and the concentrations of creatinine, albumin, BUN, and hs-CRP, but was negatively correlated with HDL-C concentration. Multiple regression analysis indicated that levels of nPNA (p < 0.001), duration of hemodialysis (p < 0.001), creatinine (p < 0.001), albumin (p = 0.006), BUN (p = 0.011), and HDL-C (p = 0.038) were independently associated with serum beta(2)-M concentration. Conclusion: Our results showed that higher serum beta(2)-M levels are associated with better survival in CHD patients and that nutritional status might be an independent predictor of serum beta(2)-M concentration in these patients.