The progression of late skin telangiectasia after radiotherapy has been studied prospectively in patients for 1, 2 and 5 fractions per week and various dose levels. The degree of telangiectasia was scored on an arbitrary scale. Skin telangiectasia was found to be a continuously progressing endpoint both in the individual patient and in terms of the number of patients who achieved a certain degree of damage. The rate of progression was dose-dependent. Dose-response analysis were performed at 3, 5 and 9 years follow-up for various endpoints: telangiectasia score greater than or equal to 1, score greater than or equal to 2 and score greater than or equal to 3. Iso-effective doses (ED50S) for score greater than or equal to 1 at 3 years, score greater than or equal to 2 at 5 years and score greater than or equal to 3 at 9 years were very similar. In an iso-effect analysis it is therefore worthwhile and time-saving to include the minimal detectable damage in the endpoint (e.g. using score greater than or equal to 1), even if this mild damage is of no clinical significance, and the dose in the endpoint (e.g. using score greater than or equal to 1), even if this mild damage is of no clinical significance, and the dose response become somewhat less steep than for more severe damage. The fact that the progression rate is dose-dependent has impact on dose-response analysis. Dose-response analysis for score greater than or equal to 3 at various follow-up times showed a very flat curve at 3 years compared to 5 and 9 years. The steepness of the dose-response curves was similar at 5 and 9 years. A minimum follow-up of 5 years is therefore necessary for reliable estimation of the late complication rates in a comparison of two dosage schedules using this endpoint. The implication of the continuous progression of telangiectasia is that the dose-response curves are shifted to the left with follow-up. The ED50 is dramatically reduced between 3 and 5 years. The ED50 is also significantly reduced between 5 and 9 years' follow-up, in spite of no change in the steepness of the dose-response curve during this period. Consequently, the time of response is the most fundamental parameter in any iso-effect analysis with progressive endpoints and the late complication rates always have to be specified at a fixed time of follow-up.