Morooka H, Tanaka A, Inaguma D, et al. - Dialysis patients were clustered by adopting an unsupervised machine learning method, using phosphate, hemoglobin, and iron-related markers. An improved prognosis in two of the clusters might be achieved due to oral use of a phosphate binder.

• This observational study included 1,175 patients with chronic kidney disease who started dialysis between October 2011 and September 2013.

• Based on phosphate, hemoglobin, logarithmic ferritin, and transferrin saturation, the cohort was split into three clusters, and cluster c showed higher mortality than cluster a.

• In addition, a lower risk of all-cause death in relation to use of phosphate binders was evident in two clusters (a and c) that were characterized by older age and higher prevalence of diabetes mellitus, among others.