OBJECTIVES The study aims were to assess the temporal stability following storage at room temperature, the effect of up to 4 freeze-thaw cycles and the effect of simulated freezer failure on measurements of canine N-terminal pro-B-type natriuretic peptide (NT-proBNP) in serum and protease-inhibited (PI) plasma. ANIMALS Twenty-five blood samples were collected from 16 dogs with myxomatous mitral valve disease. METHODS Aliquots of canine serum and PI plasma were stored at room temperature (17-26 °C) for 30 min, 6, 24, 48 and 72 h, respectively. Further aliquots were subjected to between 1 and 4 freeze-thaw cycles. A further aliquot was transferred to storage at 4 °C for 24 h while a paired aliquot remained at -80 °C. All samples were returned to storage at -80 °C until subsequent analysis. N-terminal pro-B-type natriuretic peptide was measured in serum and PI plasma samples using first- and second-generation versions of a commercially-available ELISA. Repeated measures models were used to assess change in NT-proBNP measurements. Wilcoxon signed ranks were used to compare paired measurements. RESULTS N-terminal pro-B-type natriuretic peptide concentrations declined over time in all samples stored at room temperature. Of the four situations tested, the rate of decrease was lowest for PI plasma samples measured using the second-generation assay. N-terminal pro-B-type natriuretic peptide is stable in samples subjected to up to 4 freeze-thaw cycles and in previously-frozen samples stored at 4 °C for 24 h. CONCLUSIONS Use of the second-generation assay, compared with the first-generation, resulted in significantly higher recovery of NT-proBNP measured in PI plasma stored at room temperature. Transport of serum at room temperature for NT-proBNP measurement is not recommended.