In univariate analysis, local control rate as well as hormone substitution free survival were estimated using the Kaplan Meier method. To test the statistical significant differences between survival curves, the log rank test was used. Data are given in median (range) or in percentages. Frequencies of hormone deficiencies were compared in Chi-square analysis.
Life expectancy was studied after transformation of survival time to standar-dized survival time (SST) to adjust for background mortality in the general population. SST is the quotient of observed survival time and median life expectancy in the general Dutch population matched for gender, age and year of operation. These life expectancies were derived from the data provided by the Dutch authorities (www.cbs.nl). The analytical background of this method has been reported elsewhere19. A two-sided p-value <0.05 was considered to be significant.