Current risk stratification in acute pulmonary embolism (APE) includes assessment of clinical status, right ventricular overload and plasma troponin concentrations. As impaired renal function is one of the important predictors of mortality in cardiovascular diseases, we hypothesized that it is an independent early mortality marker in APE. In prospective cohort study, we observed 220 consecutive patients (86M/134F, 64 +/- 18 years) with APE proven by spiral computed tomography (CT). On admission, echocardiography was performed and blood samples were collected for troponin and creatinine assays. The calculated glomerular filtration rate (GFR) differed significantly between 81 pts with low-, 131 pts with moderate- and 8 pts with high-risk APE [71 (19-181) vs. 55 (9-153) vs. 41 (14-68) mL min(-1); respectively P < 0.0001]. Twenty-three patients died during the 30-day observation. Importantly, GFR was lower in non-survivors than in survivors [35 (9-92) vs. 63 (14-181) mL min(-1), P < 0.0001]. The area under the curve (AUC) of the GFR receiver-operating characteristic (ROC) curve for predicting mortality was 0.760 (95% CI: 0.698-0.815). In multivariable analysis, independent mortality predictors were GFR, troponin, heart rate and history of chronic heart failure. In normotensive patients, the GFR and cardiac troponins (cTn) ROC curves for prediction of mortality showed no difference (AUC 0.789 and 0.781, respectively). However, Kaplan-Meier analysis showed an additive prognostic value of renal dysfunction. Thus, troponin-positive patients with a GFR < or = 35 mL mn(-1) showed 48% 30-day mortality, whereas troponin-positive patients with a GFR > 35 mL mn(-1) had 11% mortality, and troponin-negative patients with a GFR > 35 mL mn(-1) had good prognosis, P < 0.0001. Impaired kidney function, present in 47% of APE patients, is related to all-cause mortality. In initially normotensive patients, a GFR < 35 mL min(-1) predicts 30-day mortality. Moreover, GFR assessment can improve troponin-based risk stratification of APE.