In order to evaluate the influence of bladder filling at the beginning of a 40 minutes pad test, 71 women completed this study. Thirty-six patients were randomized to pretest filling of the bladder to 50% of the bladder capacity, and 35 patients to 75% pretest filling. In all patients, a retest was performed approximately fourteen days later. The leakage in the two groups was equal. Median leakage during the first test was 3 g in both groups (p = 0.97). The leakage was numerically larger during the second test, median 6.5 g and median 13 g, respectively (p = 0.69). The test-retest variation was calculated in both groups. In patients with 50% bladder filling test-retest variation (mean +/- 2 S.D.) was -28.4% +/- 206.8%. In patients with 75% bladder filling mean variation was -58.0% +/- 203.0%. The difference in test-retest variation between groups was not statistically significant (p = 0.17). Only the subjective evaluation by the patients revealed a preference in favor of the test with 50% bladder filling. Significantly more patients in the 50% group reported that the results of the first test and the retest corresponded well with the daily urinary leakage (p = 0.04). The most striking finding of this study was that, despite filling to either 50% or 75% of the bladder capacity at the beginning of the test, the fluid load during the test, i.e. the initial volume instilled into the bladder plus the diuresis during the test, was equal in the two groups.(ABSTRACT TRUNCATED AT 250 WORDS)