Interpreting observational studies of disordered mineral metabolism and mortality in patients on hemodialysis

Abstract

This Practice Point commentary discusses the findings and limitations of a cohort study reported by Wald and colleagues of mineral metabolism in patients on hemodialysis. The investigators' observational analysis utilized data from the 1,846 patients in the randomized, controlled Hemodialysis (HEMO) Study. Wald et al.'s advantages include the well-characterized dataset, particularly with regard to comorbid conditions. However, the authors found it impossible to analyze the potentially confounding effect of concomitant medications. Furthermore, the relatively small dataset, especially compared with those of previous studies, limits the power of this study. In summary, Wald et al.'s findings support earlier studies confirming the importance of mineral metabolism as a risk factor for mortality in patients on hemodialysis. However, because of its low statistical power, its lack of analysis of potentially important confounders, and its observational design, the study cannot provide appropriate targets and should not serve as a justification for tolerating mild hyperphosphatemia or hypercalcemia or for overlooking the importance of preventing secondary hyperparathyroidism.