Red blood cell distribution width (RDW), a widely available biomarker, independently predicts adverse outcomes in left-sided heart failure. The relation between RDW and death in pulmonary hypertension (PH) is unknown. In a prospective study of 162 consecutive patients with PH, RDW was recorded during initial diagnostic right-sided cardiac catheterization, and patients were followed for 2.1 +/- 0.8 years to determine vital status. Demographic, clinical, laboratory, and hemodynamic variables were compared by tertile of RDW. Cox proportional-hazards models were used to determine whether RDW was independently associated with death, and the prognostic utility of RDW was compared to that of other laboratory predictors, including N-terminal-pro-B-type natriuretic peptide (NT-pro-BNP). Of the 162 study patients, 78% were women, and 62% had pulmonary arterial hypertension. The mean age was 53 +/- 15 years, and most patients had severe PH (mean pulmonary artery pressure 48 +/- 13 mm Hg). The highest tertile of RDW predicted death (univariate hazard ratio 4.86, 95% confidence interval 1.37 to 17.29, p = 0.015; multivariate hazard ratio 2.4, 95% confidence interval 1.02 to 5.84, p = 0.045, after adjusting for age, gender, diabetes mellitus, connective tissue disease, diuretic use, phosphodiesterase inhibitor use, hemoglobin, mean corpuscular volume, and blood urea nitrogen [BUN]). Of the laboratory data, only RDW, BUN, and NT-pro-BNP were associated with death on univariate analysis. When RDW, BUN, and NT-pro-BNP were entered into a multivariate model, only RDW was still associated with death (p = 0.037 for RDW, p = 0.18 for BUN, and p = 0.39 for NT-pro-BNP). Adding NT-pro-BNP to RDW did not improve the prediction of mortality. In conclusion, RDW is independently associated with death in patients with PH and performs better as a prognostic indicator than NT-pro-BNP.