This article examines the distribution and prognostic importance of urinase type plasminogen activators (uPA) and of plasminogen activator inhibitors (PAI-1) in cases of primary oral squamous cell carcinoma. Tissue from the primary tumor was taken from 79 patients. In order to make an intra-individual comparison, tissue from the healthy mucous membrane of the mouth was taken from 50 patients and metastatic tissue from lymph glands in the neck from 16 patients. The content of uPA and PAI-1 was determined using ELISA. After follow-up, 58 patients with primary surgical therapy were included. Statistical evaluation was carried out using the Kruskal-Wallis test, the Mann-Whitney U-test and the Wilcoxon test. Pearson's product moment correlation was used to determine the relationship between uPA and PAI-1 levels. The median uPA value was 3.43 ng/mg in primary tumor, and for PAI-1 47.1 ng/mg ( n=79). There was a significant correlation between uPA and PAI-1 both in the cancerous as well as the healthy tissue ( P<0.01). The intra-individual comparisons showed uPA and PAI-1 differed significantly between cancerous and healthy tissue ( P<0.0001) with the mean uPA and PAI-1 values being nine times higher in the cancerous tissue ( n=58). The correlation for between uPA and PAI-1 in tumors, healthy tissue and metastatic lymph node tissue ( n=16) showed highly significant values in the tumors ( P<0.001). The comparison between cancerous tissue in the primary tumor and the lymph nodes was not significant for PAI-1. For uPA, the values in the lymph nodes were significantly lower ( P<0.049). There were also significantly higher levels in metastatic lymph node tissue compared with healthy mucous membrane ( P=0.005 for uPA and P=0.003 for PAI-1). There was no significant correlation of PAI-1 and uPA ( n=79) with the patient's sex, size of the tumor (T stage), nodal status (N stage), differentiation (grade), or residual tumor status. If the patients were divided into two groups (< or =58 years and >58 years), the older patients had higher uPA ( P<0.017) and PAI-1 ( P<0.02) levels. The was no significant association between tumor localisation and uPA content in the tumor; for PAI-1 the association was significant ( P<0.02) in the individual areas of the mouth. A total of 23 (40%) patients relapsed (local n=13, lymph node n=3, local and lymph node n=1, lymph node and skin n=1, other locations n=5). Such patients had raised uPA ( P=0.012) and PAI-1 ( P=0.014) levels in the primary tumor. The high variability of the normal clinical parameters in tumors only has a limited prognostic value because it is not taken into account in individual cases. Thus determination of the PAI-1 level directly after surgery could provide an indication of the likelihood of a relapse and thus aid in determining the adjuvant therapy. This confirms a trend in that tumor associated proteases can also play a key role in oral squamous cell carcinoma as new, independent, prognostic factors. Whether or not uPA and PAI-1 will play such a role will be determined in additional multicentre clinical studies.