In vitro cytotoxicity testing is used increasingly during the development of clinical treatment protocols. These tests are influenced by many variables, not all of which have been assessed systematically yet. We analyzed the influence of the recovery time between the end of treatments and measurements on the detection of cellular resistance. The development of resistance to cisplatin and radiation was chosen as a model since the schedule of these treatments is the objective of several ongoing clinical trials. C6 rat glioma, T98G, 86HG-39, A172 human glioma and TE671 human rhabdomyo-sarcoma cells were pretreated with radiation or cisplatin. The cellular resistance was then compared in pretreated and wild-type cells using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) test. In all cell lines, apparent drug concentrations killing 50% of the cells were dependent on the recovery time. In A172 cells this concentration was 10.3+/-02.1 microM after 48 h but decreased to 3.56+/-0.44 microM after 120 h recovery time (P < 0.001). After recovery times of more than 168 h, 53% of all pretreated cell lines were resistant to cisplatin or radiation, 25% were unchanged and 22% were more sensitive. However, only half the resistant cells could be identified when the MTT test was done with only 48 h recovery time. The sensitivity of detection increased from 0.46 to 0.83 when the recovery time of the test system was extended from 48 h to 168 h. The specificity was not dependent on the recovery time. Experiments showing resistance after short recovery times are reliable, but lack of resistance can only be shown in experiments with long recovery times. Cisplatin treatment can result in resistance to radiation in glioma cells.