A physician's perseverance uncovers problems in a key nephrology study.

The Normal Hematocrit Cardiac Trial, published in 1998, was a foundational study testing erythropoietin analog treatment to normal hematocrit targets. It served as a warning that erythropoietin replacement was not a panacea. Its large size gave it disproportionate weighting in evidence reviews and guideline development and thereby impacted treatment decisions. Coyne shows that the published results did not completely and clearly represent the study's actual results. We discuss the implications and make recommendations to prevent such occurrences.