We estimate the effect of breast cancer on labour market participation in France, up to five years after the cancer onset. The causal inference is systematically confounded by differences in age, wage at the beginning of career, past health and the past history in the labour market. We account for all these issues thanks to an administrative data set which follows individuals from their entry in the labour market. We find that the detrimental effect of the breast cancer for women increases significantly over time up to 10 percentage points after five years. We also find evidence that the effect of cancer is reduced for younger generations.