In isolated human atrial cardiomyocytes, inhibition of K2P3.1 K(+) channels results in action potential (action potential duration (APD)) prolongation. It has therefore been postulated that K2P3.1 (KCNK3), together with K2P9.1 (KCNK9), could represent novel drug targets for the treatment of atrial fibrillation (AF). However, it is unknown whether these findings in isolated cells translate to the whole heart. The purposes of this study were to investigate the expression levels of KCNK3 and KCNK9 in human hearts and two relevant rodent models and determine the antiarrhythmic potential of K2P3.1 inhibition in isolated whole-heart preparations. By quantitative PCR, we found that KCNK3 is predominantly expressed in human atria whereas KCNK9 was not detectable in heart human tissue. No differences were found between patients in AF or sinus rhythm. The expression in guinea pig heart resembled humans whereas rats displayed a more uniform expression of KCNK3 between atria and ventricle. In voltage-clamp experiments, ML365 and A293 were found to be potent and selective inhibitors of K2P3.1, but at pH 7.4, they failed to prolong atrial APD and refractory period (effective refractory period (ERP)) in isolated perfused rat and guinea pig hearts. At pH 7.8, which augments K2P3.1 currents, pharmacological channel inhibition produced a significant prolongation of atrial ERP (11.6 %, p = 0.004) without prolonging ventricular APD but did not display a significant antiarrhythmic effect in our guinea pig AF model (3/8 hearts converted on A293 vs 0/7 hearts in time-matched controls). These results suggest that when K2P3.1 current is augmented, K2P3.1 inhibition leads to atrial-specific prolongation of ERP; however, this ERP prolongation did not translate into significant antiarrhythmic effects in our AF model.