It is generally assumed that bumetanide possesses some selectivity for the renal Na-K-Cl cotransporter NKCC2, although the results are scarce in the literature and comparisons were done with extra-renal NKCC1 at its basal, almost silent state. Here we investigated NKCC2/NKCC1 selectivity of loop diuretic drugs (bumetanide, piretanide and furosemide) as a function of the NKCC1 activated state (NKCC1 was activated by hypertonic media). NKCC2 activity was measured in isolated rat medullary thick ascending limb (mTAL) and NKCC1 in rat thymocytes and erythrocytes. When NKCC2 was compared with NKCC1at its activated state, all three diuretic drugs inhibited NKCC2 and NKCC1 with the same potency (bumetanide pIC50=6.48, 6.48 and 6.47; piretanide pIC50=5.97, 5.99 and 6.29; and furosemide pIC50=5.15, 5.04 and 5.21 for mTAL NKCC2, erythrocyte NKCC1 and thymocyte NKCC1, respectively). Basal NKCC1 exhibited a lower diuretic sensitivity, although with marked differences depending on the diuretic drug and the cell type in consideration and with the notable exception of furosemide in erythrocytes. Molecular modelling showed that bumetanide and piretanide possess four potentially active groups, of which three are shared with furosemide at similar intergroup distances. Of these three common groups, one should not bind to basal NKCC1 in thymocytes. The fourth (phenoxy) group (absent in furosemide) confers higher lipophilicity and should not bind to basal NKCC1 in erythrocytes. In conclusion, loop diuretics had no NKCC2/NKCC1 selectivity, when NKCC1 is measured at its activated state. Basal NKCC1 has a reduced diuretic sensitivity, of very different magnitude depending on the diuretic drug and cell type in consideration.