Hypoxic modification by nitroimidazoles has been explored in a large number of clinical studies. Nine different drugs (misonidazole, metronidazole, benznidazole, desmethylmisonidazole, etanidazole, pimonidazole, nimorazole, ornidazole, and RSU 1069) have reached clinical evaluation. Phase I and II trials have demonstrated a relationship between drug tolerance and expected hypoxic modification. Thus, the most tolerable drugs, viz., 5-nitroimidazoles, are those with the smallest preclinical activity. However, they may be the most clinically active, due to higher plasma and tumor concentrations. The clinical evaluation of hypoxic modification is difficult, since clinically important hypoxia can be observed only indirectly. There is, however, indirect evidence of substantial hypoxia in human tumors, although with considerable heterogeneity among individual tumors. More than 7000 patients have been included in 50 randomized trials. A meta-analysis showed that modification of tumor hypoxia significantly improved the loco-regional tumor control after radiotherapy with an odds ratio of 1.17 (95% confidence limits 1.06 to 1.28). The treatment benefit could mostly be related to an improved response in head and neck with odds ratio 1.23 (95% confidence limits 1.09 to 1.37), and to a lesser extent in bladder tumors; no significant effect was observed in other tumor sites (cervix, lung, central nervous system, and esophagus). Similarly to the local control benefit, the overall survival rate was improved with an overall odds ratio of 1.13 (95% confidence limits 1.03 to 1.23). The clinical trials showed no evidence of significant chemosensitization or direct bioreductive effect. Although 50 randomized clinical trials were evaluated, the median number of patients was only 97 (range 17-620). Clinical trials of this size are not likely to detect the observed differences, which may explain the lack of significant improvement in most of the individual studies. However, the overall results indicate that the biological issue related to hypoxia appears to be a sound rationale, especially with regard to head and neck and bladder carcinoma. Future studies related to the hypoxic problem should therefore preferably be aimed towards such tumors.