The serum MUC1 markers CASA and CA 15-3 were compared with CA 125 in the serum of patients with ovarian cancer and in pregnant women. Used individually, CASA and CA 15-3 gave sensitivities of 54 and 56% in pre-operative ovarian carcinoma (n = 50), though these were lower than with CA 125 (84%). CASA levels were elevated in 3 women with a negative CA 125, while CA 15-3 was elevated in 2 of these women. The combined use of CA 125 with CASA or CA 15-3 led to the preclinical detection of recurrence in 4/5 patients, with mean lead times of 3.6 and 4.3 months, respectively. Of particular interest was the marked difference in reactivity observed with CASA and CA 15-3 in some patients, despite both assays utilising monoclonal antibodies (MAbs) that react with the MUC1 mucin. CA 15-3 and CASA showed a lower than expected correlation in patients with ovarian cancer (r = 0.70), with some patients having high concentrations of one mucin marker and low concentrations of another. Furthermore, different marker profiles were observed when monitoring the progress of patients with these markers. Marked differences between CA 15-3 and CASA were also observed in the serum of pregnant women (n = 10), where CASA showed marked elevation (mean 33.6 times cutpoint) and CA 15-3 did not (mean 0.88 times cutpoint). These data suggest that the specificities of the MAbs used in these assays affect the glycoform of MUC1 detected, and that it should not be assumed that all MUC1 assays will behave in the same manner.