To examine the serum bactericidal activity of colistin sulphate (CS) and azidothymidine (AZT) combinations, time-kill curves were performed in native and heat-inactivated human serum with five colistin-resistant and four colistin-susceptible Gram-negative strains. Serum samples were spiked according to median and minimum plasma peak concentrations measured in a phase 1 clinical study in which seven healthy subjects received three (q12h) 1-h intravenous infusions of 4, 2 and 2 MIU colistin methanesulfonate (CMS) co-administered with 200, 100 and 100 mg AZT, respectively. This trial was performed to assess pharmacokinetics and safety of CMS/AZT combination therapy. Minimum bactericidal concentrations of CS in native, but not heat-inactivated, serum were strongly reduced compared with Mueller-Hinton broth for all tested Enterobacteriaceae, except one colistin-resistant (serum-resistant) strain. For colistin-susceptible strains, the minimum CS concentration after 2 MIU CMS was already bactericidal in native and heat-inactivated serum. Median, but not minimum, CS concentrations after 2 MIU CMS were sufficient to kill the serum-resistant, colistin-resistant Escherichia coli strain in native serum. In heat-inactivated serum, even the median CS concentration after 2 MIU CMS was not bactericidal for all colistin-resistant strains. In general, combinations with AZT accelerated killing of colistin-resistant E. coli or showed bactericidal activity even if the substances alone were not bactericidal. Thus, combination with AZT potentiates the bactericidal effect of colistin against colistin-resistant E. coli. Although the dosage of 2 MIU CMS plus AZT may be sufficient to treat infections with colistin-susceptible strains, for infections caused by colistin-resistant E. coli, dosing should be further optimised.