Evaluation and management of patients with hypoxaemic respiratory failure in the intensive care unit (ICU) are difficult. The use of B-type natriuretic peptide (BNP), a quantitative marker of cardiac stress and heart failure (HF), may be helpful. The purpose of this study is to describe the prevalence of causative disorders of hypoxaemic respiratory failure in the ICU and to determine the impact of a BNP-guided diagnostic strategy.; This prospective, multi-centre, randomized, single-blind, controlled trial included 314 ICU patients with hypoxaemic respiratory failure: 159 patients were randomly assigned to a diagnostic strategy involving the measurement of BNP and 155 were assessed in a standard manner. The time to discharge and the total cost of treatment were the primary endpoints. Hypoxaemic respiratory failure was multi-causal in 27% of the patients. Heart failure was the most common diagnosis in both groups. The use of BNP levels, in conjunction with other clinical information, significantly increased the detection of HF in combination with an additional diagnosis (32 vs. 16%, P = 0.001) and also increased the application of HF-specific medical therapy (nitrates: 32 vs. 23%, P < 0.05 and diuretics: 65 vs. 50%, P < 0.01). Time to discharge (median, 13 vs.14 days, P = 0.50) and total cost of treatment (median, US-$6190 vs. 7155, P = 0.24) were comparable in both groups.; Hypoxaemic respiratory failure in the ICU is often a multi-causal disorder. The use of BNP increased the detection of HF, but did not significantly improve patient management as quantified by time to discharge or treatment cost. ClinicalTrials.gov Identifier: NCT00130559.